RU38235U1 - SYSTEM OF REMOTE IDENTIFICATION OF SMALL-SIZED OBJECTS - Google Patents

Description

Small-size Remote Detection System

The proposed utility model relates to techniques for examining, for example, minefields and, accordingly, can be used for searching, detecting and recognizing mines, including, for example, anti-tank mines and landmines, covered by a layer of soil.

A device for synthesizing a radar image containing a driver of the reference function in the form of a read-only memory (ROM), a complex multiplier, two blocks of fast Fourier transform, an amplitude detector, as well as a driver of a two-dimensional matrix signal and a storage device (see, for example, RF patent No. 2211461 with a priority of 06/18/2001, О 01 S 13/90).

In the known device, the computing environment is invariant, invariant to the trajectory and motion parameters, and the possibility of incoherent synthesis of the radar image of the probed region is provided.

However, this device does not provide the ability to conduct a search, as well as the detection and recognition of small objects.

The closest analogue prototype is a subsurface sounding system based on the principle of multi-frequency sounding of conservative media

IPC: G 03 V 37/00, G 06 K 9/00

objects

imiiGji i

(building structures, soils, etc.) and containing a radar sensor connected to a computing device connected to an indicating device (see, for example, VN Sablin “New technologies in humanitarian demining. Proceedings of the anniversary scientific and technical conference, dedicated to the 30th anniversary of the foundation of TsNIIRES, published by JSC TsNIIRES, 2001, part 1, pp. 20-23).

This system provides the possibility of scanning and, respectively, sensing the surface and allows the detection and identification of plastic and metal anti-tank and anti-personnel mines buried in the ground at (1-10) cm, however, the possibility of using it to remotely search for small-sized objects has not yet been constructively developed.

The objective of the utility model is to develop a remote system, for example, from an aircraft, search, detection and recognition installed both on the underlying surface and buried small objects, for example, different types of mines.

The essence of the invention lies in the fact that in the system for the remote detection of small objects, including a radar sensor connected to a computing device connected to the indicator by its output, a matching unit, a radar image forming unit, a classifier, two memory units, a coordinate determination unit, a synchronization unit are introduced printing and radio transmitting and receiving devices, as well as additional indicator and computing device, and the radar sensor, the first exit ohms is connected to the first input acceptance unit, its output connected to a first input of forming unit radar

image connected to the first input of the first memory block, the output connected to the first input of the radio transmitting device, the second (control) input connected to the first output of the computing device, the second output connected to the second (control) input of the first memory block, the third output connected to the control input of the radar sensor, and the fourth output connected to the input of the synchronization unit, its first and second outputs connected to the second (synchronizing) inputs, respectively about the matching unit and the radar image forming unit, and the group of outputs connected to the group of corresponding inputs of the radar sensor, the second output connected to the third input of the matching unit, while the computing device is connected to the input-output of the first memory unit by the input and output, and the fifth and sixth outputs connected respectively to the first inputs of the indicator and the coordinate determination unit, with its first and second outputs respectively connected to the first input of the computing device properties and the second input of the indicator, and the radio input associated with the radio input of the system, moreover, the radio transmitting device is connected with its radio output to the radio input of the radio receiving device, the output connected to the first input of the second memory unit connected to the first input of the classifier, the output connected to the first input of the additional computing device with the first, second and third outputs connected respectively to the second inputs of the radio receiver, the second memory block and classification ora, fourth and fifth outputs connected to first inputs of the second indicator and the printing apparatus

 1(

respectively, a input-output connected to input-output of the second memory block, while the inputs of the computing and additional computing devices are connected respectively to the first and second inputs of the system, the output of the printing device is connected to the output of the system, the optical outputs of the indicators are connected to the corresponding optical outputs of the system, and the group the radio inputs / outputs of the radar sensor is associated with the corresponding group of inputs / outputs of the system.

Moreover, the radar sensor contains a phased antenna array consisting of N independently connected nriemo-transmitting modules, where, for example, N (1, ..., 48), an adder, two switches, a generator, a modulator and a unit for generating a given distribution of amplitudes and phases signals to control the position of the antenna beam for each of the transceiver modules, and the transceiver modules are connected by the first electrical inputs to the corresponding outputs of the first switch, the first input of the generator connected to the first output a, the first (control) input connected to the input of the radar sensor, and the second input connected to the output of the modulator, the input connected to the first input of the group of inputs of the radar sensor, while the second input of the group of inputs of the radar sensor is connected to the input of the unit for generating a given distribution of signal amplitudes and phases to control the position of the antenna beam for each of the transceiver modules, its first and second outputs connected respectively to the second input of the first switch and the first input of the second switch, its second input connected to the third input of the group of inputs of the radar sensor, and the corresponding N outputs connected to the second electrical inputs

(

corresponding transceiver modules, and

the electrical outputs of the transceiver modules are connected to the corresponding inputs of the adder connected to the first (electrical) output of the radar sensor, the second output of the generator is connected to the second output of the radar sensor, the radio inputs and outputs of which are connected to the corresponding radio inputs and outputs of the transceiver modules.

The proposed system provides the ability to conduct contactless for the appropriate operator, such as mineral, search with a high probability of detection and recognition of small objects, such as mines, including those buried in the ground and, accordingly, masked.

The fact is that in solving the problem of detecting and recognizing small targets, including different types of mines, the determining factor is the ratio of the reflection powers of the radar signal from the object and directly from its environment (signal / background ratio).

The background power in the resolution element depends on the specific effective scattering surface (EPR) of the background and the linear resolution of the radar sensor (RLD) and is determined by the ratio

where k is the coefficient taking into account the attenuation of the signal during propagation in the medium, cUv is the specific EPR of the background, 5 / and SD is the linear resolution along the transverse and longitudinal coordinates of the area of the area irradiated when the object is detected, and the power of the signal reflected from the object is determined as

(

Pf k (UfdpSD,

Pc k (yo,

The features associated with the subsurface arrangement of the object (mine) significantly complicate the structure of the PC signal formation, therefore, the problem of its production is usually solved experimentally. It is known 8 that when detecting on a site with a size of m x 1 l and using radar radiation in the decimeter range of radio waves at a frequency of GHz (0.3 l) for a buried metal anti-tank mine of type M-20, the ratio PC / RF 2 (or Zdb) takes place, moreover the probability of detecting mines in such conditions is less than 0.1.

The detection probability in this case can be increased by improving the resolution of the RLD due to, for example, the use of a broadband (in the range of operating frequencies) active phased array antenna (AFAR) as a part of the RLD.

Focusing on the size of the operating frequency band (Af) of the AFAR of the order of (250-300) MHz, in accordance with the dependence

where m „is the equivalent duration of the compressed pulse, we obtain the value of the equivalent duration of the compressed pulse r„ (3.3 - 4) is clear.

Then, according to the ratio of 5rg „„ / 2, where with 3x10V / c the speed of propagation of radio waves, the resolution in radial range is 5rg (0.5-0.6) m.

In this case, the horizontal resolution (underlying surface) will be:

where wu is the angle between the direction of observation and the underlying surface (angle of observation). Given this feature, at the near boundary of the radar observation section, we have a woo 60 ° and, accordingly, we obtain 5D (1.0-1.2) g.

H 1 / Af,

SD Pr / cosdy,

The potential resolution of the RLD with a side view of the underlying surface along the path line (Sp, describes the relationship

8px / I Go / 2L sindx, (1)

where I is the length of the radio wave of the RLD sounding signals, g is the slant range to the middle of the section, L is the path length of the RLD antenna moving, and c is the angle of observation relative to the aircraft ground speed vector.

Substituting into the formula (1) the value of the carrier frequency of the probing RLD signals from a certain part of the decimeter range (0.5-1.5) GHz, for example, its average value equal to 1 GHz, and also taking the angle 0x90 °, the value of Go 85 m and the path length Z, 25 m during the flight of the aircraft for 1 s, we obtain the radial resolution Spx of 0.5 m.

However, in reality, when compensating for the influence of various factors, such as, for example, the formation of false marks due to the high level of the side lobes of coherent signal processing, which is approximately -13dB relative to the main maximum, and to increase the stability of the information processing, the use of weight smoothing functions, there is a deterioration in the resolution 11. So, taking into account such processing, the real value of the resolution in the transverse direction (cross-distance) at 90 ° m Spx (0.7-1) m can be expected, given that “observation frames and at. 30 ° the value of f in these frames will double. However, for them the range resolution will be almost 2 times better than for “observation frames at 6–90 °. On average, for all “frames of coherent processing at a selected elementary inspection interval of 1 s duration, Spx I m W. SD 1 m can be taken.

l “MIW (i (

MWC

Further improvement in the resolving power of SAR is possible by introducing incoherent processing of separately formed “observation frames during coherent processing. The receipt of these results is based on the use of reconstructive computational tomography methods, in particular, on the basis of the application of the Radon transform 12.

An assessment of the achievable resolution using such methods can be obtained, for example, by constructing a two-dimensional portrait in a Cartesian coordinate system from a set of projections — one-dimensional portraits synthesized at small angular intervals, shifted by angle relative to each other, where the corresponding values of the potential resolution for the central points synthesized two-dimensional portrait obtained in the form of the following dependencies

 Here, D is the angular interval of observation, equivalent to the total angle of projections - one-dimensional portraits synthesized at small angular intervals, shifted in angle from one another.

Hence, with D 71 we have

/.f-0.16X

2nd Presented estimates are obtained on the basis of the formation of the uncertainty function (FN), which corresponds to coherent signal processing 9 In this method of synthesizing at small angular intervals a two-dimensional portrait from its one-dimensional projections is carried out without taking into account the nonlinear dependence of the phase on the angle, which actually worsens the resolution characteristics with respect to potentially achievable .

(2)

--0.37X.

F.

 spatial selection along the segment p fx 9y) in the polar coordinate system is formed on the basis of determining the width of the FN module at the level of 0.7 (-Zdb) 10. Moreover, for circular viewing at an angular size PC A D 2, the i-potential resolution in all spatial coordinates is 6 p 0.18 L which at a hypothetically accepted wavelength D 0.3 l gives 6 p. (5 m.

When forming a quasi-hologram, the formal possibility for each radio image of a locality site is taken into account, with the help of the subsequent Fourier transform procedure, to get the original radio hologram again. In this case, such a procedure is carried out for each projection corresponding to the Radon transform of the radio image of the selected area, and for all the received projections of this area, a set of individual radio holograms is formed, each of which is characterized by the display of the area under different observation angles. Combining such radio holograms with the necessary phasing them provides a combined radio hologram of the total interval of viewing angles of observation of the selected area. However, the implementation of such a procedure is fraught with difficulties in realizing the necessary phasing and taking into account the nonlinearity of the formation of the integrated radio hologram, which becomes circular at large viewing angles. In this regard, the actually obtained representation of such a radio hologram is called a quasi-hologram.

In the absence of coherence, when “sewing separate” frames of coherent processing are performed at elementary intervals of the examination, there is a broadening of the FN and, as a result, a significant deterioration in resolution. However, this circumstance can be compensated, for example, by the telescopic survey mode proposed in this technical solution.

1 & OH1 (I

the terrain at which the resolution decreases by only 2-3 times and reaches a corresponding value of f 0.15m. The magnitude of the resolution over all spatial coordinates Sp 0.25m should be expected with a 3-fold review of the terrain with a change in the course angle of the rectilinear trajectories of the helicopter over the probed terrain, and with a single survey of the terrain and incoherent processing of information, for example, seven frames of review of coherent processing, resolution in all spatial coordinates will be SP 0.5m.

Providing a resolution of Sp 0.5m in all coordinates compared to a resolution of Sp 1m gives an additional increase in the snngal / background ratio by 4 times (by 6 dB) and its total value is 9 dB. In this case, in accordance with the data 10 that are close to this topic in the literature, the probability of detecting an anti-tank mine of the type will be 0.25. Using the next resolution level Sp 0.25m will ensure the detection of such mines with a probability of at least 0.95. In the case of smaller mines, in particular anti-personnel mines, a resolution of SP 0.25l will allow them to be detected with a probability of at least 0.95. Higher resolution Sp --0.15m in the telescopic viewing mode will make it possible to fix the radio image of the detected mines with the peculiarities of their formation “brilliant during radio exposure (radio sounding). This, in turn, when using this technical solution, allows 11 types of mines to be recognized using, for example, neurostructures that in this case recognize, in particular, anti-tank mines with a probability of at least 0.95 and anti-personnel mines with a probability of at least 0 ,9.

10

In Fig.1 shows a structural diagram of a system for remote detection of small objects, in Fig. 2 is a structural diagram of a radar sensor; FIG. 3 shows an example of conditional marking of the examined field; in FIG. 4 shows a diagram of scanning during an elementary time interval during a flyby, in FIG. 5 shows a scanning scheme for a “telescopic examination, in FIG. 6 shows an example of the excess of signals reflected from different parts of the underlying surface of the threshold level signals set to 8 dB, in FIG. 7 shows a comparison of the detection probabilities using known methods and the proposed method based on reconstructive tomography using the Radon transform.

The system for remote detection of small objects, for the implementation of the method (Fig. 1) contains a radar sensor (RLD) 1, designed to respectively provide radiation of signals, for example, the decimeter range of radio waves, receive reflected signals when sensing the underlying surface and generate appropriate signals for subsequent processing. RLD 1 with the first output connected to the first input of block 2 matching (BS), made in the form (not shown) of a series-connected frequency converter and ADC 11 and its output connected to the first input of block 3 of the formation of radar images (BFRLI), made in form, for example, processor device Baguette 14.

The output of the radar image forming unit 3 is connected to the input of the memory unit 4i, made in the form of random access memory, for example, on a PCL-6646B 15 type board, with an output connected to the first input of the radio transmitting device 5, intended for transmission

Ifimw

12

information received during the flight of the aircraft and recorded in the memory unit 4i for subsequent processing to the corresponding ground center (not shown in Fig.) and made in the form of an appropriate device like UC4 Marcad Delivery Receiver of the UURE radio system from SHURE 16.

The second (control) input of the radio transmitting device 5 is connected to the first output of the computing device 6, designed to control the operation of the associated units and devices of the system and to process and transmit signals of these units and related devices and made in the form of an industrial computer IPPC-950T by Advantech 15 its second output connected to the second (control) input of the first memory block 4i, a group of outputs connected to the first group of inputs of the RLD 1, and a third output connected to the input of the block 7 syn ronizatsii configured as a clock generator 17.

In addition, the computing device 6 is connected by its input-output to the input-output of the first memory unit 4i, and the fourth and fifth outputs are connected respectively to the first inputs of the indicator 8i and the coordinate determination unit 9, which is connected by its first and second outputs to the first input of the computing device 6i and the second input of indicator 8i, and the radio input associated with the radio input of the system.

Indicator 8i is designed to create a visual representation, for example, of an aircraft pilot, about a given minefield flight path and the degree of completion of this task and is made in the form of a corresponding device of type GV-D900 16.

Block 9 determining the coordinates (BOC) is intended to obtain information about the coordinates of the investigated sections of the underlying surface and the location of the detected small objects, as well as the coordinate reference of the aircraft and

made in the form of satellite radio navigation system SRNK-21DM 18.

The synchronization unit 7 is connected with its first and second outputs to the second (synchronizing) inputs of the matching unit 2 and the radar image forming unit 3, and the output group is connected to the group of corresponding inputs of the RLD 1, the second output connected to the third input of the matching unit 2.

The radio transmitting device 5 is connected with its radio output to the radio input of the radio receiving device 10, intended for receiving information from the aircraft for subsequent processing by the respective devices of the system and made in the form of a corresponding device of the UC1 type of the radio system of the UC series of the company SHURE 16.

The radio receiving device 10 is connected by an output to the first input of the second memory block 42, made in the form of random access memory, for example, on a PCL-6646B 15 type board and connected to the first input of the classifier 11, which is designed to implement the detection of small objects and their recognition and made in the form of a neural network microprocessor NM 6403 19.

The classifier And output is connected to the first input of an additional computing device 62, designed to process incoming information signals and build a map of the location and coordinate reference of small objects, for example, mines in the examined area, as well as to ensure the interaction of the corresponding blocks and devices of the system and executed in as an industrial computer IPPC950T by Advantech 15.

lf {

thirteen

the radio receiving device 10, the second memory unit 42 and the classifier 11, the fourth and fifth outputs to the first inputs of the additional indicator 82 and the printing device 12, respectively, and the input-output is connected to the input-output of the second memory block 42.

Additional indicator 82 is made in the form of a corresponding device of the type GV-D900 16, and the printing device 12 is made in the form of an hp LaserJet 1200 printer.

In this case, the group of radio inputs-outputs of the RLD 1 is connected with the corresponding group of inputs and outputs of the system, the inputs of the computing devices 6 and 62 are connected respectively to the first and second inputs of the system, the optical outputs of the indicators 8i and 82 are connected with the corresponding outputs of the system, and the output of the printing device connected to the system output.

The radar sensor 1 (Fig. 2) contains an antenna device (not shown in Fig.), Made in the form of an active phased antenna array (AFAR) 20, consisting of two antenna arrays (not shown in Fig.) Along M independently connected transceiver modules 13 in each, where M 1, ..., 12, and is a separately installed antenna array, as well as an adder 14, two switches 15, a generator 16, a modulator 17 and a block 18 for generating a given distribution of signal amplitudes and phases to control the antenna position beam for each of the reception o-transmitting modules 13, and the transceiving modules 13 are connected with the first electrical inputs to the corresponding outputs of the first switch 51, the first input connected to the first output of the generator 16, and the second input connected to the first output of the block 18 of the formation of a given distribution of amplitudes and phases of the signals to control the position of the antenna beam of each of the transceiver modules 13, the second electrical inputs of which are connected to

jMMWi

14

the corresponding outputs of the second switch 152, the first input connected to the second output of the unit 18 for generating a given distribution of amplitudes and phases of the signals to control the position of the antenna beam of each of the transceiver modules 13, and the second input connected to the first input of the first group of inputs of the radar sensor 1, the first output connected to the output of the adder 14, the corresponding inputs connected to the outputs of the respective transceiver modules 13, and a second output connected to the output of the generator 16, the first input of the modulator 17 connected to the output, its first input connected to the third input of the first group of inputs, and the second input connected to the first input of the second group of inputs of the radar sensor 1, while the second and third inputs of the second group of inputs of the radar sensor 1 are connected to the second inputs, respectively the generator 16 and the block 18 of the formation of a given distribution of the amplitude and phase of the signals to control the position of the antenna beam of each of the transceiver modules 13, which are radio-input-output connected are associated with the corresponding radio inputs and outputs of the radar sensor 1.

Transceiver modules 13 are designed to emit microwave signals of the decimeter range and receive the signals reflected, for example, from the underlying surface when it is sensed, and are made in the form of, for example, elements of the type M-2730 21. The adder 14 is made in the form of a microwave distributor 22 and electrical inputs connected to respective outputs.

The switches 15 are made in the form of corresponding devices 17, and the switch 15i is designed to switch modes (transmission or reception) of the operation of the transceiver modules 13 and the corresponding distribution

) Ш1С

ic

fifteen

power of the microwave signals supplied to these modules from the generator 16, made on the basis of a high-frequency transistor, for example, a transistor of the type 2T634A-2 20, and the switch 152 is designed to transmit to each of the transceiver modules 13 the corresponding parameter values of the corresponding signals from the generating unit 18 a given distribution of amplitudes and phases of the signals to control the position of the antenna beam.

The modulator 17 is designed to form parameters of the emitted pulse signal, namely: duration, repetition period and frequency modulation, and is made in the form of a chip of the type AT90S4433 21, and the block 18 of the formation of the current law of the distribution of amplitudes and phases of the signals to control the position of the antenna beam is made in the form read only memory (ROM) type 82S191 17.

The method of increasing the resolution of radar sensing is based on developments in the relevant fields of science and technology and can be briefly illustrated by the example of considering the problem of remote from an aircraft, such as a helicopter, detection, including search, detection and recognition, installed both on the underlying surface and in-depth small objects, including various types of mines.

The ability to detect small objects is determined by the ratio of the power of the reflection signals from the object and the background, determined by the size of the underlying surface area, the dimensions of which, in turn, are due to the resolving power of the radar sensor (RLD) in the longitudinal and transverse directions of its beam. In the case when the object of radar sensing is located below the surface, the passage of additional factors worsening the ratio of the signal powers from the object and the background

At 101 l.

16

radio emission through the soil layer to the object and vice versa, as well as energy dissipation of the reflected signal at the media interface 7. A generalized representation of the depth of penetration of radio waves into the soil depending on the frequency of radiation of the probing signals and soil moisture is presented in the indicated materials for radio waves mm-, cm-, dm and meter (l) -bands of radio waves 23. On a large part of the earth's surface, soil moisture is such that the signals of the sui range penetrate by units of centimeters, and the signals of the dg-range penetrate by units of decimeters. The best subsoil penetration is provided by the meter-wide radio waves characteristic of the composition of the spectrum of short-pulse radiation in the form of a single oscillation period of the probing signals. However, their use is associated with large energy losses due, in particular, to a significant deterioration in the directional properties of radio waves with a decrease in their carrier frequency. Acceptable penetration into the soil for the detection and recognition of, for example, anti-tank mines at a depth of up to 0.3 m is provided by 3 l “-band radio waves, and a significant increase in the resolution of radar sensing when using dm-band radio waves is achieved by using the aperture radar synthesizer (PCA) antenna mode in discrete viewing mode using electronic beam control of the RLD antenna, and a feature of the organization of the SAR antenna of the RLD is to provide not fulfilling the requirement that each point of the surveyed area is observed in this case for several seconds.

The information obtained on the intensity of reflection from the underlying surface at 8 positions of the AFAR beam for an elementary inspection interval of 1 s is algorithmically

(

Shts1 (M

17

transferred to a horizontal surface corresponding to the display of 7 sections of the underlying surface measuring 60mx25m, occupying a total sector ± 60 ° of the side view of the terrain when moving during this time, the real RLD antenna along the helicopter path. This information obtained (see Fig. 3) from the right and left sides of the helicopter along its path on each elementary inspection interval is recorded in the RAM of the processor device “Baguette RLD.

The stored information on the reflection intensity at 7 Zastka of the underlying surface on each side of the helicopter along its straight path during 7 elementary inspection intervals allows, as shown in Fig. 4, to compile a matrix (7x7) of data of coherent processing of reflection signals at each of these intervals. The main diagonal of such a matrix corresponds to information, in particular, for the VII section of Fig. 4, under 7 different angles of observation along the path of a direct flight of a helicopter.

A similar accumulation of coherent processing data at successive elementary inspection intervals takes place when organizing the telescopic viewing mode shown in FIG. 5. In the case where at the turn the helicopter speed is the same as in the case of a straight flight of Fig. 4, during a circular flight of one section with a diameter of 25 m, the accumulation of coherent information under 15 different angles of its radar observation from the helicopter.

By processing information on 7-15 sections of the underlying surface for an elementary inspection interval, a synthesized antenna pattern is formed with incoherent processing of information corresponding to a certain sector of its sight from the side of the helicopter. According to the above calculation results when using this technical solution and application during processing

M /

 // 7 / M 18

of the received signals, for example, the Radon transform, the sounding resolution can reach (0.15-0.5) m.

The Radon 12 transformation is applied to the case under consideration by determining the projection of some unknown function g (x, y)., Which describes the location in Cartesian coordinates of the object’s points on the ground, and the image of the desired object is formed from measurements of an available limited set of projections. These projections as functions of the observation angle ft taking into account the introduction (5-functions are determined by the expression:

p (u, y5) J g (x, y} S (x cosj3 + y sin - u} dxdy, (3)

where and is the length of the perpendicular dropped from the origin to each of the lines intersecting the function display area

g (x, y

In addition, it is assumed that one- and two-dimensional Fourier transformations of the functions p (u, ft) and g (jc,),

denoted respectively by Pj (w, p) viGi (a) cos ft, with sin ft).

Then, in accordance with the theorem, which states that by enumerating all possible values of the angle ft, we can obtain all possible values of the central section of the function Gy (with cosft, (О sinft), which is equivalent to its complete definition 27, we have:

P, (", y) G, ((, 6) smyff), (4)

and, if we then perform the inverse Fourier transform for this function, then the corresponding function g (; c, j) will be determined (restored).

A more detailed analysis of the formation of relation (4) allows us to draw an analogy for the formation of the resulting hologram for individual holograms of objects with scattering functions that coincide in a certain region. However, in contrast

(

19

from a rigorous solution to this problem 28, in this case, the concept of a quasi-hologram should be used.

A perceived analogy to the formation of the resulting hologram is that on a limited segment of each isodal of the radar image along it there is a restored partial image of the object, which, in accordance with expression (3), is its projection from the angle D The offset parameter and relative to the origin of the coordinate system of parallel lines intersecting the geometric object g (x, y), characterizes the location of the total scattering intensities at the points of the narrow band of each radar isodal.

Performing the one-dimensional Fourier transform of such an object projection returns to the complex representation area of it as quasi-holograms of recording the field scattered by the object. The subsequent combination of all such quasi-holograms based on expression (4) characterizes the formation of the resulting quasi-hologram. If, under this condition, the inverse of the two-dimensional Fourier transform of the resulting quasi-hologram is fulfilled, then the geometric image of the object g (x, y) is restored from the totality of its projections of the form p (u, / 3) from different angles.

Such an interpretation of relation (4) allows the conditional use of estimates of potential resolving power in the area of radar sensing, calculated using formulas of type (2), which are valid for coherent processing of radar information. However, this defines a convenient computational scheme for processing incoherent radar information. In this case, such a scheme is provided by using the convolution and reverse projection algorithm. The starting point of such an algorithm

Bbg /

2 luMw

the inverse Fourier transform of equality (4) is used, which corresponds to the following calculation formula

gi (x, y} P (u, fj (x cosJ3 + y sin J3 - u) dudp, (5)

where is the weight function of the influence of the “projection range / fw ;;

ADm) || d) | exp (7th) m). (6)

The image reconstruction algorithm g (x, y) is implemented on the basis of formula (5) in the form of the following stepwise procedure

Step 1: convolution or filtering. For each projection p (u, j3), a convolution is performed with the function (core) hc (u}., Forming filtered projections:

Rya (p (u, p} h () du (7)

When determining the values of the kernel hc (u) in its calculation formula (6), the integrand is multiplied by the window function W (G)). Step 2: Reverse Projection The approximation of g (x, y) to the original function gi (x, y) at each point is restored by summing the values along each line passing through this point and xcosp + ysinft:

g, (x ,, (+ y, inl3, l3} dft. (8)

Thus, the considered RWT algorithm realizes obtaining filtered values of the projections of the object for each observation angle, and then integrating the results of the observations over the variable. The combination of dependences (7) and (8) forms expression (5) and formally leads to an exact solution when for all w the value is W (o)) 1. However, in this case the integral (6) diverges and the concept of an exact solution loses its meaning. Therefore choose

moii

21

window function Scho)} such as to limit the contribution to the solution of frequencies above a certain cut-off frequency coC.

Ideally, the Fourier transform of the function hc (ti) taking into account the window function, i.e. I with I Що)}, should have an approximately linear response up to a certain boundary frequency, after which it should fall to the point at 6} (Os. This requirement is satisfied by atomic functions, which in this case provide a better approximation gi (x, y) to the original function gj (x, y) 30.

To illustrate the presented relations, an example explaining them in a simplified formulation is considered. Suppose that a number of local radar inhomogeneities (LHs) of the observed area are clearly detected by radio holographic processing of information at elementary intervals of the survey. The other part of the LV, including small-sized objects, including mines, can be detected by increasing the resolution of the RPD during the further processing of radar information based on the use of relations (7) and (8).

The task is to construct a two-dimensional portrait in the Cartesian coordinate system based on the totality of projections — one-dimensional portraits synthesized at small angular intervals, shifted by angle relative to each other 13. It is assumed that the analyzed radar images both detected at the first stage of processing and undetected are in the vicinity of the beginning coordinates of the viewing area at the current inspection interval. Relative to the point at the origin of this zone, the movement of the helicopter from the RLD along an arc of a circle of radius RO is considered.

In reality, there is a more complex displacement of the RLD relative to the origin, however, to simplify the task, we assume that the change in the angle of the observation site on the elementary inspection interval is small and then RO const const. Additionally, it is also accepted that the points of the intense radio path 22i (

of the object (“shiny points of the LN”), hereinafter referred to as BT, are located in the plane of the RLD movement along the radius Rg.

The amplitude-phase characteristic of the reflected signal model at () has a peak structure, which determines the possibility of identifying and tracking peaks due to the same BT of neighboring portraits. Joint processing of Doppler frequencies (Oj (J3i) peaks (/ is the number of the portrait of the object with the middle of the observation interval D), identified in neighboring portraits, allows one to obtain the radar image of the object in the form of an estimate of the Cartesian components of the radius vectors of its BT: pj (pjx, pjy). also, that in the measuring device, along with the useful signal y (ft), there is an additive noise distributed according to the normal law with zero mean and dispersion as.

Within the framework of the adopted model, the signal y () is a superposition of harmonic signals with unknown amplitudes, frequencies, and phases. Using the maximum likelihood method allows you to determine these unknown parameters, provided that the peaks of the harmonic signal are resolved. The obtained estimate (D) by the maximum likelihood method in asymptotics obeys the normal law with an average equal to the true value of the Doppler frequency of the BT at a given D: Mi), (E) {pjxcosj3i - pjySin i) and the dispersion B ()) equal to the potential accuracy of the frequency measurement harmonic signal 6 d / 3 / d lr, where q is the signal-to-noise ratio for a given BT and is the angular step of measuring the observation of the radar angles of the object.

The structured formulas presented in (13) correspond to expression (8). In this case, due to the linear dependence on the frequency a) (p) with its measured average value, we have the average value of the coordinate estimates pj, pjy and they coincide with their true values. The latter determine

position on the BT plane of a two-dimensional portrait of the restored object. The quality of constructing a two-dimensional portrait of an object can be characterized by a fluctuation error in measuring the coordinates of the radius vectors pj and Pjy of its BT and their potential resolution.

In the presented research materials, with some restrictions on the location of BT reflections of the surveyed limited area, the above (2) potentially achievable resolutions of the resolution of the tomography method under consideration were obtained. The obtained potentially achievable resolution of the considered signal model when using the synthesized aperture method needs to be adjusted, due to the presence of noise measurements of Doppler frequencies about when forming coordinate estimates.

The operation of the remote detection system for small-sized objects will be described using the specific conditions of its implementation as an example of the remote detection of small-sized objects using an RLD like a helicopter coherent-pulse radar complex 25.

Before conducting radar sounding of a selected area (presumably no more than (5-10) km), on which the presence of mines is expected, the generated radar image (RLI) of the underlying surface is linked to real landmarks (frames) that frame this field. There should be at least 3 such landmarks (see Fig. C) on the ground, if they are lacking, they are created artificially, while such landmarks can be radio beacons, radios or reflectors, or, finally, for example, numbered cubes or shields with distinguishable from the aircraft numbers. Regarding these landmarks, at the end of the work, they present a radar map of the location of the desired small-sized objects on

24

underlying surface, including their subsurface location.

On one of the landmarks of the selected area of the terrain, a satellite radio navigation complex (SRNK) of the SRNK-21DV type is installed (see Fig. 3) to ensure the determination of the relative coordinates between it and the aircraft, for example, a helicopter conducting radar sounding of the field and having a similar radio navigation complex (block 9 coordinates). Other landmarks framing the field are tied to a landmark with a satellite radio navigation complex, for example, with geodetic means using theodolites and portable rods. In this case, the guidelines are set around the perimeter of the field being examined, at a distance from each other equal to (0.3-0.9) the width of the sensing strip during the passage. Radio beacons are installed on selected landmarks with a geodetic reference so that when controlling the helicopter by their radiation, organize flights over the minefield with its uniform multiple coverage of the radar observation sections from the angles, providing sufficient information to complete the above procedures. At the same time, it is assumed that georeferencing is performed in an operatively formed earth coordinate system (GMS) with respect to the location of the landmark with the radio navigation complex and the discrete measurement of coordinates along the {X, Y, Z} GMS axes is 1 m, that is, they actually set the resolution of the points in ZSC, corresponding to the detection of small objects, including mines, taking into account the uncertainty of their position at the sites of resolution of the underlying surface with a size of 1m x 1m.

In addition, immediately before conducting work on probing the minefield, they write off the systematic error of determining the mutual coordinates of the helicopter

J //

25

and landmark with sRNA. This procedure is performed by “hovering the helicopter over a landmark with SRNK and visually fixing it from the helicopter’s side using, for example, a video camera (not shown in FIG.) With a viewing angle of up to 20 ° relative to the vertical axis of the helicopter and with a resolution providing fixation with one pixel of the site 1.5 x 1.5 cm in size on the underlying surface from a height of 60 m, for example, a TRV830E type camera placed on a gyro platform (not shown in Fig.) near the center of mass of the helicopter. When using 2-3 pixels to distinguish a landmark with a size on the underlying surface, such a camera with an accuracy of several cm allows you to determine the position of the landmark relative to the projection of the center of mass of the helicopter on the ground. Such an adjustment of the mutual coordinates of two objects with SRNK with the elimination of a systematic error actually allows us to regularly determine the position of the projection point of the center of mass of the helicopter on the underlying surface relative to the landmark with SRNK located on it for a long period of time sufficient to conduct mine sensing fields.

To determine the coordinates of the detected RLD of small-sized objects, it is necessary to measure with an accuracy of about 1 m the flight height of the helicopter above the probed underlying surface. In the case under consideration, these measurements are carried out by radio engineering in the framework of standard equipment and the algorithm for ensuring the operation of the RLD.

The organization of the processing of radar information from observing the terrain from the side in an operatively formed ZSC makes it possible to filter on-board measurement errors and fundamentally obtain the binding of the observed objects to their location on the ground with an accuracy of about 1 dm. This new quality enhances

26Ш) Ч1ШЬ

permitting sounding of sites on the ground when observing their iodine from different angles. In turn, this provides an increase in the signal / background ratio for small-sized objects, which gives a more reliable detection and recognition.

As an example, we consider the process of obtaining a radar image (RLI) of a minefield section from the side of a helicopter realizing a side view in relation to the direction of its flight using the on-board radar sensor 1 (RLD 1) when the helicopter is flying at a constant speed of 90 km / h at altitude ( 50-60) m. Antenna (position of the transceiver modules 13) RLD 1 of the helicopter is inclined to the underlying surface, providing in a vertical plane at an angle of radar sensing (20-60) ° overview of the terrain from 30 lg to 90 liters relative to the projection onto the underlying surface of the flight path of the helicopter (Fig.Z).

The sector of the beam deflection of the RLD 1 antenna in the horizontal plane is about ± 60 °, which determines the coverage by this sector of the length of the section of the surveyed terrain strip 175 m along the helicopter path line. Such a length corresponds to a segment of the helicopter path, during which for 7 seconds each point of a portion of the surveyed strip is in the sector of the beam deflection of the RLD 1 antenna.

Using the task of coherent radiation of the RLD 1, a synthesized aperture of the antenna mode is organized, which allows one to obtain, with a high resolution, the radar image of the probed area terrain during observation 1 s. The time of observation of the terrain using RLD for 1 s was selected based on the practice of SAR formation using coherent signal processing 11. The entire sector of the RLD viewing area of 1 ± 60 ° is divided into 8 partial sectors with a width of 15 ° under the assumption that it is possible to simultaneously receive information in all 8 sectors

within 1,024 s. This can be achieved when the AFAR RLD 1 beam, being in the same sector for 1 ms for the formation of 128 isodals in distance (in the radial direction from the radar) with a 0.5 m increment, is sequentially moved at the same time intervals to other sectors and after 7 ms, they are again returned to their original position. As a result of this, with the corresponding receipt of reflected signals, they have data for the formation of 8 radio holograms in the form of 128 x 128 square matrices for determining the reflection intensity of 8x2 points of the underlying surface. According to the general approach for reconstructing radar data from a radio hologram 26, they are processed using a two-dimensional fast Fourier transform (FFT) procedure. The obtained values of the detection intensity in the polar coordinate system approximate 10,500 nodes of the Earth’s Cartesian coordinate system, separated by cells 1mx1m the area under review) terrain size 175m hbOm.

In the next 1s time interval, the considered procedure for obtaining information for generating 8 radio holograms is repeated on a new segment of the helicopter flight path corresponding to a certain averaged observation angle. This aspect of observation of the selected area for each / -th segment of the helicopter flight path is hereinafter referred to as D (r 1, 2, ..., 7). In parallel with receiving RLD 1 data for the i-th observation angle using block 3, information is processed for the (r -1) th angle value.

The minefield section in the field of view of the RLD antenna with a width of 60 l and a length of 175 l along the helicopter path, which it overcomes in 7 s, is divided into 7 adjacent parts of strips 60 mx 25 m in size, corresponding to the 1st cycle of formation of a “terrain image frame at coherent processing of information. FIG. 4 illustrates the nature of the intersection

1

.

28

Shchk UP of these strips with the fan arrangement of the RID 1 rays of the helicopter, in particular, on the segment of its trajectory from time t to time tj4, counted from the beginning of the data accumulation time for incoherent information processing. When the helicopter moves, each 1 s receives data corresponding to those reflected from sections of the underlying surface, the numbers of which relative to the current position of the helicopter are shown in FIG. 4, signals for conducting in block 3 the formation of a radar image of their coherent processing of 49 "frames. If there is data for such a matrix, every 7 seconds it becomes possible to organize incoherent information processing for one of the strips for 7 substantially different ranges of angles of its radar exposure: + (60 ° 51 °), ± (5Г-40 °), ± (40 ° - 22 °) and ± 22 °. In FIG. 4, this case is illustrated by a matrix of “frames in size (7x7), the main diagonal of which corresponds to a strip of terrain with 7” frames for incoherent processing of information. At subsequent (1 s) time intervals (elementary intervals), the considered procedure for obtaining information for generating 8 radio holograms from each new segment of the helicopter flight path corresponding to a certain average observation angle is repeated. As a result of processing the information of 8 channels in 1 s, a synthesized antenna radiation pattern (SDN) is formed corresponding to the sector of the angle ± 60 ° of the side view of the terrain when moving the real RLD 1 antenna along the helicopter path. In addition, the length of each control section is chosen taking into account the possibility of covering such a section with the deflection sector of the radar antenna .LA, the step of the discrete rotation step of the radar beam of the radar antenna is selected accordingly

the width of this beam in the horizontal plane, and the predetermined angle of rotation of the beam is chosen corresponding to the angle of coverage of the site sector deflection of the beam of the radar antenna of the aircraft.

In this case, as signs defining either areas with the possible location of small objects, or the location of objects on them, or objects located in these places, choose the characteristics of the intensity and / or length of the discrete signals of the corresponding radio image.

Carrying out additional incoherent processing of information on 7 "frames allows several times to increase the linear resolution on the ground. If, due to the smallness of the signal / background ratios obtained in some areas, the processing results may not be sufficient to make a decision on the detection, in particular, mines, an additional examination of “suspicious areas. A natural way to further reduce the linear resolution on the ground is to re-probe these areas with a significant increase in the range of viewing angles, for example, by circular (also called telescopic) flight of these areas.

The telescopic observation mode with an RLD probe beam width of 5-15 ° allows for an inclined range of 85 m to cover a terrain area with a diameter of about 25 m and to ensure its linear resolution in the central part up to half the radio wave length of the RLD 1 probe signal.

If there is insufficient one-time inspection of the minefield and frequent calls to the telescopic viewing mode, another option is possible for organizing a search with multiple observation of sections of the minefield from different angles. In this version of the mine search organization, the helicopter repeatedly crosses this field, not

Zlt {1 & 1

thirty

changing course and making turns outside the field to cross it under a different directional direction of flight.

To maintain the helicopter heading during manual piloting, optically noticeable buoys can be previously dropped onto the surface of the minefield, which serve as the pilot's guides during helicopter flight. These buoys when conducting radar sounding of the minefield are also fixed with a video camera (not shown in Fig.), Located, for example, under the bottom of the helicopter cabin. The video allows you to organize the binding of the radar image generated on board the helicopter to landmarks on the terrain, including to the buoys in the minefield, regarding which they will subsequently carry out the operation related to mine clearance.

A three-time flight of a helicopter with a constant course over small-sized terrain objects at course angles of O, 120 ° and 240 ° is almost equivalent to a circular flight over the entire range of angles of 360 ° of a telescopic view of a selected area of terrain. In this case, subject to the above flight parameters, a regular moving 7 second cycle of processing cartographic information covers 1 hectare of terrain, which is almost 20 times larger than the area under observation in telescopic view mode. However, with such a review, it is necessary to accumulate a huge amount of information of individual “footage of the area with an area of 0.15 ha. With 3-fold coverage of a minefield with an area of 1 km, such processing will require the memory capacity of memory devices for recording radar data of almost 15,000 "frames of the terrain.

Processing is carried out in the command center (KP) computer center (not numbered in Fig. 3), where it is automatically done using radio transmitting and receiving devices

Itcht

31

 Ae w / e

(devices 5 and 10, respectively), relevant information is received.

An additional increase in linear resolution in the mode of telescopic viewing of a terrain is possible by increasing the flight speed of the helicopter so as to cover the entire range of viewing angles of small objects on a regular rolling 7 with a cycle of processing cartographic information. In this case, for an inclined range of 85 m to the center of the observed area at a helicopter flight speed of 90 km / h, the angle range does not exceed 120 °.

When organizing a search, a sequential scan of sections of a given area (the proposed minefield) is performed. The detection and recognition of small objects is carried out in accordance with the algorithm implemented in the computing device 62. Object detection

Each "frame of radar data obtained as a result of sounding and recorded in the memory unit of PSU 4i, and then transferred to PSU 42, is a matrix in which the position of the elements corresponds to the positions of elementary sites in the geodetic grid. The size of such sites is determined by the resolution of the system that it possesses at the stage of detection (for example, 1x1m). In a visual image, such a discrete area will correspond to one pixel, the brightness of which is determined by the sum of the radar signals from the earth's surface, re-reflections from subsurface layers and from an object located on or below the surface. It is quite obvious that, depending on the terrain, the slopes of the elementary sites will change, but in most cases these changes are characteristic not of one elementary site, but of a certain group of them. For such a group can be calculated

average power level of received signals Pav. Next, the PC threshold is set, which provides a given probability of false alarms, which must be kept constant for all subsequent groups of sites. The fact of object detection will be considered established if the threshold is exceeded by a signal with a power Pcij in the cell of the matrix Sy. In this case, the probability of detecting Robn depends on the ratio Ksf Ru / Psr. When observing the site in one flight above it at a certain heading angle, the CSF can be (3–9) dB 8 depending on the length of the emitted wave. At Ksf 9dB, the probability of detecting an object against a background of reflections from the underlying surface, according to 10, will be Robn (0.2-0.3). Repeated observations of this site at the same course angle of flight of the aircraft will increase the likelihood of detection. As a further consideration, we took a section from which the signal barely exceeded the threshold level and had a Ksf of 9 dB (section 4 of Fig. 6) with a resolving power of radar with a synthesized aperture equal to 1x1 m.

With an increase in the number of repeated flights over section No. 4 using the synthesized aperture method, Robn will increase (Fig. 7). The first case corresponds to N flights (1,2,3,4) over the .f 4 section, and the second to the same number of flights at different heading angles M over the same section. With the same initial correlation Cof 9 dB d.11ya of both options, with each subsequent flight over section No. 4, the resolution in the proposed system increases by 2 times in each coordinate, as a result of which the observed area decreases by 4 times (6 dB), by the same Ksf increases. This leads to a rapid increase in the probability of detection to the level of (0.95-0.98) when viewing the same area repeatedly. At the same time, Fig. 7 shows that when using existing systems with synthesized aperture, the increase in the number of spans is not

33 "WfOfS

leads to an increase in resolution, and the growth of Robn is explained only by the result of processing the accumulated information.

The object detection algorithm can be represented as follows:

The first pass of groups of sites with a resolution of 1x1m:

Defining a threshold for site groups:

Acpij a; Aij) / N; nij-l, lAepij

Acpki & Aki) / N; Py 1.1 Acpki Aeprs (gSDGZ) / N; 1.1 Aeprs where N 5 Formation of the threshold matrix, M T

34IMOM GA is significant To VAij nij- ij i || VAij the second passage of the groups of platforms with a resolution of 0.5x0.5 V;: ij 1 & AC Pts ij y + 1 retiy passage of the groups of platforms with a resolution of 0.5x0 , 5 V /: y 1 Thus, after processing the primary information, we obtain a matrix, M, the elements of which Ku will have values 0; 1; 2; 3, which determines the corresponding degree of reliability of detection of some object. After completing the detection procedure, performed with varying degrees of reliability in different areas, the automatic selection of small objects. If, when analyzing the contents of the matrix, M, a compact group of values 1 is found, then this will mean the presence of which element

extended object. In the opposite case, i.e. when disjoint values other than zero are present in the matrix, these places will correspond to the location of small objects. Moreover, the value of "1 means the lowest reliability of detection of a small-sized object, and this place on the earth's surface must be checked additionally because there could be a false alarm.

Small Object Selection Algorithm

  & (i + u, i + t2 0 V (ii, i2 € -1; 0; 1 & 1 (q i2 0)

- small object

Kij 0 & (3 (I,, 12 е -1; 0; 1 & 1 (ij- 12- 0) К ,,,; + а 0 large-sized object

After the selection of small objects above one of them having high reliability, a flyover is performed in the telescopic observation mode, in which they provide a system resolution of the order of 0.15 mx 0.15 m.

In this mode, super-resolution radios are transmitted to a neural network structure, which, by comparing with the standard it is trained in, should recognize the detected object. A small object can be, for example, a fragment of a shell, a piece of pipe or a mine of a certain type. Classifier 11 at its output should give an answer to the belonging of the detected object to one or another class of small-sized objects.

Object recognition is based on the use of neural network structures (NSS) 34. The dog consists of a Kohonen circuit and a multilayer perceptron. At the first stage of the MSS operation, the Kohonen scheme works, which is used as a detector, and at the second (recognition), a multilayer perceptron (MP).

35

structure of a natural neuron. The basis of the IN is an adder, a multiplier and a normal circuit (not shown in FIG.). The input signals arriving at the IN through the multipliers are superimposed in a certain way defined by weighting coefficients, and the total signal is fed to a threshold circuit in which a nonlinear activation function is realized. The form of this function can take the form of a sigmoid, a single jump, a quadratic, or other dependence. The output of IN can be values of O or 1 for the function of a single jump, or intermediate values between O and 1 (or -1 and 1) for other functions. As a result, the neuron will be considered excited (positive values or 1 at its output), or inhibited (negative values or O at the output). In multilayer HCC schemes, one or more intermediate layers may exist between the input and output layers, connected in a certain way with previous and subsequent layers. At the outputs of the neurons of the last layer, the result of the work of the entire neural network will be formed.

A common property for all types of NSS is that they have two operating modes - training and reproduction. Therefore, if the NSS is a learning structure, then it is not pre-programmed, but learning without programming, the NSS is an adaptive or self-organizing computer.

The operation of the Kohonen scheme is based on the principle of competition and is best suited for identifying heterogeneities that are most often found in the image during the initial processing, which is an important property for using this scheme as a detector for identifying characteristic inhomogeneities. When scanning the image with a sliding window, clusters are formed in this NSS, which correspond to the appearance of signals from the “victorious neurons”. These clusters strengthen or weaken as the window moves along the lines. After scanning everything

"JjsjG m (

image in the middle layer of the NSS clusters of heterogeneities will be formed, most often found in the image. After that, the contents of the clusters are not redirected to a multilayer perceptron, which is preliminarily trained to recognize some figures (rectangle, ellipse, square, etc.) of different orientations and their combinations. After processing the entire image, the filtered information is obtained at the MP output, in which there are no images unrelated to the reference images and their combinations.

The Kohonen (SC) scheme used in the created NSS is designed as a two-layer neural network. The input layer is actually a buffer between the input signals and the inner layer of neurons, therefore it is not considered a layer of neurons. Neurons of all layers have interconnections with certain weighting factors, for example, m, Wy. Coefficients Шц initially assigned randomly, and then at each step of the calculations set

mij mjj + С (Ei - mij) Y,, where Ej are the normalized values of the input signals YJ is the output value of the J-neuron

YJ f WjkSmkiEi - (Z ImjiEi - G),

where G is the threshold value.

In the process of functioning (self-organization) at the output of the Kohonen network, clusters are formed - groups of active neurons that characterize certain categories of input vectors corresponding to one spatial situation.

When the image obtained from the Kohonen network is similar to the corresponding reference image, information about one of the image classes is activated at the output of the perceptron, for example, either “mine, or“ shell fragment, or “piece of pipe, or“ unidentified object.

(

37 and h and g

38ДШ 1№Ш The algorithm of operation of block 3 for generating a radar image is described in the description of obtaining radio holograms and the Radon transform. Input information is received via digital communication lines (not shown in Fig.): S BSogl 2 in the form of a sequence of pulses (16 bits) with frequency 300 MHz, s BSync 7 via a synchronization channel with a pulse transmission frequency of 1 Hz; from WU 6i through the channel for transmitting trajectory data on the movement of the helicopter in the form of 16 bit words with a frequency of 1 kHz. The output of the algorithm is an array of data that is transmitted to the memory unit 4i for online storage for subsequent transmission to processing. The BFRLI algorithm includes the following sequence of actions: 1) Conducting a digital phase detection procedure for intermediate frequency signal information; 2) Formation for each burst of 8 reflected pulses with a duration of, for example, 0.2 μs and a repetition period of 1 μs of complex numbers, taking into account the correcting phase factor of signal delay compensation d.11a of the average value of the carrier frequency of the burst pulses and the overall frequency modulation index of the LFM, for example, ju 300 MHz / s; 3) Recording a sequence of complex numbers in the form of an array of the nth row of each sequentially formed nth beam field of view (, 2, ..., 8) when it moves with a time discrete of 1024 μs in a side view sector of ± 60 ° RLD relative to direct helicopter flight;

4) Formation of an elementary inspection interval of 1.048 s duration for 8 viewing areas of the beam of the matrix of source data of size (128x128);

5) Determination of the coordinates of the center of each w-th beam viewing area and the value of the slant range depending on the height of the helicopter;

6) Evaluation of the radial and tangential components of the speed of the helicopter relative to the centers of each t-th area of the beam survey in the middle of the elementary inspection interval;

7) Determination of the Doppler shift caused by the speed of the helicopter along the radial component for each of the 128 sounding frequencies;

8) Correction of the values of the complex elements of 8 matrices in size (128x128) in order to exclude the influence of the radial approximation of the helicopter;

9) Definition of the sector of the viewing angles of each center / ith beam viewing area;

10) Formation of 8 matrices in size (128x128), corresponding to discrete two-coordinate radio holograms of rotating objects;

11) Performing a two-dimensional FFT of 8 indicated matrices;

12) Obtaining discrete images of reconstructed radar images of the underlying surface areas corresponding to radio holograms;

13) Presentation of 128x128 values of the elements of 8 matrixes for viewing the X-rays of the XRDs in the polar coordinate system into the Cartesian coordinate system of the underlying surface, covered by an “observation frame 175m x 60m, and interpolation spacing this data“ frame into 7 matrices corresponding to 60m x25m sites;

14) Entering into memory of data for each of the sites recorded on the current elementary inspection interval;

Mttfic i

39

15) Formation of similar data for 7 non-sequential “frames in the form of a cell matrix (7x7) of the values of the points of the sites;

16) Isolation of the main diagonal of the formed cell matrix;

17) Carrying out the procedure for compiling projections p (u, j3) of the reconstructed image of the object, observed from 7 angles.

18) Conducting, in accordance with formula (7), the operation of convolution of the projections p (u, p) with the previously calculated core value hc (u), and the filtered value Ph (u, l3) is determined in the frequency domain using the Fourier transform of the projection p (u, p) and the kernels hc (u) and the subsequent inverse Fourier transform of the product of their images.

19) Restore, according to formula (8), the approximation g (x, y) to

of the original function gi (x, y) by the filtered values ph (u, P) at each point by summing the data values along each line xcosft + ysinp and passing through this point 31.

20) Recalculation using the 4th order interpolation splines of the data values gi (x, y) obtained for each straight line xcosfl + ysinj3 and the Cartesian coordinate system with a mesh of nodes 1m X 1m, into a network of nodes 0.1m x 0.1m 32 .

21) Determination of excesses over the adaptive threshold of the accumulated values of reflection intensities from the underlying surface for each section measuring 60m x 25m.

22) Conducting cluster analysis to determine the assessment of the location of small objects, including mines.

23) Regular repetition of the above computational procedures at each elementary inspection interval;

24) Transfer of current information to BP 4 to perform a more detailed analysis of the data obtained during the flight or after the flight to the CP.

40MOT (fi

The amount of computational costs by analogy with the calculation of the computational load of a digital computer when solving such problems 14 to ensure the operation of all BFRLI algorithms in real time will be about 45 million per second. In general, the implementation of all computational procedures for the formation of radar images of small-sized objects on the ground in real-time information flow will require about 50 million. op./s, which can be implemented, for example, by reprogrammable signal processors such as Baguette 14

Block WU 6 performs the following tasks:

The launch of the synchronization unit 7, which forms a grid of clock pulses to synchronize the operation of all blocks of the system;

Selecting, upon request, the operator from the memory unit 4i, in which the matrix Sy, of local sections of the matrix 8c, is actually filled and stored to visually evaluate the possible presence of mine-like objects in them;

Transmission of information about radar data for current terrain to a radio transmitting device 5 for its transmission to a ground processing center (L) and to a cabin indicator

Issue in RLD 1 information on controlling the current position of the antenna beam and the values of the parameters of the emitted signal for each operating mode of the system;

Management of the block 9 determine the coordinates.

The fundamental point is the implementation of the last function, the implementation of which determines the accuracy of the entire system. Indeed, the accuracy of determining the coordinates of detected objects and, ultimately, the efficiency of the system’s functioning depend on the accuracy of linking the positions of the aircraft and the AFAR beam to the terrain map. Therefore, we will gradually consider the preparatory operations and the operation of block 6i.

.vzh 1sh {

41

Stage 1.

Before the start of the minefield flight, the coordinates of the aircraft (in the dependent mode or when moving at low speed) and the beacon (RM) with an accuracy of (3-5) cm are determined by GPS satellite signals. At the same time, the camera is calibrated with the same accuracy by the reflector located on the RM (not shown in FIG.). The resulting point video image from this reflector will be taken as the origin in the visual image of the area. In this case, the GPS GPS coordinates obtained from the communication channel through the coordinate determination unit 9 are ascribed to the visual coordinate origin. Further, all objects falling into the field of view of the camera will have coordinates relative to the position of the RM. Relative to the starting point, the current position of the aircraft is highlighted, the coordinates of which are determined through the GPS system and, therefore, also tied to the visual origin.

Stage 2.

When systematically viewing the terrain using RANGE 1, the systems for detecting small-sized objects on the E1fan of indicator 8i are displayed and subsequently retained along the lengths of the aircraft increasing in length during the flight, corresponding to the current and viewed “frames of the earth’s surface. These bands also correspond to the intersection of the radiation pattern of the main lobe of the antenna beam in its current positions by the plane of the earth’s surface at a given flight height of the aircraft. In this case, the coordinates of the indicated intersection sections are calculated in WU 6i and presented on the screen of the indicator 8i. Stage 3.

A visual map of the area to improve the observability of radar reflections is eliminated by disconnecting the camcorder from the indicator. Her connection will occur only in

42

43

UR1 (

moments of observation of bright points from optical reflectors installed along the contour of the field to adjust the course of the aircraft. Thus, the indicator displays: stripes corresponding to the viewed sections of the earth's surface; memorized and current radar data; positions of reference points from optical reflectors. After the flight is completed, those areas that were missed during a radar inspection of the entire field or were viewed only along one course route (faded areas) will be visible on the screen.

Stage 4.

When radar inspection of the terrain on the screen of the indicator 82 display radar data in the form of a set of points corresponding to the detected objects in the "frames of the scanned sections of the earth's surface. Moreover, in the search mode, each point will correspond to an elementary resolution area of 1x1 m in size. The brightness of the point, in turn, will depend on the degree of reliability of the detection of the object in this place. In WU 6i, the coordinates of the detected objects are recalculated from the polar system, in which the antenna of the airborne radar device operates, into a Cartesian system, tied to a terrain map.

Block 62 as well as on-board unit 6b is a computing and control device whose main function is to control the flow of information coming from the on-board part of the system. All radar images in “frames of the earth’s surface and video images enter the memory unit of the PSU 42. The memory in the PSU 42 is divided into two areas. In one of them, incoming video images and PJOM are saved, and in the other part, radar images that have forgiven processing in the classifier I and transmitted from VU 62. The control functions of VU 62 are manifested in the transmission of commands for issuing any information from the PSU requested by the operator and its output on the indicator 82 or on the printing device PU 12.

Unit VU 62 operates in 3 main modes: Mode 1. Receiving information from unit 11 and transmitting it to indicator 82.

“Frames with radar images but the command from VU 62 are transferred to the I classifier, in which they perform all the necessary procedures for processing radar data: object detection, selection and recognition of small-sized objects. Each "frame has its own number, according to which VU 62 makes its" binding to a map of the area, i.e. to the video image received from the video camera stored in the PSU 42 and displayed on the indicator 82. Upon receipt from the classifier of the next RLI, VU 62 carries out its coordinate reference.

Mode 2. Transmission of the processed radar data to the PSU 42 Mode 3. Control of the transmission of information at the request of the operator: output of the search results to the indicator or / and to the printing device PU 12.

List of used information sources:

1.Kuchko A.S. Aerial photography and special photographic studies. M .: Pedra, 1988, S.S. 51-53.

2. Welzer V. Aerial photographs in military affairs. M .: Military publishing house, 1990, S.S. 71-80).

3. RF patent .№ 2211461 with priority dated 06/18/2001, G 01 S 13/90.

4.Sablin V.P. New technologies in humanitarian mine clearance. Proceedings of the jubilee scientific and technical conference dedicated to the 30th anniversary of the founding of 1SCHIIRES. Ed. JSC TsPIIIRES, 2001, part 1, s.p. 20-23.

5. Naval Research Laboratory USA, available at www. Science news by Miguel Vicens. com.

Ims loj

45

6.Sablin V.N. New technologies in humanitarian mine clearance, Proceedings of the jubilee scientific and technical conference dedicated to the 30th anniversary of the foundation of the Central Research Institute of Radioelectric Power Engineering. Ed. JSC SCHZHIRES, 2001, part 1, s.p. 16-18.

7. Subsurface radar. Ed. M.I. Finkelypteyna. M.: Radio and communication. 1994.

8.Check F. Lee and outhers. RCS Models and Analysis of Mines Proc. SPIE, Vol. 3079, p.p. 581-597.

9. Finkelypteyn M.I. Basics of radar. M .: Radio and communication. M., 1983 pp. 46.56.

10. Radio. Tutorial. Ed. G.S. Kondratenkova. M .: EDITION VVIA, 1998, p. 26.112.127.

I. Radar stations with synthesized antenna aperture. Ed. V.T. Goryainova. M .: Radio and communication. 1988. S.S. 19; 90.

12.Halkason S. Radon conversion. M .: Mir, M., 1983.

13. Manukyan A.A. “Determining the coordinates of local inhomogeneities on the surface of an object from the amplitude-phase backscattering diagram. // Radio engineering and electronics. Journal of the RAS. M.: Science, Volume XXX. Vol. 2, 1985.

14. Description to the patent of the Russian Federation. No. 22114161 with a priority of 06/18/2001, IPC: G 01 S 13/90.

15. Reference "Advanced automation technology. Brief product catalog. M .: compiler of the directory and supplier of products, ProSoft company, Web address - http: // www. prosoft.ru; 1999, s.p. 5.24.

16. Professional television, video and audio equipment. Catalog of the company “Window-TV. M .: 2001-2002, S.S. 8, 88.

17.Titze W. Schenk K. Semiconductor circuitry. M .: Mir, 1982, S.S. 125-127,369,284-286.

18. The satellite radio navigation system SRNK-21DV Prospect MAKS-2003. Zhukovsky city: GUL “Flight Research Center LII named after M.M. Gromova, 2003.

19.Shevchenko P.A. et al. Neuroprocessor architecture NM 6403. Collection of reports of the 5th All-Russian Conference “Neurocomputers and their Application. M., 1999, S.S. 70-90).

20. Designing phased array antennas. Ed. DI. Vosfesensky. M .: Radio engineering, 2003.

21.Levitin V.I. et al. Status and development prospects of transceiver modules and lattice fragments for the PAR. Proceedings of the scientific and technical anniversary conference TsNIIRES. M .: TsNIIRES, part 1. 2001, p. 36.

22. Prospect MAKS-2003. Zhukovsky Ed. OJSC Research Institute of Instrument Engineering and the State Ryazan Instrument Plant, 2003.

23. Aerial reconnaissance radars. Ed. G.S. Kondratenkova. M .: Military Publishing. 1983, p. 120.

24. Description of the patent of the Russian Federation No. 2099743 with a priority of 12/07/95, G 01 S 13/89.

25. Kanaschenkov A., Ratner V., Kurilkin V. Helicopter radars of the design “Fazotron-NIIR, // Journal“ Herald of Aviation and Cosmonautics, 2003, No. 4, p. 26).

26.Safronov G.S., Safronova A.P. Introduction to radio holography. M .: Soviet Radio., 1973, p. 187.

27. Herman G. Projection image restoration. Fundamentals of reconstructive tomography. M .: World. 1983.

28. Voronin E.N., Nechaev E.E., Shashenkov V.F. Reconstructive antenna measurements. M.: Science, 1995.

29.Townsend D.W. and Defnse M. Image Reconstruction Methods inp Positron Tomography. - Lectures given in the Academic Training Program of CERN, Geneva, 1993.

46/1 i / h. I 1 "

Ministry of Agriculture / 1b / W

30.Besarab M.A.,. Kravchenko V.F. Functions V.L. Rvacheva in problems of computational tomography. // Foreign radio electronics. .№3, 2002.

31.Dadzhion D., Mercero R. Digital processing of multidimensional signals. M .: Mir, 1988.

32.A.N. Tikhonov, V.Ya., Arsenin, A.A. Timonov Mathematical problems of computed tomography. M .: Nauka, 1987.

33. Guide to radar. Ed. M. Skolnik, Moscow: 1976 1, p. 204.

34.Gribkov V.F. et al. An empirical algorithm for processing radar information. Collection of reports at the 5th All-Russian Conference “Neurocomputers and their Application, M .. 1999

Claims (2)

1. A system for remote detection of small objects, including a radar sensor connected to a computing device connected to an indicator, characterized in that a matching unit, a radar image forming unit, a classifier, two memory units, a coordinate determination unit, a synchronization unit are introduced into it, a printing and radio transmitting and receiving device, as well as an additional indicator and a computing device, the radar sensor being connected to the input of the matching unit, connected with the first input of the radar image forming unit, connected to the first input of the first memory block, connected to the first input of the radio transmitting device, the second (control) input connected to the first output of the computing device, the second output connected to the second ( control) the input of the first memory block, the group of outputs connected to the first group of inputs of the radar sensor, and the third output connected to the input of the block synchronization, with its first and second outputs connected to the second (synchronizing) inputs of the matching unit and the radar imaging unit, respectively, and a group of outputs connected to the second group of inputs of the radar sensor, the second output connected to the third input of the matching unit, while the computing device is input-output connected to the input-output of the first memory block, and the fifth and sixth outputs are connected respectively to the first inputs of the indicator and the block definition dynamite, with its first and second outputs respectively connected to the first input of the computing device and the second input of the indicator, and the radio input associated with the radio input of the system, and the transmitting device with its radio output is connected to the radio input of the radio receiving device, the output connected to the first input of the second memory block, its output connected to the first input of the classifier, the output connected to the first input of the additional computing device, its first, second and third outputs sub assigned to the second inputs of the radio receiver, the second memory unit and the classifier, the fourth and fifth outputs connected to the first inputs of the second indicator and the printing device, respectively, and the input-output connected to the input-output of the second memory unit, while the inputs of the computing and additional computing devices connected respectively to the first and second inputs of the system, the output of the printing device is connected to the output of the system, the optical outputs of the indicators are connected to optical optical outputs of the system, and the group of radio inputs and outputs of the radar sensor is connected to the corresponding group of inputs and outputs of the system. 2. The system according to claim 1, characterized in that the radar sensor contains a phased antenna array consisting of N independently connected transceiver modules, where, for example, N = (1, ..., 48), an adder, two switches, a generator, a modulator and a unit for generating a predetermined distribution of amplitudes and phases of the signals to control the position of the antenna beam of each of the transceiver modules, the transceiver modules being connected by the first electrical inputs to the corresponding outputs of the first switch, the first input connected to the first output of the generator, and the second input connected to the first output of the formation unit of the specified distribution of amplitudes and phases of the signals to control the position of the antenna beam of each of the transceiver modules, the second electrical inputs of which are connected to the corresponding outputs of the second switch, the first input connected to the second output of the formation unit of the specified distribution amplitudes and phases of signals to control the position of the antenna beam of each of the transceiver modules, and the second input connected to the first input of the first group of inputs of the radar sensor, the first output connected to the output of the adder, the corresponding inputs connected to the outputs of the respective transceiver modules, and the second output connected to the output of the generator, the first input connected to the output of the modulator, its first input connected to the third input of the first group of inputs, and the second input connected to the first input of the second group of inputs of the radar sensor, while the second and third inputs of the second group of inputs of the radar The sensors are connected to second inputs respectively and a generator unit for generating a predetermined distribution of amplitudes and phases of the signals for controlling the position of the antenna beam of each of the transceiver modules, which radiovhodami-outs connected with the respective outputs radiovhodami-radar sensor.