CN108872130B - Typical aircraft Facing material recognition methods neural network based - Google Patents
Typical aircraft Facing material recognition methods neural network based Download PDFInfo
- Publication number
- CN108872130B CN108872130B CN201810662092.8A CN201810662092A CN108872130B CN 108872130 B CN108872130 B CN 108872130B CN 201810662092 A CN201810662092 A CN 201810662092A CN 108872130 B CN108872130 B CN 108872130B
- Authority
- CN
- China
- Prior art keywords
- data
- neural network
- neuron
- competition
- spectrum
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 239000000463 material Substances 0.000 title claims abstract description 56
- 238000000034 method Methods 0.000 title claims abstract description 28
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 15
- 238000001228 spectrum Methods 0.000 claims abstract description 28
- 210000002569 neuron Anatomy 0.000 claims abstract description 27
- 238000004566 IR spectroscopy Methods 0.000 claims abstract description 13
- 238000004611 spectroscopical analysis Methods 0.000 claims abstract description 13
- 238000003062 neural network model Methods 0.000 claims abstract description 9
- 238000012549 training Methods 0.000 claims abstract description 6
- 238000001514 detection method Methods 0.000 claims abstract description 5
- 238000013135 deep learning Methods 0.000 claims abstract description 4
- 238000010606 normalization Methods 0.000 claims description 7
- 238000012545 processing Methods 0.000 claims description 6
- 238000002329 infrared spectrum Methods 0.000 claims description 5
- 230000006870 function Effects 0.000 claims description 4
- 238000012360 testing method Methods 0.000 claims description 4
- 241001269238 Data Species 0.000 claims description 3
- 230000008030 elimination Effects 0.000 claims description 3
- 238000003379 elimination reaction Methods 0.000 claims description 3
- 238000002474 experimental method Methods 0.000 claims description 3
- 238000009499 grossing Methods 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims 1
- 238000002360 preparation method Methods 0.000 abstract description 2
- 239000010410 layer Substances 0.000 description 18
- 239000010959 steel Substances 0.000 description 6
- 229910000831 Steel Inorganic materials 0.000 description 5
- RTAQQCXQSZGOHL-UHFFFAOYSA-N Titanium Chemical compound [Ti] RTAQQCXQSZGOHL-UHFFFAOYSA-N 0.000 description 5
- XAGFODPZIPBFFR-UHFFFAOYSA-N aluminium Chemical compound [Al] XAGFODPZIPBFFR-UHFFFAOYSA-N 0.000 description 5
- 229910052782 aluminium Inorganic materials 0.000 description 5
- 239000010936 titanium Substances 0.000 description 5
- 229910052719 titanium Inorganic materials 0.000 description 5
- 239000004411 aluminium Substances 0.000 description 4
- 229910000737 Duralumin Inorganic materials 0.000 description 3
- 239000002131 composite material Substances 0.000 description 3
- 239000011152 fibreglass Substances 0.000 description 3
- 229910052751 metal Inorganic materials 0.000 description 3
- 239000002184 metal Substances 0.000 description 3
- 229910000838 Al alloy Inorganic materials 0.000 description 2
- 229910000851 Alloy steel Inorganic materials 0.000 description 2
- ZOXJGFHDIHLPTG-UHFFFAOYSA-N Boron Chemical compound [B] ZOXJGFHDIHLPTG-UHFFFAOYSA-N 0.000 description 2
- 229920000049 Carbon (fiber) Polymers 0.000 description 2
- 229910000746 Structural steel Inorganic materials 0.000 description 2
- 229910052796 boron Inorganic materials 0.000 description 2
- 239000004917 carbon fiber Substances 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 239000004744 fabric Substances 0.000 description 2
- 239000000835 fiber Substances 0.000 description 2
- 239000011521 glass Substances 0.000 description 2
- VNWKTOKETHGBQD-UHFFFAOYSA-N methane Chemical compound C VNWKTOKETHGBQD-UHFFFAOYSA-N 0.000 description 2
- 229920003023 plastic Polymers 0.000 description 2
- 239000010935 stainless steel Substances 0.000 description 2
- 229910001220 stainless steel Inorganic materials 0.000 description 2
- NIXOWILDQLNWCW-UHFFFAOYSA-M Acrylate Chemical compound [O-]C(=O)C=C NIXOWILDQLNWCW-UHFFFAOYSA-M 0.000 description 1
- 229910000881 Cu alloy Inorganic materials 0.000 description 1
- 239000004593 Epoxy Substances 0.000 description 1
- 101150064138 MAP1 gene Proteins 0.000 description 1
- 229910000861 Mg alloy Inorganic materials 0.000 description 1
- 240000007594 Oryza sativa Species 0.000 description 1
- 235000007164 Oryza sativa Nutrition 0.000 description 1
- 229910001069 Ti alloy Inorganic materials 0.000 description 1
- 238000004378 air conditioning Methods 0.000 description 1
- 229910045601 alloy Inorganic materials 0.000 description 1
- 239000000956 alloy Substances 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 239000011248 coating agent Substances 0.000 description 1
- 238000000576 coating method Methods 0.000 description 1
- 230000002860 competitive effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 239000004035 construction material Substances 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- PCHJSUWPFVWCPO-UHFFFAOYSA-N gold Chemical compound [Au] PCHJSUWPFVWCPO-UHFFFAOYSA-N 0.000 description 1
- 239000010931 gold Substances 0.000 description 1
- 229910052737 gold Inorganic materials 0.000 description 1
- 239000011229 interlayer Substances 0.000 description 1
- 229910021421 monocrystalline silicon Inorganic materials 0.000 description 1
- 210000005036 nerve Anatomy 0.000 description 1
- 230000007935 neutral effect Effects 0.000 description 1
- 239000004033 plastic Substances 0.000 description 1
- 230000008092 positive effect Effects 0.000 description 1
- 230000003014 reinforcing effect Effects 0.000 description 1
- 235000009566 rice Nutrition 0.000 description 1
- 239000005060 rubber Substances 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
- 238000009628 steelmaking Methods 0.000 description 1
- 239000004753 textile Substances 0.000 description 1
- 239000012780 transparent material Substances 0.000 description 1
- 239000002023 wood Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
- G01N21/31—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry
- G01N21/35—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light
- G01N21/3563—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light for analysing solids; Preparation of samples therefor
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
- G01N21/31—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry
- G01N21/35—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light
- G01N21/359—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light using near infrared light
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N2201/00—Features of devices classified in G01N21/00
- G01N2201/12—Circuits of general importance; Signal processing
- G01N2201/129—Using chemometrical methods
- G01N2201/1296—Using chemometrical methods using neural networks
Landscapes
- Physics & Mathematics (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The present invention relates to a kind of typical aircraft Facing material recognition methods neural network based, include the following steps: step 1: acquiring a large amount of diffusing reflection infrared spectroscopies of a variety of typical aircraft surfacings；Step 2: the spectrum of acquisition is pre-processed；Step 3: material being detected according to traditional material detection method, and corresponding material infrared spectroscopy is demarcated, establishes ir data and material matching model；Step 4: classifying to database infrared spectroscopy；Step 5: spectroscopic data being formatted, deep learning and training are carried out to database using neural network, establish its neural network model；The neural network is divided into input layer and competition layer, and each neuron of network competition layer obtains the respond opportunity to input pattern by competing, only one last neuron becomes victor；This method without sample preparation can outfield sample is analyzed, the accuracy rate of identification averagely reaches 85% or more.
Description
Technical field
The invention belongs to Material Identification fields, and in particular to a kind of aerospace typical aircraft table neural network based
The recognition methods of plane materiel matter.
Background technique
The range of aircraft material is wider, is divided into organism material (including structural material and non-structural material), engine material
Material and coating, wherein most importantly housing construction material and engine material.Non-structural material includes: transparent material, in cabin
Facility and ornament materials, the attachment and pipeline material of the systems such as hydraulic, air-conditioning, antenna house and electromagnetic material, tire material etc..
Non-structural material amount is few and wide in variety, has: glass, plastics, textile, rubber, aluminium alloy, magnesium alloy, copper alloy and stainless steel
Deng.
Aircraft timber, cloth and the steel making in the manned heaven of first early 20th century.Body knot is given in the appearance of duralumin
Structure brings huge variation.1910~nineteen twenty-five starts to replace timber to make fuselage framework with steel pipe, makees covering with aluminium, manufactures full gold
Belong to the aircraft of structure.Metal structure aircraft improves structural strength, improves aerodynamic configuration, and aeroplane performance is made to be improved.
The forties all-metal structure aircraft speed per hour more than 600 kilometers.The end of the fifties jet plane speed more than 2 overtones
Speed brings thermal boundary problem to aircraft materials.Aluminium alloy high temperature resistance is poor, and at 200 DEG C, intensity has dropped to room temperature value
1/2 or so, it needs to select the better titanium of heat resistance or steel.There is the full titanium high-altitude high-speed reconnaissance plane of SR-71 of 3 times of velocities of sound in the sixties
The XB-70 bomber of housing construction weight 69% is accounted for stainless steel.The 25 fighter plane wing cover of rice lattice of the Soviet Union also uses titanium
And steel.The composite material with boron fibre or fibre reinforced is used after the seventies more and more.Aluminium, titanium, steel and composite wood
Material has become the basic structure material of aircraft.
Wing is the main component of aircraft, and the wing of the dopey of early stage is timber structure, makees covering with cloth.This wing
Structural strength it is low, pneumatic efficiency is poor, already replaced metal wings.Beam inside wing is the main stressed member of wing,
Generally using superduralumin and steel or titanium alloy；The junction portion of spar and fuselage uses high-tensile structural steel.Wing cover is because upper
The stress condition of lower aerofoil is different, and the good superduralumin of compressive property and tension and the good duralumin of fatigue behaviour is respectively adopted.In order to
Mitigate weight, the front and rear edge of wing is frequently with fiberglass reinforced plastics (glass reinforced plastic) or aluminum honeycomb interlayer (core) structure.Empennage
Structural material generally uses superduralumin, and fighter plane selects boron or carbon fiber epoxy composite to mention to mitigate aft weight sometimes
High operational performance；Rudder and elevator on empennage use duralumin.
For aircraft in high-altitude flight, fuselage pressurization cockpit bears interior pressure, needs high, fatigue proof hard using tensile strength
Aluminium makees skin material.Fuselage bulkhead generally uses superduralumin, bears to use high-tensile structural steel or titanium compared with the reinforcing frame of big load
Alloy.The airborne radar of many aircrafts is mounted in fuselage head, general to be covered it using the nose cone that fiberglass reinforced plastics is made into
Firmly so that electromagnetic wave can be penetrated.The canopy and windshield of cockpit use acrylate transparent plastic (organic glass).Before
Undercarriage stress is smaller, generallys use common alloy steel or superduralumin.
The identification of aircraft surface material has great significance to the identification of aircraft, however existing aircraft surfaces material
Matter identification technology needs to pre-process the material mostly, cumbersome, time-consuming and laborious.Thus develop it is a kind of quickly, can
The method of direct field testing is of great significance.
Summary of the invention
In order to overcome the difficulty of the classification of typical aircraft surfacing with identification aspect, the present invention provides a kind of effective suitable
For the recognition methods of typical aircraft Facing material, this method can be suitble in the biggish situation of spectroscopic data amount carry out quickly,
The high material classification method of accuracy.
To achieve the goals above, the present invention is achieved by the following technical scheme:
A kind of typical aircraft Facing material recognition methods neural network based, includes the following steps:
Step 1: a large amount of diffusing reflection infrared spectroscopies of a variety of typical aircraft surfacings are acquired using near infrared spectrometer,
Establish basic database；
Step 2: the spectrum acquired to step 1 pre-processes,
2.1) ambient noise is removed:
The ir data that air background is repeatedly measured using near infrared spectrometer, seeking wavelength is 900-1700nm model
The average value of the spectrum relative intensity repeatedly measured in enclosing, it is collected a variety of common winged from step 1 as background intensity
It is cut in a large amount of diffusing reflection ir datas of row device surfacing；
2.2) discrete point is judged using Euclidean distance, remove obvious invalid spectrum；
2.3) using Noise Elimination from Wavelet Transform method to signal smoothing processing；
2.4) effective wavelength range: 936.47-1682.01nm is intercepted；
2.5) removal instrument passes letter:
It determines that instrument biography function evidence is identical as experimental data effective wavelength and resolution ratio, passes letter divided by instrument with experimental data
Data pass letter that is, multiplied by instrument fair curve to remove laboratory apparatus；
2.6) removal light source reflected light influences:
Determine that light source data is identical as experiment valid data wavelength, it is unrestrained anti-using the multiple testing standard blank of former experimental provision
Near infrared spectrum is penetrated, effective spectrum mean value is taken；With experimental data divided by standard white plate diffusing reflection near infrared spectrum data, realize
The removal of light source reflected light；
2.7) normalization data is handled:
By treated above, effectively spectroscopic data is normalized, and spectroscopic data intensity is in [- 1,1] after normalization
Between；
Step 3: will treated establishment of spectrum spectra database；According to traditional material detection method to aircraft surface material
It is detected, and corresponding material infrared spectroscopy is demarcated, establish ir data and conventional aircraft Facing material
With model；
Step 4: being classified according to aircraft surface material to database infrared spectroscopy；
Step 5: spectroscopic data is formatted, deep learning and training are carried out to database using neural network,
Establish its neural network model；The neural network is divided into input layer and competition layer, and each neuron of network competition layer passes through competing
Strive to obtain the respond opportunity to input pattern, only one last neuron becomes victor, and by with triumph neuron
Each connection weight is adjusted towards the direction for being more advantageous to its competition；
Wherein, the input layer is made of N number of neuron, and competition layer has M neuron；The connection weight of network is wij(i
=1,2 ..., N；J=1,2 ..., M), and meet constraint condition
Input layer input sample is binary set, each element value 0 or 1, the calculating side of the state of competition layer neuron j
Formula is as follows:
In formula (1), xiMaximum weighted is had in competition layer according to competition mechanism for i-th of element of input sample vector
The neuron k of value wins competition triumph, exports and is
Weight after competition is carried out as follows amendment,
In formula (2), a is learning parameter, 0 < a < < 1；M is the neuron number that input is 1 in input layer, i.e.,
In weighed value adjusting formulaItem indicates to work as xiWhen being 1, weight increases；And work as xiWhen being 0, weight reduces；Work as xiIt is living
When jump, corresponding i-th of weight is increased by, and is otherwise reduced by；When reaching preset the number of iterations, terminate algorithm；
Step 6: spectra collection being carried out to aircraft surface material according to method same as step 1；
Step 7: the spectrum of acquisition being pre-processed according to method same as step 2, and will treated spectrum number
According to the neural network model that input step 5 constructs, i.e. output material information.
Compared with existing technology, the advantages and positive effects of the present invention:
(1) present invention is by the pretreatment to ir data, and the data using lane database are foundation, constantly training
Neural network model will couple weight constantly towards the direction adjustment for being more advantageous to its competition with each of triumph neuron, make it
Preferable effect can be reached in the case where frequency of training is less, and be able to achieve and clustering is made to measure spectrum automatically；This
Normalized is used when invention pretreatment, keeps data processing faster and accurate, so that this neural network model is to aircraft
The accuracy rate of surfacing identification averagely reaches 85% or more.
(2) present invention is simple using Self-organizing Competitive Neutral Net model analysis, data processing speed is fast, meets current big
The requirement of data analysis.After adopting this method be not necessarily to sample preparation, can outfield sample is analyzed, pass through nerve later
Network compares spectroscopic data, analyzes, matches and classifies, and can reach the purpose of the identification to different aircraft.
Detailed description of the invention
Fig. 1 is the flow chart of aircraft surface Material Identification method.
Fig. 2 is sample spectrogram after removal ambient noise.
Fig. 3 is sample spectrogram after the processing of Fig. 2 data normalization.
Specific embodiment
In order to illustrate more clearly of technical solution of the present invention, to make the present invention below by specific embodiment further
Explanation, but these embodiments simply to illustrate that, rather than limiting the invention.
Parameter Map 1 is to Fig. 3, a kind of typical aircraft Facing material recognition methods neural network based, including walks as follows
It is rapid:
Step 1: a large amount of diffusing reflection infrared spectroscopies of a variety of typical aircraft surfacings are acquired using near infrared spectrometer,
Establish basic database；The typical aircraft surfacing includes: aluminium sheet, steel plate, carbon fiber board, monocrystalline silicon piece, solar energy
Solar panel etc.；
Step 2: the spectrum acquired to step 1 pre-processes,
2.1) ambient noise is removed:
The ir data that air background is repeatedly measured using near infrared spectrometer, seeking wavelength is 900-1700nm model
The average value of the spectrum relative intensity repeatedly measured in enclosing, it is collected a variety of common winged from step 1 as background intensity
It is cut in a large amount of diffusing reflection ir datas of row device surfacing；If spectral intensity goes out negative value again, taken absolute value for
0；
2.2) discrete point is judged using Euclidean distance, remove obvious invalid spectrum；
2.3) using Noise Elimination from Wavelet Transform method to signal smoothing processing；
2.4) effective wavelength range: 936.47-1682.01nm is intercepted；
2.5) removal instrument passes letter:
It determines that instrument biography function evidence is identical as experimental data effective wavelength and resolution ratio, passes letter divided by instrument with experimental data
Data pass letter that is, multiplied by instrument fair curve to remove laboratory apparatus；
2.6) removal light source reflected light influences:
Determine that light source data is identical as experiment valid data wavelength, it is unrestrained anti-using the multiple testing standard blank of former experimental provision
Near infrared spectrum is penetrated, effective spectrum mean value is taken；With experimental data divided by standard white plate diffusing reflection near infrared spectrum data, realize
The removal of light source reflected light；
2.7) normalization data is handled:
By treated above, effectively spectroscopic data is normalized, and spectroscopic data intensity is in [- 1,1] after normalization
Between；
Step 3: will treated establishment of spectrum spectra database；According to traditional material detection method to aircraft surface material
It is detected, and corresponding material infrared spectroscopy is demarcated, establish ir data and conventional aircraft Facing material
With model；
Step 4: being classified according to aircraft surface material to database infrared spectroscopy；
Step 5: spectroscopic data is formatted, deep learning and training are carried out to database using neural network,
Establish its neural network model；The neural network is divided into input layer and competition layer, and each neuron of network competition layer passes through competing
Strive to obtain the respond opportunity to input pattern, only one last neuron becomes victor, and by with triumph neuron
Each connection weight is adjusted towards the direction for being more advantageous to its competition；
Wherein, the input layer is made of N number of neuron, and competition layer has M neuron；The connection weight of network is wij(i
=1,2 ..., N；J=1,2 ..., M), and meet constraint condition
Input layer input sample is binary set, each element value 0 or 1, the calculating side of the state of competition layer neuron j
Formula is as follows:
In formula (1), xiMaximum weighted is had in competition layer according to competition mechanism for i-th of element of input sample vector
The neuron k of value wins competition triumph, exports and is
Weight after competition is carried out as follows amendment,
In formula (2), a is learning parameter, 0 < a < < 1；M is the neuron number that input is 1 in input layer, i.e.,
In weighed value adjusting formulaItem indicates to work as xiWhen being 1, weight increases；And work as xiWhen being 0, weight reduces；Work as xiIt is living
When jump, corresponding i-th of weight is increased by, and is otherwise reduced by；When reaching preset the number of iterations, terminate algorithm；
Step 6: spectra collection being carried out to aircraft surface material according to method same as step 1；
Step 7: the spectrum of acquisition being pre-processed according to method same as step 2, and will treated spectrum number
According to the neural network model that input step 5 constructs, i.e. output material information.
Claims (1)
1. a kind of typical aircraft Facing material recognition methods neural network based, which comprises the steps of:
Step 1: acquiring a large amount of diffusing reflection infrared spectroscopies of a variety of typical aircraft surfacings using near infrared spectrometer, establish
Basic database；
Step 2: the spectrum acquired to step 1 pre-processes,
2.1) ambient noise is removed:
The ir data that air background is repeatedly measured using near infrared spectrometer, seeking wavelength is within the scope of 900-1700nm
The spectrum relative intensity repeatedly measured average value, as background intensity from step 1 a variety of typical aircrafts collected
It is cut in a large amount of diffusing reflection ir datas of surfacing；
2.2) discrete point is judged using Euclidean distance, remove obvious invalid spectrum；
2.3) using Noise Elimination from Wavelet Transform method to signal smoothing processing；
2.4) effective wavelength range: 936.47-1682.01nm is intercepted；
2.5) removal instrument passes letter:
It determines that instrument biography function evidence is identical as experimental data effective wavelength and resolution ratio, passes function divided by instrument with experimental data
According to that is, multiplied by instrument fair curve, to remove laboratory apparatus biography letter；
2.6) removal light source reflected light influences:
Determine that light source data is identical as experiment valid data wavelength, it is close using the multiple testing standard blank diffusing reflection of former experimental provision
Infrared spectroscopy takes effective spectrum mean value；With experimental data divided by standard white plate diffusing reflection near infrared spectrum data, light source is realized
Reflected light removal；
2.7) normalization data is handled:
Treated that effectively spectroscopic data is normalized by above, after normalization spectroscopic data intensity [- 1,1] it
Between；
Step 3: will treated establishment of spectrum spectra database；Aircraft surface material is carried out according to traditional material detection method
Detection, and corresponding material infrared spectroscopy is demarcated, ir data, which is established, with conventional aircraft Facing material matches mould
Type；
Step 4: being classified according to aircraft surface material to database infrared spectroscopy；
Step 5: spectroscopic data being formatted, deep learning and training are carried out to database using neural network, established
Its neural network model；The neural network is divided into input layer and competition layer, each neuron of network competition layer by compete come
The respond opportunity to input pattern is obtained, only one last neuron becomes victor, and will be each with triumph neuron
Weight is connect to adjust towards the direction for being more advantageous to its competition；
Wherein, the input layer is made of N number of neuron, and competition layer has M neuron；The connection weight of network is wij(i=1,
2 ..., N；J=1,2 ..., M), and meet constraint condition
Input layer input sample is binary set, and each element value 0 or 1, the calculation of the state of competition layer neuron j is such as
Under:
In formula (1), xiMaximum weighted value is had in competition layer according to competition mechanism for i-th of element of input sample vector
Neuron k wins competition triumph, exports and is
Weight after competition is carried out as follows amendment,
In formula (2), a is learning parameter, 0 < a < < 1；M is the neuron number that input is 1 in input layer, i.e.,
In weighed value adjusting formulaItem indicates to work as xiWhen being 1, weight increases；And work as xiWhen being 0, weight reduces；Work as xiWhen enlivening,
Corresponding i-th of weight is increased by, and is otherwise reduced by；When reaching preset the number of iterations, terminate algorithm；
Step 6: spectra collection being carried out to aircraft surface material according to method same as step 1；
Step 7: the spectrum of acquisition being pre-processed according to method same as step 2, and spectroscopic data is defeated by treated
Enter the neural network model of step 5 building, i.e. output material information.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810662092.8A CN108872130B (en) | 2018-06-25 | 2018-06-25 | Typical aircraft Facing material recognition methods neural network based |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810662092.8A CN108872130B (en) | 2018-06-25 | 2018-06-25 | Typical aircraft Facing material recognition methods neural network based |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108872130A CN108872130A (en) | 2018-11-23 |
| CN108872130B true CN108872130B (en) | 2019-08-23 |
Family
ID=64294596
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810662092.8A Active CN108872130B (en) | 2018-06-25 | 2018-06-25 | Typical aircraft Facing material recognition methods neural network based |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108872130B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110717510A (en) * | 2019-09-03 | 2020-01-21 | 天津大学 | Material distinguishing method based on deep learning and atomic force microscope force curve |
| CN112182965B (en) * | 2020-09-25 | 2024-02-02 | 苏州融睿纳米复材科技有限公司 | Method for designing infrared phase change material based on deep learning |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015195746A1 (en) * | 2014-06-18 | 2015-12-23 | Innopix, Inc. | Spectral imaging system for remote and noninvasive detection of target substances using spectral filter arrays and image capture arrays |
| FR3037429B1 (en) * | 2015-06-15 | 2018-09-07 | Donecle | SYSTEM AND METHOD FOR AUTOMATIC SURFACE INSPECTION |
| CN107844760A (en) * | 2017-10-24 | 2018-03-27 | 西安交通大学 | Three-dimensional face identification method based on curved surface normal direction component map Neural Networks Representation |
-
2018
- 2018-06-25 CN CN201810662092.8A patent/CN108872130B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN108872130A (en) | 2018-11-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108872130B (en) | Typical aircraft Facing material recognition methods neural network based | |
| Obert | Aerodynamic design of transport aircraft | |
| CN110363151A (en) | Based on the controllable radar target detection method of binary channels convolutional neural networks false-alarm | |
| Streett et al. | Aerodynamic noise reduction for high-lift devices on a swept wing model | |
| Zhou et al. | Acoustic and radar integrated stealth design for ducted tail rotor based on comprehensive optimization method | |
| CN109164429B (en) | Multi-fractal characteristic aircraft target classification method based on principal component analysis | |
| CN113362604B (en) | Controller workload assessment method based on sector dynamic traffic characteristics | |
| Tao et al. | A novel optimization method for maintaining aerodynamic performances in noise reduction design | |
| Moshkov et al. | Effect of spacing between the pusher propeller and wing on environmental light aircraft noise | |
| CN112623254B (en) | Hybrid laminar flow wing air suction energy loss engineering calculation method | |
| CN117622552A (en) | Long-endurance unmanned aerial vehicle and pneumatic optimization method thereof | |
| Zimmer | Integral design and optimisation process for a highly flexible generic long range jet transport with flight mechanic derivative constraints | |
| Flores-Mezarina et al. | Application of wavy leading edge to enhance winglet aerodynamic performance | |
| CN110163227B (en) | Airport runway pavement airworthiness discrimination method based on pattern recognition | |
| George et al. | Noise due to tip vortex formation on lifting rotors | |
| Candeloro et al. | Experimental Fluid Dynamic Characterization of Serrated Rotors for Drone Propulsion | |
| Davari et al. | Impact of Canard on the Flowfield over the Wing in Various Flow Regimes | |
| Ye et al. | Intelligent Recognition of Aircraft Type Using Audio Feature | |
| Lin et al. | Semispan Test Results of a Conventional High-Lift Common Research Model in Landing Configuration | |
| UENO et al. | Propulsion/airframe Integration Considering Low Drag and Low Sonic Boom | |
| Sedky et al. | Experiments in transverse gust mitigation using open-loop pitch maneuvers | |
| CN118626894B (en) | Aircraft stall scene recognition method and system based on deep learning | |
| Kanazaki et al. | Efficient global optimization applied to design and knowledge discovery of supersonic wing | |
| Kay | Aerodynamic Response of a Two-Dimensional Cambered Wing at Low Reynolds Numbers in Steady-State Onset Turbulence | |
| Pinindriya et al. | Impact of Airfoil Section on Winglet Design for Enhancing Aerodynamics Performance of Aircraft using CFD Analysis |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |