JP2010197249A - Vehicle weight measuring system and method for vehicle passing on bridge, and computer program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a vehicle weight measuring system for calculating a vehicle weight with high precision without having adverse effect upon a vehicle running speed. <P>SOLUTION: A vehicle detection processing part 21 outputs detection vehicle image data 33 based on monitor image data 31 photographed by a camera 5 and entry data 32 measured by a strain gauge 9 for axle detection. An axle position identification processing part 23 traces the detection vehicle image data 33 from the monitor image data 31 at every determined time and performs position identification of a leading axis based on axis interval data. An interference line data conversion processing part 24 converts time axis influence line data 40 to time axis influence line data 35 based on the axis interval data. A vehicle weight calculation part 25 performs vehicle weight calculation processing based on axis weight calculation data 36 measured by the strain gauge 7 for axis weight calculation and the time axis influence line data 35, thereby to determine the weight of a vehicle 10. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a vehicle traveling on a road, and more particularly to a vehicle weight calculation system, a vehicle weight calculation method, and a program for calculating a vehicle weight of a heavy vehicle.

  In maintaining and managing a bridge, the axle weight (axle weight) of a heavy vehicle passing through the bridge is important information for predicting damage to the bridge. In order to measure axle weight, a technique (Weigh In Motion) has been proposed in which the strain value when passing through a vehicle is continuously measured from a strain gauge installed in the main girder of the bridge (hereinafter referred to as an axle weight calculation strain gauge) and the axle weight is calculated. ing. In this method, it is necessary to obtain a time T at which the vehicle passes immediately above the axle load calculation strain gauge. FIG. 28 is a diagram showing an axle weight calculation strain gauge and an axle detection strain gauge installed in the main girder of the bridge. For example, in the technique described in Non-Patent Document 1, for each lane, in addition to the axle load calculation strain gauge 2001, axle detection strain gauges 2002a and 2002b are additionally installed at two locations sensitive to axle passage. Axle passing time T is specified by a technique.

Strain increases gradually before the axle passes, and gradually decreases after the axle passes. FIG. 29 is a diagram showing the time change of the strain detected by the axle detection strain gauges 2002a and 2002b. Peak detection is performed from the distorted waveform shown in FIG. 29, and the first peak time is set as the first axis passing timing T1, and the second peak time is set as the second axis passing timing T2.
As shown in FIG. 28, when the installation intervals of the two axle detection strain gauges 2002a and 2002b and the axle load calculation strain gauge 2001 are D1 and D2, respectively, the axle calculates the axle load from the passage timings T1 and T2. The time T that has passed right above the strain gauge is calculated by the following equations (1) and (2), for example.

  However, in Formula (1) and Formula (2), v is a vehicle passage speed.

"Long-term traffic weighted monitoring by real-time fully automatic processing weight-in-motion" Japan Society of Civil Engineers, October 2004, No. 773 / I-69, pp. 99-112

  However, the method described in Non-Patent Document 1 is a method for obtaining the time T on the assumption that the vehicle travels at a constant speed, and when acceleration / deceleration occurs in the vehicle speed, such as a decrease in the passing speed due to traffic congestion on the road. There is a problem that it is difficult to apply. FIG. 30 is a diagram showing a temporal change in strain detected by the axle detection strain gauges 2002a and 2002b during a traffic jam. As shown in FIG. 30, when the traffic is congested, the measurement results of the axle detection strain gauges 2002a and 2002b have a shape in which the time axis is extended, and the peak when passing through the axle becomes dull, so the passing timing (T1, T2) is correctly detected. It becomes impossible. As a result, it becomes impossible to correctly calculate the time T when the axle passes immediately above the strain gauge for calculating the axle load, and there arises a problem that an accurate vehicle weight cannot be calculated.

  The present invention has been made in view of the above points, and is a vehicle weight calculation system, a vehicle weight calculation method, and a program capable of accurately calculating the vehicle weight without being affected by the traveling speed of the vehicle. The purpose is to provide.

  In order to achieve the above-described object, the present invention includes a strain meter that measures road strain, and a vehicle weight calculation device that calculates a vehicle weight of a vehicle having a plurality of axles that travel on the road by the strain. In the vehicle weight calculation system, the vehicle weight calculation device measures strain by an axis interval data storage unit that stores in advance axis interval data corresponding to the vehicle type of the vehicle in association with the vehicle type, and the strain gauge. The vehicle type of the vehicle is identified using monitoring video data obtained by photographing the vehicle that has entered the load measurement range at predetermined time intervals, and the axis interval data of the vehicle is referred to the axis interval data storage unit. A shaft position characteristic that identifies the position of the axle of the vehicle at each predetermined time based on the shaft distance specification processing unit for specifying the position, the monitoring video data, and the shaft distance data. The influence line data indicating, for each position of the axle, the distortion when the axle having a reference axle load passes through the road at a constant speed is processed in advance. An influence line data conversion processing unit that generates time axis influence line data that is converted based on the position of the axle for each determined time and reflects the influence of the time required for passing the vehicle in the influence line data; and A vehicle weight calculation system comprising: a vehicle weight calculation unit configured to calculate a vehicle weight of the vehicle based on the strain of the vehicle measured by a strain gauge and the time axis influence line data.

  In the invention described above, the present invention further includes a vehicle image storage unit that stores vehicle images of a plurality of vehicle types, and the shaft interval identification processing unit detects a vehicle image of the vehicle detected from the monitoring video data. Then, the similarity with the vehicle image stored in the vehicle image storage unit is calculated, and the vehicle type of the vehicle image with the highest similarity is specified as the vehicle type of the vehicle.

  The present invention may further include a vehicle detection processing unit that obtains a detected vehicle image of the vehicle from the monitoring video data when the vehicle enters the load measurement range according to the invention described above, The processing unit detects a region similar to the detected vehicle image obtained by the vehicle detection processing unit in the monitoring video data photographed at the predetermined time, and the moving pixel of the region at the predetermined time The number is obtained, and the position of the axle for each predetermined time is specified from the distance corresponding to the number of moving pixels.

  In the invention described above, the vehicle weight calculation unit obtains a parameter that best fits the time axis influence line data for each axle to the strain measured by the strain gauge, and determines the axle from the parameter. The weight of each vehicle is calculated, and the sum of the calculated weights for each axle is used as the vehicle weight of the vehicle.

  Further, the present invention is a vehicle weight calculation method for calculating a vehicle weight of a vehicle having a plurality of axles traveling on the road by a strain gauge that measures the strain of the road, and an axial interval corresponding to the vehicle type of the vehicle The step of storing the data in advance in the axis interval data storage unit in association with the vehicle type, and the axis interval specifying processing unit is predetermined for the vehicle that has entered the load measurement range in which the strain is measured by the strain gauge The step of identifying the vehicle type of the vehicle using monitoring video data taken every time and identifying the axis interval data of the vehicle with reference to the axis interval data storage unit; Based on the video data and the axis interval data, the step of identifying the axle position of the vehicle for each predetermined time, and the influence line data conversion processing unit The influence line data indicating, for each position of the axle, the distortion when the axle has passed through the road at a constant speed is set to the axle position at the predetermined time specified by the shaft position specifying processing unit. Generating time-axis influence line data in which the influence of the time required for passing the vehicle is reflected in the influence line data, and a vehicle weight calculation unit of the vehicle measured by the strain gauge Calculating the vehicle weight of the vehicle based on the strain and the time-axis influence line data.

  Further, the present invention provides a computer that calculates the vehicle weight of a vehicle having a plurality of axles traveling on the road by a strain gauge that measures the strain of the road, the axis distance data corresponding to the vehicle type of the vehicle, Using the monitoring video data obtained by capturing the vehicle that has entered the load measurement range in which the strain is measured by the strain gauge at predetermined time intervals. The vehicle type of the vehicle is specified, the axis interval data of the vehicle is specified with reference to the axis interval data storage unit, and the predetermined time interval is determined based on the monitoring video data and the axis interval data. The step of identifying the position of the axle of the vehicle and the distortion when the axle having the reference axle weight passes through the road at a constant speed are shown for each position of the axle. The sound ray data is converted based on the position of the axle for each predetermined time specified by the shaft position specifying processing unit, and the influence of the time required for passing the vehicle is reflected in the influence line data. A program for executing a step of generating time axis influence line data and a step of calculating a vehicle weight of the vehicle based on the strain of the vehicle measured by the strain gauge and the time axis influence line data It is.

  According to the present invention, it is possible to accurately calculate the vehicle weight without being affected by the traveling speed of the vehicle.

1 is a schematic configuration diagram of a vehicle weight calculation system according to an embodiment of the present invention. It is a figure which shows the measurement of the distortion | strain of the bridge by the axial load which concerns on the same embodiment. It is a figure which shows the example of installation of the strain gauge for axle load calculation and the strain gauge for axle detection in the road of 2 lanes concerning the embodiment. It is a figure which shows the definition of the coordinate in the road of 2 lanes concerning the embodiment. It is a figure which shows the influence data which concern on the same embodiment. It is a figure which shows the preparation method of the dictionary image data memorize | stored in the vehicle image database which concerns on the embodiment. It is a figure which shows the partial space method utilized when creating the dictionary image data based on the embodiment. It is a figure which shows the axial space | interval data which concern on the same embodiment. It is a flowchart which shows operation | movement of the vehicle weight calculation system which concerns on the same embodiment. It is a flowchart which shows the flow of the vehicle detection process which concerns on the same embodiment. It is a figure which shows the vehicle detection process which concerns on the same embodiment. It is a figure which shows how to cut out the detection vehicle image in the vehicle detection process which concerns on the embodiment. It is a flowchart which shows the flow of the axial distance specific process which concerns on the same embodiment. It is a figure which shows the axial space specific process which concerns on the same embodiment. It is a flowchart which shows the flow of the axial position specific process which concerns on the same embodiment. It is a figure which shows the tracking method of the detected vehicle image in the axial position specific process which concerns on the embodiment. It is a figure (the 1) which shows the concept of the influence line data which concern on the same embodiment. It is a figure (the 1) which shows the concept of the influence line data which concern on the same embodiment. It is a flowchart which shows the flow of the influence line data conversion process which concerns on the same embodiment. It is a figure showing an example of influence line data concerning the embodiment. It is a figure which shows an example of the time-axis influence line data corresponding to the axle shaft concerning the embodiment. It is a flowchart which shows the flow of the vehicle weight calculation process which concerns on the same embodiment. It is a figure which shows the data for a comparison which concerns on the same embodiment. It is a figure which shows the time-axis influence line data corresponding to the axle shaft which concerns on the same embodiment. It is a figure which shows the relationship of the time-axis influence line data corresponding to the axle shaft and parameter concerning the embodiment. It is the figure which showed an example of the statistical data according to vehicle weight obtained by the same embodiment. It is a figure which shows an example of installation of the strain gauge in the multiple lane in other embodiment which concerns on this invention. It is a figure which shows the example of installation of the strain gauge in a bridge. It is a figure which shows the measurement data of the strain gauge for axle detection. It is a figure which shows the measurement data of the strain gauge for axle detection of the vehicle which decelerated.

  Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings.

  FIG. 1 is a schematic configuration diagram of a vehicle weight calculation system 1. The vehicle weight calculation system 1 includes a vehicle weight calculation device 3, a camera 5, an axle weight calculation strain gauge 7, and an axle detection strain gauge 9. Each of the camera 5, the axle weight calculation strain gauge 7, the axle detection strain gauge 9, and the vehicle weight calculation device 3 can transmit and receive data via the network 2. The network 2 is, for example, the Internet, and data transmission / reception may be wireless communication or wired communication.

  The vehicle weight calculation device 3 is a computer, for example, and is installed in a monitoring center or the like, and includes a control unit 11, an input unit 12, a display unit 13, a communication unit 14, and a storage unit 15. The control unit 11 is a central processing unit (CPU), a microprocessor, and the like, and controls each unit of the vehicle weight calculation device 3 and includes functional units described later. The input unit 12 is, for example, a keyboard or a mouse, and inputs information in response to an operation of the operator of the vehicle weight calculation device 3. The display unit 13 is, for example, a display device or a speaker, and outputs data stored in the storage unit 15. The communication unit 14 transmits / receives data to / from the camera 5, the axle load calculation strain gauge 7, and the axle detection strain gauge 9 via the network 2. The storage unit 15 is, for example, a nonvolatile memory, a volatile memory, a hard disk, or the like, and stores data to be described later.

The camera 5, the axle load calculation strain gauge 7, and the axle detection strain gauge 9 are installed on the road. The camera 5 captures the vehicle 10 traveling on the road, and transmits the captured image to the vehicle weight calculation device 3 as monitoring video data 31.
As described above, what is important in maintaining the bridge is to obtain the vehicle weight of the heavy vehicle passing through the bridge. Therefore, the vehicle 10 detected by the present system is mainly a heavy vehicle such as a trailer.

  FIG. 2 is a diagram showing the principle according to the present embodiment. In this embodiment, the axial load calculating strain gauge 7 is installed on the lower surface of the bridge girder, and the elongation strain of the portion is measured. Specifically, as shown in the upper diagram of FIG. 2, the passage of the vehicle 10 causes the bridge girder to bend downward, thereby extending the lower surface of the bridge girder and measuring the elongation strain. The lower diagram of FIG. 2 is a graph showing the strain actually measured. The axial load calculating strain gauge 7 according to the present embodiment is configured by, for example, an optical fiber strain gauge, and measures strain in 1/100 seconds (that is, 100 Hz).

  Here, the strain is a deformation of the shape of the substance, and is locally an amount of expansion / contraction of the measurement location. For example, when the strain at a certain location is 0.1, it indicates that the location is 0.9 times longer. This indicates that a reduction of 10% has occurred. Therefore, the strain is a ratio and is a value having no unit. Although there is no unit, in general, 1/100 strain is often expressed as “%”, 1/1000000 strain is often expressed as “μ”, and strain in the extension direction is often expressed as “+”. The strain in this embodiment also follows this notation.

Next, with reference to FIGS. 3 and 4, an installation example of the axle load calculation strain gauge 7 and the axle detection strain gauge 9 on a two-lane road and the definition of coordinate axes used in the following description will be described.
As shown in FIG. 3, the axial load calculating strain gauge 7 is installed in a girder 63 between bridge piers 61 a and 61 b that support a road. The strain measured by the axle load calculating strain gauge 7 is transmitted as axle load calculating data 36 to the vehicle weight calculating device 3.
The axle detection strain gauge 9 is installed at the end of the beam 63. As shown in FIG. 4, when the coordinate axis X is set for each vehicle direction for each lane, the axle detection strain gauge 9 has a vehicle load measurement range (X = 0 [m] to X = L [m] range). ) Is measured and transmitted to the vehicle weight calculation device 3 as the entry data 32. Here, the load measurement range is a range in which strain is measured by the axial load calculation strain gauge 7 on the bridge, which is equal to the interval between the members supporting the load, and is a girder in a simple structure composed of a girder and a pier. It becomes the length of. Therefore, in this case, the length L [m] of the beam 63 between the piers 61a and 61b is obtained.

  Returning to FIG. 1, the control unit 11 of the vehicle weight calculation device 3 includes a vehicle detection processing unit 21, an axis interval specification processing unit 22, an axis position specification processing unit 23, an influence line data conversion processing unit 24, and a vehicle weight calculation unit 25. Have. Each processing unit is configured by the control unit 11 reading and executing a program for performing each process. For example, it is assumed that the program is stored in the storage unit 15.

  The vehicle detection processing unit 21 detects the vehicle 10 from the monitoring video data 31 captured by the camera 5. The shaft interval identification processing unit 22 identifies the vehicle type from the detected vehicle image that is the detected image of the vehicle 10, and identifies the interval between the axles from the identified vehicle type. The shaft position specification processing unit 23 specifies the position of the head axle of the vehicle 10 on the road from the detected vehicle image. The influence line data conversion processing unit 24 is a time axis that reflects the influence of the time required for the vehicle 10 to pass through the influence line data 40 indicating the change in the reference strain based on the position of the leading axis and the distance between the axles. The influence line data 35 is converted. The vehicle weight calculation unit 25 calculates the vehicle weight of the vehicle 10 from the time axis influence line data 35 and the axle weight calculation data measured by the axle weight calculation strain gauge 7.

  The storage unit 15 includes monitoring video data 31, approach data 32, detected vehicle image data 33, head axis position data 34, time axis influence line data 35, axle load calculation data 36, passing vehicle weight storage part 37, influence line data. 40, a vehicle image database (DB) 41, an axis interval database (DB) 42, and position correspondence table data 43 are stored.

  Among these data, the monitoring video data 31, the approach data 32, the detected vehicle image data 33, the head axis position data 34, the time axis influence line data 35, the axle weight calculation data 36, and the passing vehicle weight storage unit 37 are stored. The vehicle weight is input data or output data of processing by the processing unit of the camera 5, the axle load calculation strain gauge 7, the axle detection strain gauge 9, and the control unit 11.

As shown in FIG. 5, the influence line data 40 is obtained when a 1 [t (ton)] axle is placed at a position X (0 ≦ X ≦ L) on the bridge girder shown in FIGS. This is data obtained by defining the strain ε measured by the axle load calculating strain gauge 7 and defining it as a function of the position X. It is “0” outside the load measurement range. The influence line data 40 stored in the storage unit 15 is obtained by converting the strain obtained by passing a test vehicle having a known axle weight in advance when a 1-axle axle is mounted.
In addition, since the influence line data 40 changes for every load measurement range of a bridge, when processing about several load measurement ranges, it is necessary to memorize | store each influence line data.

The vehicle image database (DB) 41 stores in advance dictionary image data for each vehicle type that is referred to when the vehicle detection processing unit 21 specifies the vehicle type.
Here, generation of dictionary image data stored in the vehicle image database 41 will be described with reference to FIGS. 6 and 7. As shown in FIG. 6, a plurality of learning sample image data of a certain vehicle type G, for example, “trailer” is collected and processed into a feature image such as binarization by edge point extraction, for example. All processed images are regarded as one-dimensional vectors, individual covariance matrices are created and added together, and eigenvalues and eigenvectors are obtained for the added matrix. The d eigenvectors are added in descending order of the obtained eigenvalues, and the added vectors are converted into the original image size to obtain the dictionary image data of the vehicle type G.

The calculation in FIG. 6 is a calculation method called a subspace method, and FIG. 7 is a diagram showing the concept of this calculation. As shown in FIG. 7, the sample image data for learning the trailer is distributed in a certain space (in FIG. 7, an example is a two-dimensional space), and a vector that best represents this distribution is selected. Then, the dictionary image data is generated by converting the selected vector into the size of the original image.
The above processing is performed on all vehicle types, for example, “large trucks”, “tank trucks”, “mixer vehicles”, “large dump trucks”, etc., and dictionary image data for all vehicle types is created and stored in the vehicle image database 41. Remember.

  The axis interval database (DB) 42 has a data configuration as shown in FIG. 8 and stores the vehicle type and axis interval data for each vehicle type. For example, if the vehicle type is a “3-axle large truck”, there are three shafts, the axis spacing between the 1-2 axes is “5870 [mm]”, and the axis spacing between the 2-3 axes is “7070 [mm]. ] ".

The position correspondence table data 43 is data indicating the correspondence between the number of moving pixels (pixels) in an image frame and the actual distance (position). In general, it is considered that the relative positional relationship between the monitoring camera 5 and the road surface is constant. And the direction of the lane is known. Under this assumption, by measuring the actual distance on the lane, if A [pixel] moves in the lane direction in the image frame of the monitoring video data 31, it means that B [m] has moved on the actual road surface. It is possible to obtain a correspondence relationship in advance. The correspondence between the pixel A and the actual distance (position) B is stored in advance in the storage unit 15 as position correspondence table data 43 starting from the installation position of the axle detection strain gauge 9.
Note that when the shooting direction and zoom of the monitoring camera 5 can be remotely operated and the relative positional relationship with the road surface changes, the information of the position correspondence table data 43 is reproduced from the information such as the shooting direction and zoom. Will be calculated.

(Operation of vehicle weight calculation system)
Next, the operation of the vehicle weight calculation system 1 will be described.
FIG. 9 is a schematic flowchart showing the flow of the entire process of the vehicle weight calculation system 1.
The vehicle detection processing unit 21 performs vehicle detection processing of the vehicle 10 photographed by the camera 5 based on the monitoring video data 31 photographed by the camera 5 and the approach data 32 measured by the axle detection strain gauge 9 (step S101). The detected vehicle image data 33 is output and stored in the storage unit 15.

  The shaft interval identification processing unit 22 identifies the vehicle type of the vehicle 10 from the detected vehicle image data 33 detected by the vehicle detection processing unit 21 and the dictionary image data stored in the vehicle image database 41, and the axis based on the identified vehicle type. With reference to the interval database 42, an axis interval specifying process for detecting the axis interval data of the vehicle 10 is performed (step S102). The axis position specification processing unit 23 performs axis position specification processing based on the monitoring video data 31 captured by the camera 5 and the axis interval data detected by the axis interval specification processing unit 22, and outputs leading axis position data 34. It memorize | stores in the memory | storage part 15 (step S103).

  The influence line data conversion processing unit 24 is stored in advance in the storage unit 15 based on the head axis position data 34 obtained by the axis position identification processing unit 23 and the axis interval data obtained by the axis interval identification processing unit 22. The influence line data 40 is converted to perform influence line data conversion processing for generating time axis influence line data 35 corresponding to time, and the generated time axis influence line data 35 is stored in the storage unit 15 (step S104). The vehicle weight calculation unit 25 performs vehicle weight calculation processing based on the axle load calculation data 36 measured by the axle load calculation strain gauge 7 and the time axis influence line data 35 generated by the influence line data conversion processing unit 24. The vehicle weight calculation process for calculating the vehicle weight of the vehicle 10 is performed (step S105). Hereinafter, each process of step S101-S105 is demonstrated in detail.

(Vehicle detection process S101)
Next, the vehicle detection process (step S101) will be described with reference to FIGS.
As shown in FIG. 10, first, the vehicle detection processing unit 21 reads the monitoring video data 31 captured by the camera 5 from the storage unit 15 and creates a background image (step S201).
Here, the shooting direction of the monitoring camera 5 that acquires the monitoring video data 31 is fixed with respect to the road surface serving as the monitoring area, and the relative positional relationship is unchanged, or the shooting direction and the camera are fixed. It is assumed that the position of can be grasped. The monitoring video data 31 is obtained by acquiring image frames at regular intervals, and an acquisition time is given to each image frame. The monitoring video that is normally used complies with, for example, the NTSC (National Television System Committee) standard, and 30 image frames are acquired per second.

As shown in FIG. 11, the vehicle detection processing unit 21 reads a certain number of image frames from the monitoring video data 31 at certain time intervals and creates an average image of the image frames, so that there is a change between the image frames. In other words, the moving vehicle part is erased from the monitoring video data 31, and a background image 81 of only the background road surface is created.
The background image 81 is created at regular intervals and is constantly updated regardless of other processes. By updating the background image 81 at regular intervals, it is possible to appropriately reflect changes in brightness due to weather and time.

The vehicle detection processing unit 21 reads the entry data 32 of the vehicle 10 measured by the axle detection strain gauge 9 from the storage unit 15, determines whether or not the vehicle is a predetermined threshold value, and determines whether the vehicle is a heavy vehicle. Is determined (step S202).
In the bridge shown in FIG. 3 described above, the load measurement range for measuring the approach data is the entire bridge girder 63, and the axle detection strain gauge 9 is installed at the end of the girder in the approach direction of the vehicle 10. That is, it is considered that the vehicle 10 has entered the load measurement range when the axle detection strain gauge 9 detects the entry of a heavy vehicle. The load measurement range is from the position X = 0 [m] to X = L [m] on the beam 63 when the X axis parallel to the beam 63 is considered.

Even when the axle detection strain gauge 9 is installed at a point before the load measurement range, a practical problem does not occur because the time when the load measurement range is entered can be specified by the shaft position specifying process described later.
In addition, it is assumed that the measurement timings of the monitoring video data 31 and the approach data 32 are synchronized, and a fixed number (1 / Δt) of data is acquired every second at a constant interval Δt [seconds].
The threshold value of the entry data 32 serving as a determination criterion is determined by, for example, passing a vehicle whose axle load is known in advance through a bridge on which the axle detection strain gauge 9 is installed, and is input from the input unit 12. Accepted and stored in advance in the storage unit 15.

  In step S202, when the vehicle detection processing unit 21 does not determine that the vehicle 10 is a heavy vehicle, the vehicle detection processing unit 21 acquires the next entry data 32 and determines whether the vehicle 10 is a heavy vehicle. On the other hand, if the vehicle detection processing unit 21 determines in step 202 that the vehicle 10 is a heavy vehicle, the monitoring video data 31 at the time when the vehicle 10 entered the load measurement range is read from the storage unit 15 and shown in FIG. In addition, the detected vehicle image data 33 of only the vehicle portion is extracted by subtracting the background image 81 at that time (step S203).

In addition, about the vehicle image extraction process of step S203, since there are other moving vehicles, it is necessary to cut out only the detected vehicle. The following method is used for this.
As shown in FIG. 12, since the positional relationship between the camera 5 and the road surface is fixed, the installation position of the axle detection strain gauge 9 in the monitoring image frame can be specified in advance. Since it is found from the acquisition of the entry data 32 that the leading axle of the heavy vehicle is on the axle detection strain gauge 9, a region of a certain rearward range is cut out from the position in the monitoring image frame. Since the size of the vehicle is not known at the time of this process, the entire detected vehicle may not enter. Even in such a case, the purpose of using the detected vehicle image is to use the detected vehicle image to specify the vehicle model number by collating with the vehicle image database 41 and to specify the axle position by tracking the vehicle area in the monitoring video data 31. Therefore, for this purpose of use, the entire vehicle 10 does not have to be included in the detected vehicle image, and it may be created in a certain range of size defined above.

(Axis spacing identification process S103)
Next, with reference to FIGS. 13 and 14, the shaft interval specifying process by the shaft interval specifying processing unit 22 will be described.
As shown in FIG. 13, first, the shaft interval identification processing unit 22 collates the detected vehicle image data 33 obtained by the vehicle detection process with the dictionary image data stored in the vehicle image database 41 to determine the detected vehicle. A vehicle type is specified (step S301). The vehicle type is specified by the method shown in FIG. 14. The shaft interval specifying processing unit 22 performs the same processing as the dictionary image data on the detected vehicle image data 33 to obtain a feature image. And Then, the axis interval identification processing unit 22 adds the axis interval identification processing unit 22 vector A and the vector B of all dictionary image data stored in the vehicle image database 41 (a vector obtained by adding d eigenvectors in descending order of eigenvalues). ) And dividing this by the product of the magnitude value of vector A (| A |) and the magnitude value of vector B (| B |), the similarity (cos θ) is calculated. calculate.

  The shaft interval identification processing unit 22 determines the vehicle type having the highest similarity as the detected vehicle image data 33, that is, the vehicle type of the vehicle 10. For example, in the similarity shown in FIG. 14, the similarity of “trailer” is the largest “0.71”, so the vehicle type of the detected vehicle image data 33 is specified as “trailer”. Then, the shaft interval identification processing unit 22 collates the identified vehicle type with the axis interval database 42, identifies the axis interval, and acquires the axis interval data (step S302).

(Axis position specifying process S103)
Next, the shaft position specifying process in step S103 will be described with reference to FIGS.
The shaft position specifying process detects the position of the vehicle, that is, the detected position of the axle at regular intervals by tracking the detected vehicle image data 33 in time series with respect to the monitoring video data 31. . If the position of the leading axle can be detected, the other axles can be calculated from the axis interval data, and therefore, only the leading axle of the detected vehicle is subject to the shaft position specifying process.

  As shown in FIG. 15, the shaft position specifying processing unit 23 tracks the detected vehicle image data 33 in the monitoring video data 31 for each time, that is, the monitoring video data 31 for the vehicle 10 indicated in the detected vehicle image data 33 for each time. Position data is obtained by detecting the position in the image indicated by (step S401). The processing for obtaining the position data is performed according to the procedure shown in FIG.

In FIG. 16, it is assumed that an image frame is stored every Δt seconds in the monitoring video data 31, and the image frame stored at time T is indicated by F (T) in the following description. First, when the storage time of the image frame R0 from which the detected vehicle image data 33 to be input is first extracted is T = t, the axis position specifying processing unit 23 performs the following operation on the next image frame F (t + Δt). A region R1 that matches the detected vehicle image data 33 is found. This image matching process is performed by finding an area where the difference from the detected vehicle image data 33 is the smallest from the image frames of the monitoring video data 31 at each time. The number of moving pixels is the distance on the image between the position (x ₀ , y ₀ ) of the detected vehicle image data 33 at F (t) and the position (x ₁ , y ₁ ) at F (t + Δt). To obtain position data.
Note that (x ₀ , y ₀ ) is the coordinates in the monitoring video data 31 when the detected vehicle image data 33 to be input is first cut out, that is, the first axis of the vehicle 10 is the axle detection strain meter 9. The distance from the axle detection strain gauge 9 of the leading axis of the vehicle can be found by obtaining the distance from the origin. From this, the position data to be output is a value indicating the distance from the axle detection strain gauge 9 to the head axle by the number of moving pixels.
When the axle detection strain gauge 9 is installed in front of the load measurement range as described above, a deviation to the load measurement range is corrected in the processing.

  Then, the shaft position specification processing unit 23 specifies and specifies the actual position of the front axle by referring to the position correspondence table data 43 stored in advance in the storage unit 15 based on the obtained position data. First axis position data 34 indicating the position of the first axle is stored in the storage unit 15 (step S402).

Next, the shaft position specification processing unit 23 reads the axial distance data of the vehicle 10 and calculates the distance D [m] between the head axle and the tail axle. Assuming that the position where the load measurement range ends is L [m], by determining whether or not the leading axis position X [m] obtained in the immediately preceding axis position specifying process is larger than (L + D) [m], The shaft position specification processing unit 23 determines whether or not the vehicle 10 has left the load measurement range (step S403). If it is determined that the vehicle 10 has not left, the process returns to step S401 to perform image tracking at the next time step. If it is determined that the vehicle 10 has left the load measurement range, the process is terminated.
The processes of steps S401, S402, and S403 are repeatedly executed until the vehicle 10 leaves the load measurement range, and position data indicating the position of the leading axis is additionally stored in the storage unit 15.

  In addition, when the traveling region and direction of the vehicle in the image frame are known, the search region at the time of matching can be limited only to the traveling direction on the lane, and processing can be reduced.

(Influence line data conversion processing S104)
Next, the influence line data conversion process in step S101 will be described with reference to FIGS. As described above, the influence line data 40 is the strain gauge 7 for calculating the axle load when a 1 [t (ton)] axle is placed at the position X (0 ≦ X ≦ L) on the bridge girder. This is data obtained by indicating the measured strain ε and defining it as a function of the position X. And the influence line data 40 memorize | stored in the memory | storage part 15 are calculated | required by converting the data of the distortion at the time of passing a bridge through the test vehicle whose load is known in the case of 1 axle of 1 [t]. It is what As shown in FIG. 17, this conversion is performed when the 1 [t] single axle and the W [t] single axle are mounted. Use the relationship of becoming. For example, as shown in FIG. 18, when measurement is performed by passing a test vehicle having a shaft weight of 4 [t], the influence line data 40 is obtained by setting the obtained strain data to a value of 1/4. Will be able to.

  FIG. 19 is a flowchart showing the flow of influence line data conversion processing. The influence line data conversion processing unit 24 reads the axis interval data, the head axis position data 34, and the influence line data 40 of the vehicle 10 obtained by the above processing from the storage unit 15, converts the influence line data 40, and The time axis influence line data 35 is generated (step S501).

  As described in the above-described shaft position specifying process, the storage unit 15 stores n pieces of data as Δt intervals from the time t when the vehicle 10 enters the load measurement range to the exit as the head axis position data 34. Yes. The influence line data conversion processing unit 24 performs time T (T = t + Δt × n, n = 0, 1, 2,..., I,..., N tracked by the axis position specifying process in the influence line data 40. ) At position Pi, and is stored in the storage unit 15 as time axis influence line data 35 of the first axis. For the other axles, the influence line data conversion processing unit 24 reads the distance D between the axle and the head axis based on the axis interval data, sets the position as (Pi + D), performs the same processing as above, Is generated and stored in the storage unit 15.

  20 and 21 are diagrams for explaining the influence line data conversion processing. The influence line data conversion processing unit 24 acquires strain values at times T1, T2, T3, T4, T5, and T6 tracked by the axis position specifying process from the influence line data 40 shown in FIG. 20, and as shown in FIG. The time axis influence line data 35 is generated by performing conversion by arranging the strain values of T1, T2, T3, T4, T5, and T6 with the time axis as the horizontal axis. Thereby, the time axis influence line data 35 reflecting the influence of the time required for the passage of the vehicle 10 in the influence line data 40 can be generated.

(Vehicle weight calculation processing step S105)
Next, the vehicle weight calculation process will be described with reference to FIGS.
In the vehicle weight calculation process, the vehicle weight calculation method widely known as “Weight In Motion” is used, and the time axis influence line data 35 created by the influence line data conversion process is arranged on the time axis by the number of axles. The weight of each axle is calculated by finding a magnification that matches the strain data actually measured by the least square method. Finally, each axle weight is summed to obtain the vehicle weight.

FIG. 22 is a flowchart showing the flow of the vehicle weight calculation process.
The vehicle weight calculation unit 25 obtains a time range (t ≦ T ≦ t + Δt × n) in which the data is stored from the time axis influence line data 35, and uses the corresponding range data as the strain for calculating the axial load. The comparison data is generated by cutting out from the axial weight calculation data 36 measured in total 7 and the generated comparison data is stored in the storage unit 15 (step S601). Then, the vehicle weight calculation unit 25 obtains each magnification of the time axis influence line data 35 of each axle created by the influence line data conversion process most suitable for the comparison data by using the least square method, and calculates each axis. The mass is calculated (step S602), the total axle load is obtained, and the vehicle weight of the vehicle 10 is calculated (step S603).

  FIG. 23 is a diagram showing an example of the comparison data. In the following description, the comparison data is J (t). FIG. 24 is a diagram showing the time axis influence line data 35 created by the influence line data conversion process. The vehicle 10 is a triaxial vehicle (axes A, B, C), and the time with respect to each axle arranged on the time axis. The axis influence line data 35 are shown as JA (t), JB (t), and JC (t), respectively. In the least square optimization process of step S602, the comparison data is best approximated by J (t) using JA (t), JB (t), and JC (t) of the time axis influence line data 35 of each axle. Parameters a, b, and c are obtained. The least square error E is defined by the following equation (3), and parameters a, b, and c that minimize this E are searched by a numerical calculation optimization method.

  FIG. 25 is a diagram showing time axis influence line data 35 of each axle doubled at each magnification. The influence line data is strain data measured per ton. As a result of multiplying it by a times, b times, and c times, it becomes a value closest to the actually measured data. The masses are calculated as a [ton], b [ton], and c [ton], respectively, and at least the date and time when the vehicle 10 passed and the vehicle type of the vehicle 10 passed, the calculated vehicle weight value is passed through the vehicle weight storage unit 37. Remember me.

  As described above, in this embodiment, the position of the axle is specified in time series from the monitoring video data 31, and the time axis influence line data 35 reflecting the change in the position of the axle is created, whereby the axle weight of each axle is generated. Can be calculated with high accuracy and the vehicle weight can be obtained. And in this embodiment, since the influence line conversion process which considered the time required for the vehicle 10 to pass is performed, there exists the same effect also at the time of traffic congestion.

  FIG. 26 is a graph showing the number of vehicles for each vehicle weight for each date using the vehicle weight obtained by the vehicle weight calculation system 1 according to the present embodiment. By using such a graph, It is possible to compare and grasp the trend of the passing amount, and eventually determine the bridge deterioration prediction, inspection, repair necessary location, prediction of the construction time of the specified location, road expansion plan, regulation, guidance strengthening policy It can be used for

In the above-described embodiment, as illustrated in FIG. 3, a one-lane road has been described as an example. However, the present invention can also be applied to a multi-lane road. FIG. 27 is a diagram showing an installation example of a strain gauge when there are six main girders in two lanes on one side and four lanes. As shown in this example, the axle load calculation strain gauge 7 and the axle detection strain gauge 9 are installed in the main girder corresponding to each lane, and the same processing as described above is performed, so that a road with multiple lanes can be obtained. It is possible to determine the vehicle weight of the vehicle that passes.
Further, in the present embodiment, the processing has been described on the assumption that there is one vehicle 10 within the load measurement range. However, when considering a plurality of heavy vehicles at the same time, a plurality of processing in the case of the above-mentioned one vehicle are performed. They can be executed in parallel.

  It is also possible to create a program that realizes the processing performed by the control unit 11 and read the program by a general-purpose computer to realize the vehicle weight calculation device 3. This program may be recorded on a recording medium such as a CD-ROM, or distributed via a network.

DESCRIPTION OF SYMBOLS 1 ......... Vehicle weight calculation system 3 ......... Vehicle weight calculation apparatus 5 ......... Camera 7 ......... Axial load calculation strain gauge 9 ......... Axle detection strain gauge 10 ......... Vehicle 11 ......... Control Unit 15 ............ Storage unit 21 ............ Vehicle detection processing unit 22 ............ Axis interval identification processing unit 23 ............ Axis position identification processing unit 24 ............ Influence line data conversion processing unit 25 ............ Vehicle weight calculation Section 31... Surveillance video data 32... Incoming data 33... Detected vehicle image data 34... Lead axis position data 35 ... Time axis influence line data 36. …… Passing vehicle weight storage unit 40 ………… Influence line data 41 ………… Vehicle image database 42 ………… Axis interval database 43 ……… Position correspondence table data

Claims (6)

A vehicle weight calculation system comprising a strain gauge that measures road strain, and a vehicle weight calculation device that calculates a vehicle weight of a vehicle having a plurality of axles that travel on the road by the strain,
The vehicle weight calculation device includes:
An axis interval data storage unit for storing axis interval data corresponding to the vehicle type of the vehicle in advance in association with the vehicle type;
The vehicle type of the vehicle is identified using monitoring video data taken at predetermined intervals of the vehicle that has entered the load measurement range in which strain is measured by the strain gauge, and the shaft interval data storage unit is referred to. An axis interval identification processing unit for identifying axis interval data of the vehicle,
Based on the monitoring video data and the axis interval data, an axis position specifying processing unit that specifies the position of the axle of the vehicle for each predetermined time;
The predetermined time specified by the shaft position specifying processing unit is the influence line data indicating, for each position of the axle, the strain when the axle having a reference axle weight passes through the road at a constant speed. An influence line data conversion processing unit that generates time axis influence line data that is converted based on the position of each axle and that reflects the influence of the time required to pass the vehicle in the influence line data;
A vehicle weight calculation unit that calculates the vehicle weight of the vehicle based on the strain of the vehicle measured by the strain gauge and the time-axis influence line data;
A vehicle weight calculation system comprising: A vehicle image storage unit for storing vehicle images of a plurality of vehicle types;
The axis interval identification processing unit
The similarity between the vehicle image of the vehicle detected from the monitoring video data and the vehicle image stored in the vehicle image storage unit is calculated, and the vehicle type of the vehicle image with the highest similarity is identified as the vehicle type of the vehicle The vehicle weight calculation system according to claim 1, wherein: A vehicle detection processing unit that obtains a detection vehicle image of the vehicle from the monitoring video data when the vehicle enters the load measurement range;
The axis position specifying processing unit
In the monitoring video data taken at the predetermined time, a region similar to the detected vehicle image obtained by the vehicle detection processing unit is detected, and the number of moving pixels in the region at the predetermined time is obtained. 3. The vehicle weight calculation system according to claim 1, wherein a position of an axle for each predetermined time is specified from a distance corresponding to the number of moving pixels. 4.   The vehicle weight calculation unit obtains a parameter that best fits the time axis influence line data for each axle to the strain measured by the strain gauge, calculates the weight for each axle from the parameter, and calculates the calculated weight for each axle. The vehicle weight calculation system according to any one of claims 1 to 3, wherein a sum of weights is used as a vehicle weight of the vehicle. A vehicle weight calculation method for calculating a vehicle weight of a vehicle having a plurality of axles traveling on the road by a strain meter that measures road strain,
Storing the axis interval data corresponding to the vehicle type of the vehicle in advance in the axis interval data storage unit in association with the vehicle type;
The shaft interval identification processing unit identifies the vehicle type of the vehicle using monitoring video data obtained by photographing the vehicle that has entered the load measurement range in which strain is measured by the strain gauge at predetermined time intervals, and the shaft Identifying axis distance data of the vehicle with reference to the distance data storage unit;
An axis position specifying processing unit specifying the position of the axle of the vehicle for each predetermined time based on the monitoring video data and the axis interval data;
The influence line data conversion processing unit specifies the influence line data indicating, for each position of the axle, the distortion line when the axle having a reference axle weight passes through the road at a constant speed by the axis position specification processing unit. Converting based on the position of the axle for each predetermined time, and generating time axis influence line data reflecting the influence of the time required for passing the vehicle in the influence line data;
A vehicle weight calculation unit calculating the vehicle weight of the vehicle based on the strain of the vehicle measured by the strain gauge and the time axis influence line data;
The vehicle weight calculation method characterized by including. A computer that calculates the vehicle weight of a vehicle having a plurality of axles traveling on the road by a strain gauge that measures the strain of the road,
Storing the axis interval data corresponding to the vehicle type of the vehicle in advance in the axis interval data storage unit in association with the vehicle type;
The vehicle type of the vehicle is identified using monitoring video data taken at predetermined intervals of the vehicle that has entered the load measurement range in which strain is measured by the strain gauge, and the shaft interval data storage unit is referred to. Identifying the axle distance data of the vehicle,
Identifying the position of the axle of the vehicle for each predetermined time based on the monitoring video data and the axis spacing data;
The predetermined time specified by the shaft position specifying processing unit is the influence line data indicating, for each position of the axle, the strain when the axle having a reference axle weight passes through the road at a constant speed. Converting based on the position of each axle and generating time axis influence line data reflecting the influence of the time required for passing the vehicle in the influence line data;
Calculating the vehicle weight of the vehicle based on the strain of the vehicle measured by the strain gauge and the time axis influence line data;
A program for running

Images