JP6750967B2 - Vehicle type identification device and vehicle type identification method - Google Patents

Description

Embodiments of the present invention relate to a vehicle type identification device and a vehicle type identification method.

The type of vehicle to travel is determined for the purpose of charging road tolls and grasping the amount of traffic. Conventionally, as a method of determining the vehicle type of a vehicle traveling on a road, based on a technique of imaging or measuring the vehicle with a camera or a laser scanner, measuring the number of vehicle axles, vehicle height, and vehicle width, and the measurement result. A technique for discriminating a vehicle type is disclosed.

JP, 2014-215719, A Japanese Unexamined Patent Publication No. 9-89640 Japanese Patent No. 5478419 JP-A-11-232586

However, since the vehicle shape including the vehicle height and the vehicle width used in the conventional vehicle type identification technology changes with the passage of time, there is a demand for a more stable vehicle type identification method.

Further, in the conventional vehicle type discrimination technology, it is required to improve the accuracy of discriminating between a large vehicle and an oversized vehicle.

The vehicle type identification device according to the embodiment includes an acquisition unit, an image generation unit, a panoramic image estimation unit, an interpolation unit, an axle detection unit, and a determination unit. The acquisition unit acquires a detection point distance detected by irradiating a vehicle passing a predetermined position with a laser, which is detected by the scanner unit, and a reflection intensity of the laser reflected at the detection point. A plurality of still images obtained by continuously capturing positions are acquired. The image generation unit generates a reflectance image which is an image at a predetermined position including the detection point, in which the density of the detection point is changed according to the reflection intensity of the laser reflected at the detection point, and the reflection image is generated . Normalization processing is performed on the chest image to match the aspect ratio of the reflectance image with the actual aspect ratio of the vehicle . The panoramic image estimation unit estimates a panoramic image, which is an image showing the shape of the entire vehicle, based on a plurality of still images continuously captured by the image capturing unit . The interpolator performs matching processing for matching the feature points of the panoramic image with the detection points of the normalized reflectance image, and based on the result of the matching processing and the detection point distance, the feature points of the panoramic image are compared. Is interpolated. The axle detection unit detects an axle from a panoramic image in which feature points are interpolated, and measures the number of axles and the distance between the axles. The determination unit determines the vehicle type of the vehicle based on the number of axles and the distance between the axles.

1 is a diagram showing an example of a configuration of a vehicle type identification system according to a first embodiment. 3 is a block diagram showing an example of a functional configuration of the vehicle type identification device of the first embodiment. FIG. FIG. 3 is a diagram for explaining an example of a method for estimating a three-dimensional shape of a vehicle in the vehicle type identification device according to the first embodiment. FIG. 3 is a diagram showing an example of a panoramic image of a vehicle in the first embodiment. FIG. 3 is a diagram showing an example of a panoramic image of a vehicle after rotation processing according to the first embodiment. FIG. 5 is a diagram showing an example of point cloud data generated from detection point distances according to the first embodiment. FIG. 3 is a diagram showing an example of a reflectance image according to the first embodiment. FIG. 4 is a first diagram for explaining an example of a normalization process of a vehicle image included in a reflectance image in the vehicle type identification device of the first embodiment. FIG. 6 is a second diagram for explaining an example of a normalization process of a vehicle image included in a reflectance image in the vehicle type identification device of the first embodiment. FIG. 6 is a diagram for specifically explaining an example of a normalization process of a vehicle image included in a reflectance image in the vehicle type identification device of the first embodiment. FIG. 6 is a diagram for explaining an example of a procedure of image interpolation processing in the vehicle type identification device of the first embodiment. FIG. 3 is a diagram for explaining an example of matching processing in the vehicle type identification device of the first embodiment. FIG. 3 is a diagram for explaining an example of image interpolation processing in the vehicle type identification device of the first embodiment. FIG. 3 is a diagram for explaining an example of axle detection processing in the vehicle type identification device of the first embodiment. FIG. 3 is a diagram for explaining an example of a method of measuring the size of the axle and the center position of the axle in the vehicle type identification device of the first embodiment. FIG. 3 is a view for explaining an example of a lift axle function for lifting one axis in the vehicle type identification device of the first embodiment. FIG. 3 is a view for explaining an example of a lift axle function for lifting two axes in the vehicle type identification device of the first embodiment. FIG. 3 is a diagram showing an example of vehicle type identification information in the vehicle type identification device of the first embodiment. 6 is a flowchart showing an example of a procedure of vehicle type identification processing in the vehicle type identification device of the first embodiment. The figure which shows an example of a structure of the vehicle type discrimination system of Embodiment 2. 3 is a block diagram showing an example of a functional configuration of a vehicle type identification device according to a second embodiment. FIG. FIG. 6 is a diagram for explaining an example of normalization processing in the vehicle type identification device according to the second embodiment. The block diagram which shows an example of a function structure of the vehicle type discrimination device of Embodiment 3. FIG. 8 is a diagram for explaining an example of a counting process for each vehicle type classification in the vehicle type determination device of the third embodiment. 7 is a graph showing an example of a counting result of the number of passing vehicles per hour by vehicle type classification in the vehicle type identification device of the third embodiment. 9 is a flowchart showing an example of a procedure of vehicle type identification processing in the vehicle type identification device of the third embodiment. The figure which shows an example of a structure of the vehicle type discrimination system of Embodiment 4. FIG. 16 is a diagram showing an example of a plurality of still images captured by the camera of Embodiment 4 continuously. FIG. 13 is a diagram showing an example of a detection result of the laser scanner of the fourth embodiment. The block diagram which shows an example of a function structure of the vehicle classification device of Embodiment 4. The figure which shows an example of the weight analysis for every axle in the vehicle type identification device of Embodiment 4. The flowchart which shows an example of the procedure of the vehicle type discrimination process in the vehicle type discrimination apparatus of Embodiment 4.

(Embodiment 1)
The vehicle type determination system of the present embodiment uses the image of the vehicle captured by the camera and the detection point distance and the reflection intensity obtained by scanning the vehicle with the laser scanner to generate a high-resolution panoramic image of the vehicle, A more stable vehicle type determination is realized by measuring the number of axles and the distance between the axles from this panoramic image and determining the vehicle type classification and vehicle type of the vehicle based on the number of axles and the distance between the axles. Is. The details of this embodiment will be described below.

FIG. 1 is a diagram showing an example of the configuration of a vehicle type identification system 1a of this embodiment. As shown in FIG. 1, the vehicle type identification system 1a of the present embodiment includes a camera 101, a laser scanner 102, and a vehicle type identification device 103. The vehicle type identification system 1a is provided on a road through which large vehicles pass, an entrance/exit of a parking lot, a tollgate of an expressway, and the like.

The camera 101 of this embodiment is a monocular camera or the like, and captures an image of a predetermined position T through which the vehicle 10 passes. The camera 101 is an example of the image capturing unit in this embodiment.

The camera 101 of this embodiment is provided at a position diagonally above the vehicle 10 that has entered the predetermined position T. Since the legal vehicle height of the vehicle 10 is 3.8 m, specifically, the camera 101 is provided 5 to 7 m above the road surface of the road at the predetermined position T.

Further, the camera 101 of the present embodiment is provided at a depression angle of 25 to 45 degrees with reference to the road surface of the road at the predetermined position T. Further, the camera 101 of the present embodiment is provided so as to be able to capture the entire view of the vehicle 10 passing through the predetermined position T. Here, the panoramic view of the vehicle 10 refers to the overall shape of the vehicle 10.

Specifically, the camera 101 has a front region of the vehicle 10 (for example, a front end of a tractor portion of a large vehicle) and a rear region of the vehicle 10 (for example, a rear end of a trailer portion of a large vehicle) of the camera 101. It is provided to fit within the angle of view. Further, when the length of the vehicle 10 to be imaged is long, the camera 101 may adopt a configuration in which the camera 101 can be dynamically tilted and imaged in order to widen the angle of view of the camera 101.

Although the vehicle type identification system 1a of the present embodiment includes at least one camera 101, a configuration including a plurality of cameras 101 that image the front and rear or side surfaces of the vehicle 10 may be adopted. The installation position of the camera 101 is not limited to the above-described installation position as long as the camera 101 can image the vehicle 10. The installation angle of the camera 101 is not limited to the angle described above.

In this embodiment, the camera 101 captures an image of the vehicle 10 passing through the predetermined position T when an image capture trigger signal is input from the laser scanner 102. Alternatively, the camera 101 may be configured to continue capturing images at regular intervals, or the vehicle type identification system 1a may include a vehicle detection device (not shown) separately, and the vehicle detection device may transmit an image capture trigger signal to the camera 101. It is also possible to adopt a configuration for sending the.

The camera 101 continuously captures the vehicle 10 passing through the predetermined position T a plurality of times. The camera 101 transmits a captured image obtained by capturing the image of the vehicle 10 to the vehicle type identification device 103.

The laser scanner 102 irradiates an object with a laser such as infrared rays to scan the object. The laser scanner 102 irradiates a laser at a predetermined position T through which the vehicle 10 passes, and detects the detection point distance and the reflection intensity (reflectance) of the laser reflected at the detection point. Here, the detection point distance refers to the distance from the laser irradiation position to the detection point. The laser scanner 102 is an example of a scanner unit.

In the present embodiment, the laser scanner 102 is installed at a position where the vehicle 10 can be scanned from the side surface when the vehicle 10 enters the predetermined position T. Although the vehicle type identification system 1a of this embodiment includes at least one laser scanner 102, a configuration including a plurality of laser scanners 102 may be adopted.

When the laser scanner 102 detects that the vehicle 10 has entered the predetermined position T, the laser scanner 102 transmits to the camera 101 an imaging trigger signal instructing imaging of the vehicle 10. Further, the laser scanner 102 irradiates the vehicle 10 with laser until the vehicle 10 is no longer detected after the vehicle 10 is detected, that is, until the vehicle 10 enters the predetermined position T and then exits. Specifically, the laser scanner 102 detects the detection point distances and reflection intensities at a plurality of detection points while scanning the laser in a direction intersecting with the traveling direction of the vehicle 10 (hereinafter referred to as a scanning direction). The laser scanner 102 transmits the detection point distance and the reflection intensity detected for each detection point to the vehicle type identification device 103.

The vehicle type identification device 103 is an example of an electronic computer that executes vehicle type identification processing. The vehicle type identification device 103 is electrically connected to the camera 101 and the laser scanner 102.

The vehicle type identification device 103 of the present embodiment is installed on the side of the road adjacent to the predetermined position T as shown in FIG. 1, but the invention is not limited to this and may be installed in a management center or the like.

The vehicle type identification device 103 acquires a captured image from the camera 101. Further, the vehicle type identification device 103 acquires the detection point distance and the reflection intensity of each detection point from the laser scanner 102. Then, the vehicle type identification device 103 identifies the vehicle type of the vehicle 10 passing through the predetermined position T using the captured image, the detection point distance, and the reflection intensity.

FIG. 2 is a block diagram showing an example of the functional configuration of the vehicle type identification device 103 of this embodiment. As shown in FIG. 2, the vehicle type identification device 103 of this embodiment includes a storage unit 90, an acquisition unit 91, an image generation unit 92, a panoramic image estimation unit 93, an interpolation unit 94, and an axle detection unit 95. , And a determination unit 96.

The storage unit 90 stores the vehicle type identification information 20 which is a reference for identifying the vehicle type of the vehicle 10. The content of the vehicle type identification information 20 will be described later. In addition, the storage unit 90 can store information used for processing of each functional unit of the vehicle type identification device 103 according to the present embodiment. The storage unit 90 is a storage medium such as a hard disk drive (HDD), a solid state drive (SSD), or a memory.

The acquisition unit 91 acquires the detection point distance detected by the laser scanner 102 by irradiating the vehicle 10 passing through the predetermined position T with the laser, and the reflection intensity of the laser reflected at the detection point. Specifically, when the laser scanner 102 detects that the vehicle 10 has entered the predetermined position T, the acquisition unit 91 acquires the detection point distances and the reflection intensities of the plurality of detection points from the laser scanner 102. Further, the acquisition unit 91 acquires, from the camera 101, a captured image obtained by capturing an image of the predetermined position T.

The acquisition unit 91 sends the acquired detection point distance and reflection intensity to the image generation unit 92. The acquisition unit 91 also sends the acquired captured image to the panoramic image estimation unit 93.

The panoramic image estimation unit 93 estimates (generates) a three-dimensional image (hereinafter referred to as panoramic image D1) that shows the overall shape of the vehicle 10, based on the captured image obtained by the camera 101.

FIG. 3 is a diagram for explaining an example of a method of estimating the three-dimensional shape of the vehicle in the vehicle type identification device 103 of this embodiment. In the present embodiment, the panoramic image estimation unit 93 acquires from the acquisition unit 91 a plurality of captured images obtained by capturing a predetermined position T with the camera 101 multiple times. Then, the panoramic image estimation unit 93 estimates the panoramic image D1 of the vehicle 10 according to the plurality of captured images using a technique such as SfM (Structured from Motion).

The plurality of images 200 to 203 shown in FIG. 3 are examples of images in which the camera 101 continuously captures the vehicle 10 passing through the predetermined position T. For example, the leftmost image 200 is an image obtained by the camera 101 capturing the predetermined position T at time t0 when the vehicle 10 enters the predetermined position T. The image 203 at the right end is an image obtained by the camera 101 capturing the predetermined position T at time t3 when the vehicle 10 exits from the predetermined position T. The two central images 201 to 202 are images in which the camera 101 images the predetermined position T at predetermined times t1 and t2 after the vehicle 10 enters the predetermined position T and exits.

The lower part of FIG. 3 shows a method of estimating the amount of movement of the feature point P of the vehicle 10 between times t0 and t1 from the images 200 to 201. The characteristic points Pa and Pb are examples of the characteristic points P of the vehicle 10 included in the images 200 to 201. The locations used as the characteristic points Pa and Pb include the edges (corners) of the vehicle 10 included in the images 200 to 201.

A point a indicates a position where the characteristic point Pa included in the image 200 is projected on the projection plane 205 with the position 204 of the camera 101 as a reference. The point b indicates the position where the characteristic point Pb included in the image 200 is projected on the projection plane 205 with the position 204 of the camera 101 as a reference. Point a″ and point b″ indicate the positions where the characteristic points Pa and Pb included in the image 201 are projected on the projection plane 205 with the position 204 of the camera 101 as a reference. The projection plane 205 is a virtual plane defined for estimating the positions of the characteristic points Pa and Pb with the position of the camera 101 as a reference (origin).

The position of the projection plane 205 is fixed, and as the feature points Pa and Pb move between the image 200 and the image 201, the point a becomes the position of the point a″ and the point b becomes the position of the point b″. , On the projection plane 205. The panoramic image estimation unit 93 detects the movements of the characteristic points Pa and Pb in this manner and associates them with each other, thereby determining the movement amounts of the characteristic points Pa and the characteristic points Pb between the images 200 and 201 that are continuously captured. Can be estimated.

The panoramic image estimation unit 93 determines the feature points Pa and Pb from the movement amounts of the points a and b projected on the projection plane 205 based on the positional relationship between the vehicle 10 that is the subject and the position 204 of the camera 101. Estimate the 3D position.

The panoramic image estimation unit 93 associates the positions of the characteristic points Pa and Pb between the image 201 and the image 202 and between the image 202 and the image 203, and estimates the three-dimensional positions of the characteristic points Pa and Pb. To do. The panoramic image estimation unit 93 estimates the three-dimensional shape of the panoramic view of the vehicle 10 by sequentially estimating the amount of movement of the feature points between consecutive images in this manner.

When the camera 101 is not fixed and the vehicle 10 is imaged while moving, the panoramic image estimation unit 93 estimates the three-dimensional positions of the feature points Pa and Pb based on the moving amount of the camera 101. You may adopt it.

In this way, the panoramic image estimation unit 93 estimates the panoramic image D1 from a plurality of still images continuously captured by the camera 101. The panoramic image estimation unit 93 estimates a panoramic image D1 which is an image showing a three-dimensional shape of the panoramic view of the vehicle 10. In addition, the panoramic image estimation unit 93 may adopt a configuration in which the images of the panoramic image D1 are overlapped by pasting the captured images in the vicinity of the respective feature points to further improve the resolution of the panoramic image D1.

Although the four images 200 to 203 are illustrated in FIG. 3 as an example, the camera 101 may employ a configuration capable of capturing more images until the vehicle 10 passes the predetermined position T. The larger the number of captured images, the more the panoramic image estimation unit 93 can estimate the panoramic image D1 with higher accuracy. The feature points Pa and Pb are examples of the feature points P in the images 200 to 203, and the number of feature points used for alignment is not limited.

FIG. 4 is a diagram showing an example of the panoramic image D1 of the vehicle in the present embodiment. Point P indicates a feature point forming the panoramic image D1. Here, as described in FIG. 1, since the camera 101 images the vehicle 10 from an obliquely upper position, the estimated panoramic image D1 has a shape in which the vehicle 10 is viewed obliquely as shown in FIG. Represents.

Here, the panoramic image estimation unit 93 rotates the panoramic image D1 and estimates an image viewed from the side surface of the vehicle 10.
FIG. 5 is a diagram showing an example of the panoramic image D2 of the vehicle after the rotation processing in the present embodiment. By performing the process of rotating the panoramic image D1, the panoramic image estimating unit 93 estimates the panoramic image D2 of the vehicle 10 viewed from the same direction as the radiation position of the laser scanner 102.

The panoramic image estimation unit 93 measures the vehicle length of the vehicle 10 from the panoramic image D2 and sends it to the image generation unit 92. Further, the panoramic image estimation unit 93 sends the estimated panoramic image D2 to the interpolation unit 94. The panoramic image estimation unit 93 may employ a configuration that further measures the vehicle width and vehicle height of the vehicle 10 from the panoramic image D1 and the panoramic image D2.

Returning to FIG. 2, the image generation unit 92 generates the point cloud data g from the detection point distance acquired from the acquisition unit 91. The point cloud data g is an image at a predetermined position T, in which the densities of the detection points are changed according to the detection point distance of the detection points.
FIG. 6 is a diagram showing an example of the point cloud data g generated from the detection point distance. In FIG. 6, the vertical axis represents the scanning angle when the laser scans each detection point p of the vehicle 10, and the horizontal axis represents the time when the laser scans each detection point p of the vehicle 10. As shown in FIG. 6, in the point cloud data g, a detection point p having a shorter detection point distance is displayed in a lighter color, and a detection point p having a longer detection point distance is displayed in a darker color.

The image generation unit 92 also generates a reflectance image G from the reflection intensity acquired from the acquisition unit 91. The reflectance image G is an image at a predetermined position T, and is an image in which the density of the detection point p is changed according to the reflection intensity of the laser reflected at the detection point p.
FIG. 7 is a diagram showing an example of the reflectance image G in this embodiment. In FIG. 7, the vertical axis represents the scanning angle when the laser scans each detection point p of the vehicle 10, and the horizontal axis represents the time when the laser scans each detection point p of the vehicle 10. As shown in FIG. 7, in the reflectance image G, the detection point p having a higher laser reflection intensity is displayed in a lighter color, and the detection point p having a lower reflection intensity is displayed in a darker color.

Since the distance and the reflection intensity are obtained in a one-to-one correspondence with each detection point p, the point cloud data g and each detection point p displayed in the reflectance image G have a one-to-one correspondence. Become involved. That is, the point cloud data g and the reflectance image G are images including the same vehicle 10.

A characteristic of the image obtained by the laser scanner 102 is that the detection points p are densely detected in the scanning direction of the laser scanner 102, but the detection points p are sparsely detected in the time axis direction. .. That is, in the point cloud data g and the reflectance image G, the detection points p are densely displayed in the vehicle height direction of the vehicle 10, but the detection points p are periodically missing in the vehicle length direction of the vehicle 10. ing.

Here, in the vehicle type determination process in the vehicle type determination device 103 of the present embodiment, the vehicle type is determined using the number of axles of the vehicle 10 measured from the image and the distance between the axles. Therefore, in the vehicle type identification device 103 of the present embodiment, the point group data g and the reflectance image G are used to perform image interpolation processing of the image captured by the camera 101 to generate a high-resolution image. Improve the accuracy of measuring the number of cars and the distance between axles.

In order to perform the image interpolation process, it is necessary to perform a matching process in which an image captured by the camera 101 is collated with an image obtained by the laser scanner 102, corresponding points are detected, and alignment is performed. Here, the image captured by the camera 101 has a higher affinity for the reflectance image G than for the point cloud data g. Therefore, using the reflectance image G for matching processing with the image captured by the camera 101 makes it easier to detect corresponding points than using the point cloud data g. For the above reason, in the present embodiment, the reflectance image G is used for the matching process with the image captured by the camera 101.

Here, the image obtained by the laser scanner 102 has the property of expanding and contracting in the time axis direction depending on the speed of the object. That is, as the passing speed of the vehicle 10 traveling in front of the laser scanner 102 increases, the number of detection points p in the time axis direction that can be detected by the laser scanner 102 decreases, and the horizontal length of the generated reflectance image G is increased. Becomes shorter.

If the aspect ratio of the reflectance image G remains changed depending on the speed of the vehicle 10, it is difficult to perform matching processing with the image captured by the camera 101. Therefore, the image generation unit 92 performs a normalization process on the reflectance image G so that the aspect ratio of the reflectance image G matches the actual aspect ratio of the vehicle 10.

8 and 9 are diagrams for explaining an example of the normalization process of the vehicle image included in the reflectance image G in the vehicle type identification device 103 of the present embodiment. The vehicle length 1 is the vehicle length of the vehicle 10 in the reflectance image G before the normalization processing, and the vehicle length L is the vehicle length of the vehicle 10 in the reflectance image G after the normalization processing. The vehicle 10 in FIG. 8 has passed the predetermined position T at a speed V1, and the vehicle 10 in FIG. 9 has passed the predetermined position T at a speed V2 faster than the speed V1.

8 and 9 are reflectance images G including images of the vehicle 10 having the same vehicle length. However, the vehicle length l of the image of the vehicle 10 included in the reflectance image G when the vehicle 10 passes at the speed V2 (see FIG. 6) is the reflectance image G when the vehicle 10 passes at the speed V1 (see FIG. 5). The vehicle length is shorter than the vehicle length 1 of the image of the vehicle 10 included in (see). That is, the vehicle length 1 of the image of the vehicle 10 included in the reflectance image G becomes shorter as the speed of the vehicle 10 passing through the predetermined position T becomes faster. Therefore, as shown in FIGS. 8 and 9, the image generation unit 92 performs a process of normalizing the vehicle length l of the vehicle image included in the reflectance image G to the vehicle length L.

Next, a specific example of the process of normalizing the reflectance image G will be described with reference to FIG. FIG. 10 is a diagram for specifically explaining an example of the normalization process of the image of the vehicle 10 included in the reflectance image G in the vehicle type identification device 103 of the present embodiment.

As shown in FIGS. 10A and 10B, the image generation unit 92 first identifies the vehicle length 1 of the vehicle image included in the reflectance image G. Further, as shown in FIG. 10C, the image generation unit 92 detects the front axle FS and the rear axle BS included in the reflectance image G. Next, the image generator 92 measures the width of the front axle FS and the width of the rear axle BS. The width of the front axle FS can be acquired by measuring the distance from the left end to the right end of the detected front axle FS. The same applies to the method of measuring the width of the rear axle BS. The image generation unit 92 can detect the front axle FS and the rear axle BS by the same method as the axle detection by the axle detection unit 95. A method of axle detection by the axle detector 95 will be described later.

Further, as shown in FIG. 10D, the image generation unit 92 calculates the ratio of the width of the front axle FS and the width of the rear axle BS as the speed change rate when the vehicle 10 passes the predetermined position T. .. Then, when the rate of change in speed is “1” and there is a high possibility that the vehicle 10 passes through the predetermined position T at a constant speed, as shown in FIG. The reflectance image G is uniformly expanded and contracted in the vehicle length direction of the vehicle image, and the vehicle length of the vehicle image included in the reflectance image G is normalized to a predetermined length L.

Here, in the present embodiment, the image generation unit 92 uses the vehicle length L of the panoramic image D2 estimated (generated) by the panoramic image estimation unit 93 from the image captured by the camera 101 as the predetermined length L.

On the other hand, when the speed change rate is not “1”, the image generation unit 92 is likely to have accelerated or decelerated when the vehicle 10 passed the predetermined position T. Therefore, the portion of the reflectance image G indicates that the vehicle of the vehicle image Change the stretch ratio in the long direction. For example, when the rate of change in speed is smaller than “1” and there is a high possibility that the vehicle 10 has decelerated when passing the predetermined position T, the image generation unit 92 moves toward the vehicle length direction of the vehicle image, The distance between the detection points p of the reflectance image G is made smaller toward the rear end. On the other hand, when the rate of change in speed is greater than “1” and there is a high possibility that the vehicle 10 has accelerated when passing through the predetermined position T, the image generation unit 92 moves toward the vehicle length direction of the vehicle image, The distance between the detection points p of the reflectance image G is increased toward the rear end.

The image generation unit 92 sends the point cloud data g and the normalized reflectance image G to the interpolation unit 94.

Returning to FIG. 2, the interpolation unit 94 performs a matching process of matching the feature point P of the panoramic image D2 with the detection point p of the normalized reflectance image G. Then, the interpolation unit 94 interpolates the feature point P of the panoramic image D2 based on the result of the matching process and the detection point p included in the point group data g generated from the detection point distance.

FIG. 11 is a diagram for explaining an example of image interpolation processing in the vehicle type identification device 103 of the present embodiment. As shown in FIG. 11A, the interpolation unit 94 executes a matching process of matching the feature point P included in the panoramic image D2 with the detection point p of the reflectance image G. Next, as shown in FIG. 11B, the interpolation unit 94 interpolates between the feature points P collated with the detection points p in the panoramic image D2 by the point group data g existing between the feature points P. Interpolation processing is executed to obtain a panoramic image A having dense feature points P.

The matching process performed by the interpolation unit 94 will be specifically described.
FIG. 12 is a diagram for explaining an example of matching processing in the vehicle type identification device 103 of the present embodiment. As illustrated in FIG. 12, the interpolation unit 94 identifies the detection point p of the normalized reflectance image G that matches the feature point P of the panoramic image D2.

Next, the image interpolation processing performed by the interpolation unit 94 will be specifically described.
FIG. 13 is a diagram for explaining an example of image interpolation processing in the vehicle type identification device 103 of the present embodiment. As described in FIGS. 6 and 7, the point cloud data g has the detection points p densely in the scanning direction of the vehicle 10. However, in the point cloud data g, the detection points p are sparse in the time axis direction of the vehicle 10.

On the other hand, as described with reference to FIGS. 3 to 5, the panoramic image D2 of the vehicle 10 is estimated based on the captured image obtained by capturing the image with the camera 101. Therefore, the panoramic image D2 is likely to be affected by the subject (vehicle 10), the imaging conditions when capturing the subject, and the like. In general, the characteristic points P included in the panoramic image D2 have a property of being sparser than the detection points p included in the point cloud data g.

As described with reference to FIG. 7, since the detection point p of the reflectance image G and the detection point p of the point cloud data g have a one-to-one correspondence, the point cloud data g that matches the feature point P of the panoramic image D2. The detection point p of is specified by the result of the matching processing described in FIG.

As shown in FIG. 13, the interpolation unit 94 interpolates (expands) the feature points P that match (match) the detection points p with the detection points p existing between the feature points P in the panoramic image D2. As a result, as shown in FIG. 13, the interpolation unit 94 can obtain the panoramic image A having dense feature points P (in other words, the panoramic image A in which the feature points P are expanded).

The interpolation unit 94 sends the panoramic image A to the axle detection unit 95.

Returning to FIG. 2, the axle detection unit 95 detects axles from the panoramic image A in which the characteristic points P are interpolated by the point group data g, and measures the number of axles and the distance between the axles.

FIG. 14 is a diagram for explaining an example of axle detection processing in the vehicle type identification device 103 of the present embodiment. The semi-trailer panoramic image A1 that has passed the predetermined position T shown in FIG. 14 is an example of the interpolated panoramic image A.

In FIG. 14, the lower limit in the height direction of the range in which the axle detection unit 95 detects the axle is the road surface 300. The upper limit position 301 in the height direction of the range in which the axle detection unit 95 detects the axle is the height of the bottom surface of the vehicle body of the vehicle 10 included in the panoramic image A1.

It is desirable that the range in which the axle detection unit 95 performs the axle detection be set so that the structures other than the axles 11a to 11d are included as little as possible. For example, the upper limit position 301 in the height direction of the range in which the axle detection is performed may be the value of the minimum ground altitude of the vehicle type that is assumed to travel at the predetermined position T. Alternatively, the height from the road surface to the upper limit position 301 in the height direction of the range where the axle detection is performed may be a value that is ½ of the assumed diameter of the axles 11a to 11d. The storage unit 90 may store the assumed minimum ground altitude of the vehicle type and the assumed diameters of the axles 11a to 11d.

The axle detection unit 95 detects an area in which the frequency of occurrence of the characteristic points P is higher than a predetermined threshold value in the range in which the axle detection is performed in the panoramic image A1. In the region where the axles 11a to 11d exist, the frequency of occurrence of the feature point P is higher than that in the space portion where the axles 11a to 11d do not exist. Therefore, the axle detection unit 95 can detect the positions of the axles 11a to 11d by detecting an area in which the occurrence frequency of the characteristic points P is higher than a predetermined threshold value. In this way, the axle detection unit 95 detects the positions of the axles 11a to 11d included in the panoramic image A1 shown in FIG. The storage unit 90 stores a predetermined threshold value of the frequency of occurrence of the feature points P.

The method of detecting the axles 11a to 11d by the axle detector 95 is not limited to this. For example, since it is assumed that the axles 11a to 11d are generally circular and black, the axle detection unit 95 matches the circular and black image pattern (template) from the panoramic image A1 by pattern matching (template matching). You may employ the structure which detects an area|region as axle 11a-11d. In this case, the storage unit 90 may store the circular black image pattern (template).

Further, as described with reference to FIGS. 8 and 9, the reflectance image G after the normalization process shows that the vehicle 10 in the vehicle length direction (time axis direction) increases as the speed of the vehicle 10 passing through the predetermined position T increases. The number of the detection points p to the points is reduced. That is, with respect to the vehicle 10 having a high traveling speed, the horizontal resolution of the reflectance image G may be insufficient, and the image may be expanded in the horizontal direction by the normalization process, resulting in a large error in the position of the detection point p. .. Therefore, the axle detection unit 95 sets, for example, a lower limit to the horizontal resolution of the reflectance image G before the normalization process, and when the lower limit is not reached, the axle detection unit 95 detects the axle from the uninterpolated panoramic image D2. Alternatively, a configuration for measuring the number of axles and the distance between the axles may be adopted. The lower limit of the horizontal resolution of the reflectance image G may be based on, for example, the case where the width of the axle in the image of the vehicle 10 included in the reflectance image G before the normalization process is equal to or less than a certain value.

The axle detection method shown in FIG. 14 can also be applied to the axle detection in the normalization process of the reflectance image G. That is, the image generation unit 92 may employ a configuration in which the axle detection unit 95 performs axle processing from the panoramic image A and detects the axle from the reflectance image G.

Further, FIG. 15 is a diagram for explaining an example of a method of measuring the size of the axle and the center position of the axle in the vehicle type identification device 103 of the present embodiment. As shown in FIG. 15A, the axle detection unit 95 detects the axle 310 from the panoramic image A. FIG. 15B is an enlarged view of a region including the axle 310. The axle detection unit 95 can acquire the diameter (size) of the axle 310 by measuring the distance from the road surface 312 to the highest position 311 of the axle 310. In addition, the axle detection unit 95 measures an intermediate position 313 between the road surface 312 and the highest position 311 of the axle 310 from the road surface, and measures the distance from the road surface 312 to the intermediate position 313, thereby measuring the axle 310. You can get the radius.

Further, the distance from the left end 314 to the right end 315 of the axle 310 indicates the width of the axle 310. An intermediate position 316 between the left end 314 and the right end 315 of the axle 310 is the center of the axle 310.

As shown in FIG. 14, the axle detection unit 95 measures the distance between the axles 11a to 11d after detecting the positions of the axles 11a to 11d. In FIG. 14, reference symbol L1 is the distance from the center of the first axle 11a to the center of the second axle 11b, reference symbol L2 is the distance from the center of the second axle 11b to the center of the third axle 11c, and reference symbol L3 is The distances from the center of the third axle 11c to the center of the fourth axle 11d are shown respectively. The axle detection unit 95 converts the distances L1 to L3 between the axles 11a to 11d on the panoramic image A1 into the actual size of the vehicle 10 and measures the actual distance between the axles of the vehicle 10.

Here, as a matter that needs to be taken into consideration when detecting the axle, a case where the vehicle 10 having the lift axle function passes by will be described with reference to FIGS. 16 and 17. The lift axle function refers to a function of lifting (lifting) a part of the axles from the road surface so as not to touch the ground when the load is light, for example, on a trailer. An axle that floats from the road surface is not an axle according to the vehicle type classification and vehicle type determination criteria. Therefore, the lift axle function is often used to reduce the number of axles for the purpose of lowering the vehicle classification from oversized vehicles to large vehicles to save highway tolls.

Specifically, as shown in FIG. 16, there are cases where one axle is floated from the road surface, and where two axles are floated from the road surface as shown in FIG. There is a space from the road surface 400 shown in FIGS. 16 and 17 to a lower limit 401 of the height of the lifted axle. Since there is a difference in the distribution of the feature points P in the panoramic image A1 between the axle portion and the space portion, when the axle detection unit 95 recognizes a space below the axle from the distribution of the feature points P in the panoramic image A1, the axle is Judge that it is not an axle. The method of determining the lift of the axle is not limited to this, and the axle detection unit 95 may identify the space under the axle based on the detection point distance of the detection point p measured by the laser scanner 102. As a result of the pattern matching, the axle detection unit 95 may recognize that the position of the circular image estimated to be the axle is distant from the road surface 400 and may determine the lift of the axle.

When the axle detection unit 95 determines that the axle is floating above the road surface, the axle detection unit 95 measures the number of axles and the distance between the axles excluding the axles floating above the road surface.

The axle detection unit 95 sends the measured number of axles of the vehicle 10 and the distance between the axles to the determination unit 96.

Returning to FIG. 2, the determination unit 96 determines the vehicle type classification and vehicle type of the vehicle 10 based on the number of axles and the distance between the axles acquired from the axle detection unit 95. Here, the process in which the determination unit 96 determines the vehicle type classification and the vehicle type is referred to as a vehicle type determination process.

Here, the vehicle type classification refers to a classification for classifying vehicle types. In the present embodiment, the vehicle type classification adopts a configuration in which the vehicle type is classified into a light vehicle, a standard vehicle, a medium-sized vehicle, a large vehicle, and an oversized vehicle in accordance with the vehicle type classification used in the highway toll standard.

The determination unit 96 executes the vehicle type determination process according to the standard defined in the vehicle type determination information 20. The vehicle type identification information 20 is information that defines a vehicle type classification and a standard for identifying a vehicle type based on the number of axles and the distance between the axles.

For example, the determination unit 96 determines whether the vehicle 10 has three or more axles according to the criteria defined in the vehicle type identification information 20 and the distance between the axles of the vehicle 10 is two or more. When it is determined that there are more than one location, the vehicle 10 is determined to be an oversized vehicle. In addition, when the determination unit 96 determines that the number of axles of the vehicle 10 is five or more, the determination unit 96 determines that the vehicle 10 is an oversized vehicle.

Details of the criteria defined in the vehicle type identification information 20 will be described with reference to FIG. FIG. 18 is a diagram showing an example of the vehicle type identification information 20 in the vehicle type identification device 103 of this embodiment. In the present embodiment, the vehicle type identification information 20 is stored in the storage unit 90. Alternatively, the vehicle type identification information 20 may be stored in an external storage device (not shown) connected to the vehicle type identification device 103 via a network.

As shown in FIG. 18, the vehicle type identification information 20 includes at least information of “number of axles”, “inter-axle distance condition (m)”, “vehicle type classification”, and “vehicle type”. Note that the item “image diagram” is a model of the position of the axle based on the inter-axle distance condition and is displayed, and is not essential as a configuration of the vehicle type identification information 20.

The item “number of axles” of the vehicle type identification information 20 indicates the number of axles. According to the vehicle type identification information 20 in FIG. 18, the number of axles is classified into four, that is, two axes, three axes, four axes, and five or more axes. In the present embodiment, if the number of axles of the vehicle 10 is two, the determination unit 96 determines that the vehicle type of the vehicle 10 is “light vehicle, ordinary vehicle, medium-sized vehicle” or “medium-sized vehicle” depending on the distance between the axles. , Large-sized vehicle”. If the number of axles of the vehicle 10 is three, the determining unit 96 determines that the vehicle type classification of the vehicle 10 is either “extra large vehicle” or “medium-sized vehicle or large vehicle” depending on the combination of the distances between the axles. Determine. If the number of axles of the vehicle 10 is four, the determination unit 96 determines that the vehicle type classification of the vehicle 10 is either “large vehicle” or “large vehicle” depending on the combination of the distances between the axles. When the number of axles of the vehicle 10 is five or more, the determination unit 96 determines that the vehicle type classification of the vehicle 10 is “oversized vehicle” regardless of the distance between the axles.

The “number of axes” of the vehicle type identification information 20 shown in FIG. 18 is an example, and the present invention is not limited to this.

The item “inter-axle distance condition (m)” of the vehicle type identification information 20 indicates a threshold value of the distance between the axles in units of meters. When the number of axles is three or more, the combination of the distances between the axles is a condition for determining the vehicle type classification and the vehicle type. The items L1 to L3 of the “inter-axle distance condition (m)” of the vehicle type identification information 20 shown in FIG. 18 measure the distance between two adjacent axles of the vehicle 10 from the center of each axle and convert the actual dimensions. Indicates the length. Specifically, as in FIG. 14, the item L1 in FIG. 18 is the distance between the frontmost axle (first axle) of the vehicle 10 and the second axle (second axle) from the front side. Indicates. Item L2 indicates the distance between the second axle of the vehicle 10 and the third axle (third axle) from the front side. Item L3 indicates the distance between the third axle of the vehicle 10 and the fourth axle (fourth axle) from the front side.

The items L1 to L3 of the “inter-axle distance condition (m)” of the vehicle type discrimination information 20 are AND conditions, and it is considered that the conditions are met only when all combinations of distances are satisfied. The values n1 to n3 set in the items L1 to L3 are examples of predetermined threshold values for the distance between the axles. Further, the magnitude relationship between the values n1 to n3 is n1<n2<n3.

For example, No. of the vehicle type identification information 20 of FIG. 6, the number of axles is four, and L1 of the “inter-axle distance condition (m)” is “n2≦L1<n3”, L2 is “n2≦L2<n3”, and L3 is “0<L3<. n2”. That is, the distance between the first axle and the second axle of the vehicle 10 is n2 meters or more and less than n3 meters, and the distance between the second axle and the third axle is n2 meters or more and less than n3 meters. If the distance between the third axle and the fourth axle is longer than 0 m and less than n2 meters, the determination unit 96 determines that the vehicle 10 is No. It is determined that the “inter-axle distance condition (m) of 6” is satisfied.

Further, the number of axles of the vehicle 10 is four, and the number of the vehicle type identification information 20 of FIG. When none of the "inter-axle distance conditions (m)" of 6 to 9 is satisfied, No. As shown in FIG. 10, the “vehicle type classification” of the vehicle 10 is a large vehicle, and the “vehicle type” is determined to be an “other four-axle vehicle” other than an ordinary truck, a motorcycle, and a semi-trailer.

The item “vehicle type classification” of the vehicle type identification information 20 indicates a vehicle type classification of the vehicle 10 specified from the “number of axles” and the “inter-axle distance condition (m)”.

The item “vehicle type” of the vehicle type identification information 20 indicates the vehicle type of the vehicle 10 specified from the “number of axles” and the “inter-axle distance condition (m)”. For example, No. of the vehicle type discrimination information 20 shown in FIG. As defined in 1, when the number of axles is two and the distance L1 between the axles is n1 meters or less, the vehicle type classification is one of a light vehicle, a standard vehicle, and a medium-sized vehicle. There are passenger cars and rough terrain cranes. In addition, No. of the vehicle type identification information 20. As defined in 2, if the number of axles is two and the distance L1 between the axles is longer than n1 meters, the vehicle type classification is a medium-sized vehicle or a large vehicle, and the assumed vehicle type is a single vehicle. , Ordinary trucks, buses, tractors. Here, the single vehicle is a general term for a freight vehicle having a structure that cannot be separated into a towed portion and a towed portion.

The vehicle type shown in FIG. 18 is an example, and the vehicle type identification information 20 may be configured to store a larger number of vehicle types.

Here, the content of the vehicle type discrimination processing will be specifically described with reference to FIGS. 14 and 18. As described in FIG. 14, the axle detection unit 95 detects the axles 11a to 11d from the panoramic image A1 and measures the distances L1 to L3 between the axles. The determination unit 96 acquires the detection results of the axles 11a to 11d from the axle detection unit 95, searches the vehicle type determination information 20 illustrated in FIG. 18, and determines the “number of axles” and the “inter-axle distance condition (m) to which the vehicle 10 applies. )” is acquired. It is assumed that the number of axles 11a to 11d of the vehicle 10 shown in FIG. 14 is four and the distances between the axles are n2≦L1<n3, n2≦L2<n3, and 0<L3<n2. In this case, the panoramic image A1 shown in FIG. The condition of the combination of “number of axles” and “inter-axle distance condition (m)” shown in 6 is satisfied. Therefore, the determination unit 96 determines that the “vehicle type classification” of the panoramic image A1 is an oversized vehicle and the “vehicle type” is a single vehicle or a semi-trailer.

On the other hand, when the vehicle 10 shown in the panoramic image A1 separates the towed vehicle that is the rear load-carrying portion and independently travels only by the front towed vehicle, the number of axles of the vehicle 10 is 2 axles 11a and 11b. In this case, the determination unit 96 searches the vehicle type determination information 20 and determines that “the number of axles” is applicable to the vehicle 10 in which the “number of axles” is two axes and the distance L1 between the axles is “n2≦L1<n3”. A combination of "inter-axle distance condition (m)" is acquired. In this case, the vehicle 10 has the No. of the vehicle type identification information 20. The condition of the combination of “number of axles” and “inter-axle distance condition (m)” shown in 2 is satisfied. Therefore, the determination unit 96 determines that the “vehicle type” is a medium-sized vehicle or a large vehicle, and the “vehicle type” is a single vehicle, an ordinary truck, a bus, or a tractor.

Further, according to the vehicle type identification information 20 shown in FIG. 18, the vehicle type classification is the oversized vehicle when the vehicle 10 is No. 3, No. 6, No. 7, No. This is the case where any one of the definitions of 11 is satisfied. That is, as described above, in the determination unit 96, the vehicle 10 of the vehicle type determination information 20 has three or more axles, and there are two or more locations where the distance between the axles of the vehicle 10 is n2 meters or more. If it is determined that the number of axles of the vehicle 10 is five or more, the vehicle 10 is determined to be an oversized vehicle.

The determination unit 96 outputs the result of the vehicle type process to an external device (not shown). Alternatively, the determination unit 96 may employ a configuration in which the result of the vehicle type determination process is displayed on a display unit (not shown) of the vehicle type determination device 103. Further, when the discriminating unit 96 discriminates the vehicle 10 of a specific vehicle type classification or vehicle type, a notifying unit (not shown) of the vehicle type discriminating apparatus 103 may be used to notify.

As described above, in the vehicle type identification device 103 according to the present embodiment, the determination unit 96 determines the vehicle type based on the number of axles and the distance between the axles, and therefore the shapes of other parts of the vehicle 10 influence the vehicle type identification. do not do. For this reason, the interpolation unit 94 may adopt a configuration in which, instead of uniformly interpolating the entire panoramic image D2, a specific range of the panoramic image D2 is interpolated so as to improve the resolution around the axle. For example, the interpolation unit 94 may employ a configuration that interpolates a range from the road surface on the panoramic image D2 to the maximum value of the assumed axle height. In this case, the configuration may be such that the storage unit 90 stores the maximum value of the assumed axle height. Alternatively, the interpolation unit 94 may adopt a configuration in which the axle is detected from the panoramic image D2 by pattern matching before the interpolation, and the range near the axle is interpolated. By limiting the interpolation range, the load of the matching process and the image interpolation process performed by the interpolation unit 94 can be reduced as compared with the case of interpolating the entire panoramic image D2.

Next, the flow of the vehicle type discrimination processing of the present embodiment configured as described above will be described. FIG. 19 is a flowchart showing an example of the procedure of the vehicle type identification process in the vehicle type identification device 103 of this embodiment. The process of this flowchart starts, for example, when the laser scanner 102 detects that the vehicle 10 has entered the predetermined position T.

In S10, the acquisition unit 91 acquires a captured image from the camera 101. The acquisition unit 91 sends the acquired captured image to the panoramic image estimation unit 93.

In S11, the panoramic image estimation unit 93 extracts the feature point P of the vehicle image included in the captured image acquired by the acquisition unit 91. Furthermore, the panoramic image estimation unit 93 estimates the panoramic image D1 of the vehicle 10 by specifying the positions of the extracted feature points P based on the positional relationship between the vehicle 10 and the camera 101.

In S12, the panoramic image estimation unit 93 estimates the panoramic image D2, which is an image in the lateral direction of the vehicle 10, by rotating the panoramic image D1. The panoramic image estimation unit 93 sends the panoramic image D2 to the interpolation unit 94. Further, the panoramic image estimation unit 93 sends the vehicle length L of the vehicle 10 measured from the panoramic image D2 to the image generation unit 92.

In S13, the acquisition unit 91 acquires the detection point distance and the reflection intensity of each detection point p from the laser scanner 102. The acquisition unit 91 sends the acquired detection point distance and reflection intensity of each detection point p to the image generation unit 92.

In S14, the image generation unit 92 determines whether the density of the detection point p of the vehicle 10 is different according to the detection point distance of each detection point p based on the detection point distance of each detection point p acquired by the acquisition unit 91. The point cloud data g, which is the generated image, is generated. In addition, the image generation unit 92 changes the density of the detection point p of the vehicle 10 based on the reflection intensity of each detection point p acquired by the acquisition unit 91 according to the reflection intensity of the detection point p. Of the reflectance image G is generated.

In S15, the image generation unit 92 performs a process of normalizing the reflectance image G using the vehicle length L acquired from the panoramic image estimation unit 93. The image generation unit 92 sends the point cloud data g and the normalized reflectance image G to the interpolation unit 94.

In S16, the interpolation unit 94 executes a matching process of matching the feature point P of the panoramic image D2 with the detection point p of the normalized reflectance image G. The interpolation unit 94 performs an image interpolation process of interpolating the feature points P of the panoramic image D2 using the detection points p included in the point group data g based on the result of the matching process. The interpolation unit 94 generates the panoramic image A whose resolution is improved by the image interpolation processing. The interpolation unit 94 sends the panoramic image A to the axle detection unit 95.

In S17, the axle detection unit 95 detects the number and position of axles from the panoramic image A. Further, the axle detection unit 95 measures the distance between the axles converted into the actual size of the vehicle 10. The axle detection unit 95 sends the measurement result of the number of axles and the distance between the axles to the determination unit 96.

In S18, the determination unit 96 refers to the vehicle type determination information 20 stored in the storage unit 90, and identifies the vehicle type classification and vehicle type whose conditions match the number of axles and the distance between the axles acquired from the axle detection unit 95. Car type discrimination processing is performed.

In S19, the determination unit 96 outputs the result of the vehicle type determination process to an external device (not shown).

As described above, according to the vehicle type identification device 103 of the present embodiment, the identification unit 96 performs the vehicle type identification process based on the number of axles and the distance between the axles. Therefore, the vehicle type identification is performed based on the shape of the vehicle body or the like. Compared with the case where it is performed, it is less likely to be affected by the change in the vehicle body shape due to aging, and a more stable result of the vehicle type determination processing can be obtained. Further, since the difference in the shape of the vehicle 10 due to the difference in the load is not affected, the vehicle type identification device 103 according to the present embodiment can improve the accuracy of the vehicle type identification process.

Further, depending on the color of the vehicle body of the vehicle 10, it may be difficult to scan with the laser scanner 102. Particularly for a black vehicle body, the laser light output from the laser scanner 102 is mirror-reflected, so that the light returns to the light receiving side. The distance measurement value may not be acquired correctly. On the other hand, the diffused reflection of the tire (axle) portion allows the laser scanner 102 to measure the distance relatively stably. According to the vehicle type identification device 103 according to the present embodiment, the vehicle type identification process is performed based on the number of axles and the distance between the axles, and thus the influence of the specular reflection of the vehicle body of the vehicle 10 is small and more stable. Vehicle type discrimination processing can be performed.

Further, according to the vehicle type identification device 103 according to the present embodiment, the acquisition unit 91 acquires the detection point distance from the laser scanner 102 to the detection point p and the reflection intensity of the laser at the detection point p, and the acquisition unit 91 outputs from the camera 101. A captured image of the predetermined position T is acquired. Further, the image generation unit 92 generates the reflectance image G, and the panoramic image estimation unit 93 estimates the panoramic image D1 and the panoramic image D2 from the captured image. Further, the interpolation unit 94 executes a matching process between the reflectance image G and the panoramic image D2, and interpolates the feature point P of the panoramic image D2 based on the result of the matching process and the detection point distance. The axle detection unit 95 detects an axle from the panoramic image A in which the feature points P are interpolated, measures the number of axles and the distance between the axles, and the determination unit 96 determines, based on the number of axles and the distance between the axles. The vehicle type of the vehicle 10 is determined.

Therefore, according to the vehicle type identification device 103 according to the present embodiment, it is generated based on only the detection point distance and the reflection intensity estimated based on only the image captured by the camera 101 or acquired by the laser scanner 102. The axle can be detected using a higher resolution image compared to the captured image. That is, the accuracy of the axle detection by the axle detection unit 95 and the measurement of the distance between the axles can be improved, and the determination unit 96 can perform the vehicle type determination processing with high accuracy.

Further, according to the vehicle type identification device 103 according to the present embodiment, the image generation unit 92 performs a normalization process on the reflectance image G so that the aspect ratio of the reflectance image G matches the actual aspect ratio of the vehicle 10. Give. Further, the panoramic image estimation unit 93 estimates the panoramic image D1 and the panoramic image D2 from a plurality of still images continuously captured by the camera 101. Then, the interpolation unit 94 performs a matching process of matching the feature points P of the panoramic image D2 with the detection points p of the normalized reflectance image G. Therefore, according to the vehicle type identification apparatus 103 according to the present embodiment, the positions of the characteristic points P of the panoramic image D2 and the detection points p of the point cloud data g can be linked with high accuracy, and the image interpolation can be performed with higher accuracy. Processing can be performed.

Further, according to the vehicle type identification device 103 according to the present embodiment, in the identification unit 96, the vehicle 10 has three or more axles, and the distance between the axles is two or more locations apart from each other by a predetermined distance or more. If it is determined that the vehicle 10 is an oversized vehicle, it is possible to efficiently identify the oversized vehicle.

Further, according to the vehicle type identification device 103 according to the present embodiment, the identification unit 96 performs the vehicle type identification process according to the condition of the number of axles defined in the vehicle type identification information 20 and the threshold value of the distance between the axles. Compared with the case where the vehicle type determination is performed based on the shape, the vehicle type determination process rarely results in the vehicle type determination being impossible. Further, the vehicle type identification information 20 is configured by the numerical data of the number of axles and the distance between the axles, the vehicle type classification, and the vehicle type, and therefore is stored in the storage unit 90 as compared with the case where the vehicle body shape of each vehicle type is stored. The amount of data can be kept small.

The axle detection unit 95 of the vehicle type identification device 103 according to the present embodiment may further adopt a configuration in which the turning radius of the vehicle 10 is determined based on the number of axles and the distance between the axles. In this case, the vehicle type identification system 1a of the present embodiment is installed near the entrance of a mountain road or the like where there is a sharp curve, and issues a warning from a notification unit (not shown) when a vehicle 10 that exceeds a predetermined turning radius enters. A configuration may be adopted. Alternatively, the vehicle type identification system 1a of the present embodiment may be configured to add a function of closing a gate (not shown) to stop the passage of the vehicle 10. The storage unit 90 may store the threshold value of the predetermined turning radius. With this configuration, according to the vehicle type identification system 1a of the present embodiment, it is possible to prevent the vehicle 10 that has a large turning radius and cannot turn a sharp curve from entering the mountain road or the like, thereby preventing a danger. can do.

(Embodiment 2)
In the vehicle type identification device 103 of the first embodiment, the reflectance image G is normalized based on the vehicle length of the vehicle 10 estimated from the image captured by the camera 101. However, in the vehicle type identification device 1103 of the present embodiment, the speedometer is used. The reflectance image G is normalized based on the speed of the vehicle 10 measured by 104.

FIG. 20 is a diagram showing an example of the configuration of the vehicle type identification system 1b of this embodiment. As shown in FIG. 20, the vehicle type identification system 1b of the present embodiment includes a camera 101, a laser scanner 102, a vehicle type identification device 1103, and at least one speedometer 104.

The camera 101 and the laser scanner 102 of this embodiment have the same configuration as that of the first embodiment described with reference to FIG.

The speedometer 104 measures the speed of the vehicle 10. For example, as shown in FIG. 20, the speedometer 104 measures the speed when the vehicle 10 enters the predetermined position T (reference numeral 10b) from the position immediately before the predetermined position T (reference numeral 10a). The speedometer 104 transmits the measured speed of the vehicle 10 to the vehicle type identification device 1103. The speedometer 104 is an example of a speed measuring unit.

The vehicle type identification device 1103 is an example of an electronic computer that executes vehicle type identification processing. The vehicle type identification device 1103 is electrically connected to the camera 101, the laser scanner 102, and the speedometer 104. The installation location of the vehicle type identification device 1103 is the same as the installation location of the vehicle type identification device 103 described in FIG.

FIG. 21 is a block diagram showing an example of the functional configuration of the vehicle type identification device 1103 of this embodiment. As shown in FIG. 21, the vehicle type identification device 1103 of this embodiment includes a storage unit 90, an acquisition unit 1091, an image generation unit 1092, a panoramic image estimation unit 93, an interpolation unit 94, and an axle detection unit 95. , And a determination unit 96.

The storage unit 90, the panoramic image estimation unit 93, the interpolation unit 94, the axle detection unit 95, and the determination unit 96 of this embodiment have the same configurations as those of the first embodiment described with reference to FIG.

The acquisition unit 1091 of the present embodiment acquires the speed of the vehicle 10 from the speedometer 104 in addition to the function of the acquisition unit 91 of the first embodiment. The acquisition unit 1091 sends the acquired speed to the image generation unit 1092.

The image generation unit 1092 acquires the speed of the vehicle 10 from the acquisition unit 1091. The image generation unit 1092 normalizes the reflectance image G based on the speed of the vehicle 10.

As described with reference to FIGS. 8 and 9, as the speed of the vehicle 10 increases, the vehicle length l of the image of the vehicle 10 included in the reflectance image G decreases. That is, by measuring the speed of the vehicle 10, the expansion/contraction ratio of the vehicle length l of the vehicle 10 included in the reflectance image G compared with the actual vehicle length of the vehicle 10 can be calculated.

The image generation unit 1092 calculates the expansion/contraction rate of the reflectance image G in the time axis direction from the speed of the vehicle 10, and corrects the length of the reflectance image G in the time axis direction based on the calculation result. As a result, the aspect ratio of the reflectance image G becomes the aspect ratio of the actual vehicle 10.

FIG. 22 is a diagram for explaining an example of normalization processing in the vehicle type identification device 1103 of this embodiment. FIG. 22 shows a model 500 of the vehicle length 1 and the axle position of the vehicle 10 in the reflectance image G before normalization, and a model 501 of the vehicle length L and the axle position of the vehicle 10 after normalization. The image generation unit 1092 calculates a scaling factor for making the vehicle length 1 the vehicle length L that matches the actual aspect ratio of the vehicle 10, from the expansion/contraction rate calculated from the speed of the vehicle 10. Then, as shown in models 500 to 501, the image generation unit 1092 expands the horizontal width of the reflectance image G before normalization in accordance with the magnification of the calculation result to perform the normalization of the reflectance image G.

Further, the vehicle type identification system 1b may employ a configuration including a plurality of speedometers 104 arranged at regular intervals. In this case, it is possible to measure that the vehicle 10 has accelerated or decelerated while passing through the predetermined position T. Therefore, the entire reflectance image G is not expanded at the same ratio as in FIG. 22, but is changed in speed. It is possible to change the degree of expansion for each certain area in accordance with. Further, in the normalization processing of the first embodiment, the speed change rate of the vehicle 10 is calculated from the change in the width of the axles of the front wheels and the rear wheels, but the vehicle type determination system 1b of the present embodiment is arranged at constant distance intervals. When a plurality of speedometers 104 are provided, the speed change rate can be calculated based on the measurement result of the speedometers 104.

The flow of the vehicle type determination process in the vehicle type determination device 1103 of the present embodiment configured as described above is the same as the flow of the vehicle type determination process of the first embodiment described in FIG. 19, but in the process of S13, the acquisition unit 1091 Further, the speed of the vehicle 10 is acquired from the speedometer 104. Further, in the process of S15, the image generation unit 1092 performs the normalization process of the reflectance image G based on the speed of the vehicle 10 acquired by the acquisition unit 1091.

Thus, according to the vehicle type identification device 1103 of the present embodiment, the acquisition unit 1091 acquires the speed of the vehicle 10 from the speedometer 104, and the image generation unit 1092 is based on the speed of the vehicle 10 acquired by the acquisition unit 1091. Since the reflectance image G is normalized, the reflectance image G can be expanded or contracted according to the actual speed of the vehicle 10. Therefore, according to the vehicle type identification device 1103 of the present embodiment, more accurate normalization processing can be performed, and the accuracy of the subsequent matching processing with the panoramic image D2 can be improved. That is, according to the vehicle type identification device 1103 of the present embodiment, by generating a higher-resolution panoramic image A, the accuracy of measurement of the number of axles and the distance between the axles is improved, and more accurate vehicle type identification is performed. It can be carried out.

(Embodiment 3)
Although the vehicle type identification device 103 of the first embodiment determines the vehicle type of the vehicle 10, the vehicle type identification device 2103 of the present embodiment further counts the vehicle type classification of the vehicle 10 or the number of passing vehicles for each vehicle type.

The vehicle type identification system 1c of this embodiment includes a camera 101, a laser scanner 102, and a vehicle type identification device 2103.

The camera 101 and the laser scanner 102 of this embodiment have the same configuration as that of the first embodiment described with reference to FIG.

The vehicle type identification device 2103 is an example of an electronic computer that executes vehicle type identification processing. The vehicle type identification device 2103 is electrically connected to the camera 101 and the laser scanner 102. The installation location of the vehicle type identification device 2103 is the same as the installation location of the vehicle type identification device 103 described in FIG.

FIG. 23 is a block diagram showing an example of the functional configuration of the vehicle type identification device 2103 of this embodiment. As shown in FIG. 23, the vehicle type identification device 2103 according to the present embodiment includes a storage unit 90, an acquisition unit 91, an image generation unit 92, a panoramic image estimation unit 93, an interpolation unit 94, and an axle detection unit 95. The discriminating unit 96 and the counter 97 are provided.

The storage unit 90, the acquisition unit 91, the image generation unit 92, the panoramic image estimation unit 93, the interpolation unit 94, the axle detection unit 95, and the determination unit 96 according to the present embodiment are the same as those described in FIG. The configuration is the same as that of 1.

The counter 97 acquires the vehicle type determination result of the determination unit 96 and counts the number of passing vehicles for each vehicle type classification or each vehicle type. The counter 97 may count the number of passing vehicles based on either the vehicle type classification or the vehicle type classification, or may count based on both the classifications. Alternatively, the counter 97 may be configured to count only the vehicles 10 of a specific vehicle type classification or vehicle type.

Further, the counter 97 associates the count result of each vehicle type or the vehicle type of the vehicle 10 with the time when each vehicle 10 is counted. The counter 97 may be configured to calculate a total value of the number of passing vehicles for each vehicle type classification or each vehicle type at predetermined time intervals. Alternatively, the counter 97 may employ a configuration in which the vehicle type classification or the vehicle type is associated with the time each time one vehicle 10 is counted.

The counter 97 outputs the count result to an external device (not shown). Alternatively, the counter 97 may be configured to display a count result on a display unit (not shown) of the vehicle type identification device 2103.

FIG. 24 is a diagram for explaining an example of a counting process for each vehicle type classification in the vehicle type determination device 2103 of the present embodiment. The axle detecting unit 95 measures the number of axles and the distance of the axles from the panoramic image A, and the discriminating unit 96 identifies the vehicle type category to which the vehicle 10 corresponding to the panoramic image A corresponds by the vehicle type discriminating process. The counter 97 acquires the determination result of the determination unit 96 and counts the number of passing vehicles for each vehicle type classification.

FIG. 25 is a graph showing an example of the counting result of the number of passing vehicles per hour by vehicle type classification in the vehicle type identification device 2103 of the present embodiment. In FIG. 25, the horizontal axis represents time and the vertical axis represents the number of passing vehicles. A solid line 600 in FIG. 25 indicates a change in the value of the number of passing vehicles that is the total of all vehicle types, and a broken line 601 indicates a change in the value of the passing vehicles of the vehicle 10 whose vehicle type classification is an oversized vehicle. In general, oversized vehicles such as lorries often travel in the middle of the night, and as shown in FIG. 25, the peaks of the total number of vehicles and the number of oversized vehicles may be in different time zones. is there. Even if the total number of passing vehicles is the same, the road may be more congested if more oversized vehicles pass. In the vehicle type discriminating apparatus 2103 of the present embodiment, the counter 97 counts the number of passing vehicles for each vehicle type classification, so that it is more accurate than when the road congestion situation is grasped only from the total number of passing vehicles of all vehicle types. It is possible to grasp the road congestion situation.

Next, the flow of the vehicle type discrimination processing of the present embodiment configured as described above will be described.
FIG. 26 is a flow chart showing an example of the procedure of the vehicle type identification process in the vehicle type identification device 2103 of this embodiment. The process from the acquisition of the image captured by the camera in S20 to the vehicle type determination process in S28 is performed in the same manner as in S10 to S18 of the vehicle type determination process of the first embodiment described in FIG.

When the determination unit 96 performs the vehicle type determination process in S28, the counter 97 acquires the vehicle type category or vehicle type determined by the determination unit 96 and counts the number of passing vehicles for each vehicle type category or vehicle type in S29. Further, the counter 97 associates the count result with the time at the time of counting.

In S30, the determination unit 96 outputs the result of the vehicle type determination process to an external device (not shown). The counter 97 also outputs the count result and the time at the time of counting to an external device (not shown).

As described above, according to the vehicle type identification device 2103 according to the present embodiment, the counter 97 counts the number of passing vehicles for each vehicle type classification or each vehicle type, so that the number of passing vehicles for each vehicle type classification can be measured. Therefore, according to the vehicle type identification device 2103 according to the present embodiment, the road congestion state can be more accurately grasped as compared with the case where the road congestion state is grasped only from the total number of passing vehicles of all vehicle types.

Further, according to the vehicle type identification device 2103 according to the present embodiment, the counter 97 associates the vehicle type classification of each vehicle 10 or the counting result for each vehicle type with the time when each vehicle 10 is counted, so that the vehicle type classification per hour is performed. It is possible to measure the number of passing vehicles for each vehicle type or vehicle type.

The vehicle type identification device 2103 of the present embodiment can provide the count result by the counter 97 and the time at the time of count as, for example, road traffic information. As described with reference to FIG. 25, by using the number of passing vehicles for each vehicle type classification or each vehicle type per hour, more accurate roads can be obtained as compared with the case where the road congestion situation is grasped from the total number of passing vehicles of all vehicle types. The congestion status can be provided as road traffic information. By installing the vehicle type identification system 1c of the present embodiment on a plurality of roads and counting the number of passing vehicles for each vehicle type classification or vehicle type per hour at each installation location, the accuracy of road traffic information is further improved, and Accurate traffic congestion can be predicted.

Further, the number of passing vehicles for each vehicle type classification per hour or for each vehicle type at the installation location of the vehicle type identification system 1c may be integrated with the map information. In this case, at each point corresponding to the installation location of the vehicle type identification system 1c on the map, it is possible to grasp the number of passing vehicles for each vehicle type category or each vehicle type per hour. For example, by providing the driver of the oversized vehicle with the information, the driver can know a more accurate road congestion situation and can drive while avoiding a congested road.

In addition, when the vehicle 10 of the oversized vehicle type is not traveling on the specific road at all, there is a possibility that the oversized vehicle cannot travel due to a narrow road width or the like. In this case, the driver of the oversized vehicle can avoid the road and travel by knowing the number of passing vehicles for each vehicle type category. That is, according to the vehicle type identification system 1c of the present embodiment, the driver can be assisted to select an appropriate route. The integration of the count result and the map information may be performed by an external device (not shown), or the vehicle type identification device 2103 may further acquire the map information and perform an integration process.

In addition, oversized vehicles generally have a higher load on the road than vehicles 10 of other vehicle categories. According to the vehicle type identification device 2103 of the present embodiment, it is possible to grasp the number of passing vehicles in a specific vehicle type section, and therefore it can be used for analysis of the load condition on the road. For example, the output result of the vehicle type identification device 2103 can be used for calculation of the maintenance time so that the maintenance of the road on which many oversized vehicles pass is performed more frequently than the road on which less oversized vehicles pass. ..

(Embodiment 4)
Although the vehicle type identification device 103 of the first embodiment determines the vehicle type of the vehicle 10, the vehicle type identification device 3103 of the present embodiment further measures the weight of the vehicle 10 by the weight measuring devices 105a to 105c.

FIG. 27: is a figure which shows an example of a structure of the vehicle type discrimination system 1d of this embodiment. As shown in FIG. 27, the vehicle type identification system 1d of the present embodiment includes a camera 101, a laser scanner 102, a vehicle type identification device 3103, and weight measuring devices 105a to 105c.

The camera 101 and the laser scanner 102 of this embodiment have the same configuration as that of the first embodiment described with reference to FIG.

In FIG. 27, the vehicle type identification system 1d includes three weight measuring devices 105a to 105c, but this is an example, and the vehicle type identification system 1d includes at least one weight measuring device 105. Good. Here, the weight measuring device 105 is a general term for the weight measuring devices 105a to 105c. The weight measuring device 105 is installed on a road (vehicle passage) at a predetermined position T, and measures the weight of the vehicle 10 when the axle of the vehicle 10 passes over the weight measuring device 105. The weight measuring device 105 may be a plate-like device installed on the road, or may be embedded in the road. The weight measuring device 105 transmits the measured weight of the vehicle 10 to the vehicle type identification device 3103. The weight measuring device 105 is an example of a weight measuring unit.

At least one weight measuring device 105 is preferably installed in front of the laser scanner 102 like the weight measuring device 105b in FIG. Further, since there is a change in the value of the measured weight due to the shaking of the vehicle 10, more accurate weight can be measured by measuring with a plurality of weight measuring devices 105a to 105c in FIG. When any of the weight measuring devices 105a to 105c fails, the measurement value of the weight measuring device 105a to 105c that is operating normally is adopted, and the configuration for continuing the operation of the vehicle type identification system 1d is adopted. May be.

The reason why it is desirable to install the weight measuring device 105 in the front position of the laser scanner 102 will be described with reference to FIGS. 28 and 29.

FIG. 28 is a diagram showing an example of a plurality of still images continuously captured by the camera 101 of the present embodiment. It is assumed that the camera 101 images the continuous images 1200 to 1203 at times t0 to t3, respectively. At this time, the time when the vehicle 10 enters and leaves the field of view of the camera 101 can be acquired, but it is difficult to detect the accurate position of the vehicle 10 at any time between times t0 and t3.

On the other hand, since the laser scanner 102 acquires the position information of the vehicle 10 in a shorter cycle than the camera 101, it is possible to acquire the time when each axle passes in front of the laser scanner 102.
FIG. 29 is a diagram showing an example of the detection result of the laser scanner 102 of this embodiment. As shown in FIG. 29, the point cloud data g generated from the distance information acquired by the laser scanner 102 scanning the vehicle 10 includes an axle 701. Further, as indicated by reference numeral 702, the weight measuring device 105b measures the weight of the vehicle 10 at time t1.

When the weight measuring device 105b is installed in the front position of the laser scanner 102 as shown in FIG. 27, the time t1 when the axle of the vehicle 10 passes in front of the laser scanner 102 and the weight measuring device 105b measures the weight of the axle. The time t1 is almost the same time. Therefore, the weight measuring instrument 105b measures the weight by synchronizing the time t1 when the laser scanner 102 scans the axle 701 shown in FIG. 29 with the time t1 when the weight measuring instrument 105b measures the weight of the axle 701. It can be specified that the axle 701 is the axle 701 scanned by the laser scanner 102. For the above reasons, it is desirable that at least one weight measuring device 105 is installed in a position in front of the laser scanner 102.

Returning to FIG. 27, the vehicle type identification device 3103 is an example of an electronic computer that executes vehicle type identification processing. The vehicle type identification device 3103 is electrically connected to the camera 101, the laser scanner 102, and the weight measuring devices 105a to 105c. The installation location of the vehicle type identification device 3103 is the same as the installation location of the vehicle type identification device 103 described in FIG.

FIG. 30 is a block diagram showing an example of the functional configuration of the vehicle type identification device 3103 of this embodiment. As shown in FIG. 30, the vehicle type identification device 3103 according to the present embodiment includes a storage unit 1090, an acquisition unit 2091, an image generation unit 92, a panoramic image estimation unit 93, an interpolation unit 94, and an axle detection unit 1095. A determination unit 1096, an axle weight measurement unit 98, and an overload determination unit 99 are provided.

The image generation unit 92, the panoramic image estimation unit 93, and the interpolation unit 94 of the present embodiment have the same configurations as those of the first embodiment described with reference to FIG.

In addition to the stored contents of the storage unit 90 of the first embodiment, the storage unit 1090 of the present embodiment further assumes an upper limit value of the weight of the loaded load for each vehicle type category or vehicle type, and the assumption of the vehicle body of the vehicle type category or vehicle type. The overload reference value, which is the weight obtained by adding the upper limit of the weight, is stored. In addition, the storage unit 1090 further stores a limit weight, which is an upper limit value of the weight per axle that can be safely supported, for each vehicle type classification or vehicle type. Different values may be defined for the limit weight depending on the size of the axle. The method for calculating the overload reference value and the limit weight described above is an example, and the present invention is not limited to this.

In addition to the function of the acquisition unit 91 of the first embodiment, the acquisition unit 2091 of the present embodiment further acquires the weight of the vehicle 10 and the time when the weight of the vehicle 10 was measured from the weight measuring devices 105a to 105c. The acquisition unit 2091 sends the acquired weight to the axle-by-axle weight measurement unit 98.

When detecting an axle from the panoramic image A interpolated by the point cloud data g, the axle detection unit 1095 associates each axle with the time when the detection point p corresponding to each axle was detected by the laser scanner 102. In addition to the number of axles and the distance between the axles, the axle detection unit 1095 sends the time when each axle was scanned by the laser scanner 102 to the axle-by-axle weight measurement unit 98. The method of associating the time when each axle is scanned with the detection result of each axle is not limited to this, and a configuration using the point cloud data g or the reflectance image G may be adopted.

The axle-by-axle weight measuring unit 98 acquires the number of axles, the distance between the axles, and the time when each axle is scanned by the laser scanner 102 from the axle detecting unit 1095. The axle-by-axle weight measuring unit 98 also acquires, from the acquiring unit 2091, the weight of the vehicle 10 acquired by the acquiring unit 2091 from the weight measuring devices 105a to 105c and the time at which the weight of the vehicle 10 was measured. The axle-by-axle weight measurement unit 98 includes the number of axles acquired from the axle detection unit 1095, the distance between the axles, the time when each axle is scanned by the laser scanner 102, the weight of the vehicle 10 acquired by the acquisition unit 2091, and the vehicle 10. The weight of each axle is measured by comparing with the time when the weight of the vehicle is measured.

Here, as described with reference to FIGS. 16 and 17, the vehicle 10 may cause some axles to float from the road surface by the lift axle function. There is a space from the road surface 400 shown in FIGS. 16 and 17 to a lower limit 401 of the height of the lifted axle. However, there is no specific standard for the distance from the road surface 400 to the lower limit 401 of the height of the lifted axle. Therefore, the distance from the road surface 400 to the lower limit 401 of the height of the lifted axle becomes extremely short, and it is judged from the shape information of the vehicle 10 such as the panoramic image A that the axle is floating above the road surface 400. Can be difficult. In this case, although the axle of part of the vehicle 10 is actually floating from the road surface 400, the axle detection unit 1095 may determine that the floating axle is also an axle that is in contact with the road surface 400.

Therefore, the axle-by-axle weight measuring unit 98 of the present embodiment determines whether or not each axle is floating from the road surface based on the measurement result of the weight of each axle of the vehicle 10. That is, even if the axle-by-axle weight measuring unit 98 determines that the axle detecting unit 1095 is an axle that is grounded on the road surface, if the weight measuring devices 105a to 105c do not measure the weight of the axle, it is actually Is determined to be an axle that floats from the road surface.

When it is determined that the axle is floating above the road surface, the axle-by-axle weight measuring unit 98 again measures the number of axles and the distance between the axles excluding the axles floating above the road surface.

The axle weight measuring unit 98 sends the measured number of axles of the vehicle 10 and the distance between the axles to the determining unit 1096. If there is no axle that is floating above the road surface, the axle weight measuring unit 98 sends the number of axles measured by the axle detecting unit 1095 and the distance between the axles to the determining unit 1096. The axle-by-axle weight measuring unit 98 also sends the weight of each axle of the vehicle 10 to the overload determining unit 99.

Returning to FIG. 30, the determination unit 1096 determines the vehicle type classification and vehicle type of the vehicle 10 based on the number of axles and the distance between the axles measured by the axle weight measurement unit 98. The determination unit 1096 performs the vehicle type determination process based on the criteria defined in the vehicle type determination information 20, similarly to the determination unit 96 of the first embodiment described in FIG. The determination unit 1096 sends the result of the vehicle type determination process to the overload determination unit 99.

Returning to FIG. 30, the overload determination unit 99 acquires the result of the vehicle type determination processing from the determination unit 1096. Further, the overload determination unit 99 acquires the weight of each axle of the vehicle 10 from the axle weight measurement unit 98. The overload determination unit 99 determines whether the vehicle 10 is overloaded based on the vehicle type determination result of the determination unit 1096 and the weight of the vehicle 10.

Specifically, the overload determination unit 99 acquires, from the storage unit 1090, an overload reference value that matches the vehicle type classification or the vehicle type of the vehicle type determination result of the determination unit 1096, and compares it with the weight of the vehicle 10. That is, when the overload determination unit 99 determines that the weight of the vehicle 10 acquired by the acquisition unit 2091 exceeds the overload reference value set for the vehicle type classification or vehicle type of the vehicle 10 determined by the determination unit 1096. First, it is determined that the vehicle 10 is overloaded. Alternatively, the overload determination unit 99 may adopt a configuration in which the vehicle 10 having a weight that is close to the overload reference value for a certain amount or more even if the overload reference value has not reached the overload reference value.

Further, the overload determination unit 99 measures the vehicle type measured by the axle weight measurement unit 98 based on the vehicle type determination result acquired from the determination unit 1096 and the weight for each axle of the vehicle 10 acquired from the axle weight measurement unit 98. It is determined whether the weight of each axle 10 has reached the limit weight, which is the upper limit of the weight of each axle.

Specifically, the overload determination unit 99 obtains from the storage unit 1090 the limit weight that matches the vehicle type classification or vehicle type of the vehicle type determination result of the determination unit 1096, and compares it with the weight of each axle of the vehicle 10. , It is determined whether or not the weight of each axle of the vehicle 10 has reached the limit weight. Further, the overload determination unit 99 may adopt a configuration to be a determination target even when the weight of each axle of the vehicle 10 does not reach the limit weight, when the difference in weight of each axle is significant. The threshold value of the weight difference between the axles may be stored in the storage unit 1090 as a predetermined weight difference.

Further, the overload determination unit 99 may acquire the size (diameter) of the axle measured by the axle detection unit 95 as described in FIG. In that case, the limit weight defined according to the size of the axle is acquired for each vehicle type classification or vehicle type from the storage unit 1090, and it is determined whether the weight for each axle reaches the limit weight.

The overload determination unit 99 outputs an overload determination result, a determination result of whether or not each axle has reached a limit weight, and a weight measurement result for each axle to an external device (not shown). Alternatively, the overload determination unit 99 employs a configuration that sends a signal to a not-shown notifying unit to issue an alarm when the vehicle 10 is overloaded or when one of the axles reaches the limit weight. You may do it. Further, the weight measurement result for each axle may be directly output by the axle weight measurement unit 98 to an external device (not shown).

FIG. 31 is a diagram showing an example of weight analysis for each axle in the vehicle type identification device 3103 of the present embodiment. As shown in FIG. 31, the panoramic image A of the vehicle 10 includes an axle 801 which is a front wheel of the vehicle 10 and an axle 802 which is a rear wheel of the vehicle 10. The lower part of FIG. 31 shows the weight measurement result for each axle by the axle-by-axle weight measurement unit 98. A rectangle 803 below the axle 801 shows the weight of the axle 801 and a rectangle 804 below the axle 802 shows the weight of the axle 802, respectively. When there is a difference in the measurement results by the weight measuring devices 105a to 105c, the weight value measured at the bottom of the rectangles 803 to 804 is the lightest, and the weight value measured at the upper side of the rectangles 803 to 804 is the heaviest. Show.

As shown in FIG. 31, the weight of the axle 801 corresponding to the front wheels of the vehicle 10 has not reached the limit weight. However, it is understood that the weight of the axle 802, which is the rear wheel of the vehicle 10, has reached the limit weight.

Next, the flow of the vehicle type discrimination processing of the present embodiment configured as described above will be described. FIG. 32 is a flow chart showing an example of the procedure of the vehicle type identification process in the vehicle type identification device 3103 of this embodiment. The process from acquisition of the image captured by the camera in S40 to detection of the axle in S47 is performed in the same manner as in S10 to S17 of the vehicle type determination process of the first embodiment described in FIG.

After the axle detection unit 1095 measures the number of axles and the distance between the axles in S47, the acquisition unit 2091 determines the weight of the vehicle 10 from the weight measuring devices 105a to 105c and the time at which the weight of the vehicle 10 is measured in S48. It is acquired and sent to the axle-by-axle weight measuring unit 98.

In S49, the axle weight measurement unit 98 collates the weight of the vehicle 10 acquired from the acquisition unit 2091, the time when the weight of the vehicle 10 was measured, and the axle detection result acquired from the axle detection unit 1095.

In S50, the axle weight measuring unit 98 measures the weight of each axle. Further, when the axle-by-axle weight measuring unit 98 detects an axle that is floating from the road surface, it measures the number of axles and the distance between the axles excluding the axle.

In S51, the determination unit 1096 determines the vehicle type classification and vehicle type of the vehicle 10 based on the number of axles acquired from the axle weight measurement unit 98 and the measurement result of the distance between the axles.

In S52, the overload determination unit 99 determines whether the vehicle 10 is overloaded. Further, the overload determination unit 99 determines whether or not the weight of each axle has reached the limit weight.

In S53, the determination unit 1096 outputs the vehicle type determination result to an external device (not shown). Further, the overload determination unit 99 outputs the overload determination result, the determination result of whether or not each axle has reached the limit weight, and the weight measurement result for each axle to an external device (not shown).

As described above, according to the vehicle type identification device 3103 of the present embodiment, the acquisition unit 2091 acquires the weight of the vehicle 10 from the weight measuring devices 105a to 105c, and the overload determination unit 99 determines the vehicle type determination result of the determination unit 1096. Since it is determined whether the vehicle 10 is overloaded based on the weight of the vehicle 10, the overloaded vehicle 10 can be efficiently determined. Although the upper limit value of the weight naturally varies depending on the vehicle type classification and the vehicle type, according to the vehicle type determination device 3103 of the present embodiment, the overload determination unit 99 determines that the overloading is overloaded after the determination unit 1096 determines the vehicle type classification and the vehicle type. By performing the determination, it is possible to appropriately determine whether or not the overloading is performed.

Further, according to the vehicle type identification device 3103 of the present embodiment, the acquisition unit 2091 acquires the weight of the vehicle 10 and the time when the weight of the vehicle 10 is measured from the weight measuring devices 105a to 105c, and measures the weight by axle. The unit 98 collates the weight of the vehicle 10 and the time when the weight of the vehicle 10 is measured with the axle detection result of the axle detection unit 1095 to measure the weight of each axle, and thus the load applied to each axle is grasped. can do. Therefore, according to the vehicle type identification device 3103 of the present embodiment, the vehicle 10 that is in a dangerous traveling state even if the total load capacity of the load is within the standard, or the vehicle 10 that has a bad load balance of the load is determined. be able to.

Further, according to the vehicle type identification device 3103 of the present embodiment, the axle-by-axle weight measuring unit 98 determines whether or not the axle is floating from the road surface based on the weight measurement result of the axle, and the axle that is floating from the road surface. The lift axle function of the vehicle 10 in order to measure the number of axles and the distance between the axles, and the determination unit 1096 determines the vehicle type classification and vehicle type of the vehicle 10 based on the measurement result of the axle weight measurement unit 98. Thus, even if the distance between the lifted axle and the road surface is small, the vehicle type classification and the vehicle type determination can be performed with high accuracy.

As described above, according to the vehicle type identification systems 1a to 1d and the vehicle type identification devices 103 to 3103 of the first to fourth embodiments, it is possible to improve the accuracy of the vehicle type classification of the vehicle 10 and the determination of the vehicle type.

The vehicle type identification devices 103 to 3103 of the above-described first to fourth embodiments are ordinary computers including a CPU (Central Processing Unit), a RAM (Random Access Memory), a ROM (Read Only Memory), an HDD (Hard Disk Drive), and the like. It has a hardware configuration.

In addition, the vehicle type identification processing executed by the vehicle type identification devices 103 to 3103 of the above-described first to fourth embodiments is provided as a computer program product as a vehicle type identification program preinstalled in a ROM or the like. The vehicle type discriminating program executed by the vehicle type discriminating apparatuses 103 to 3103 according to the first to fourth exemplary embodiments is a CD-ROM, a flexible disk (FD), a CD-R, a DVD (a file in an installable format or an executable format). It may be configured to be recorded in a computer-readable recording medium such as a Digital Versatile Disk) and provided as a computer program product.

Further, the vehicle type identification programs executed by the vehicle type identification devices 103 to 3103 of the above-described first to fourth embodiments are stored on a computer connected to a network such as the Internet and provided as a computer program product by downloading via the network. It may be configured to do so. Further, the vehicle type identification program executed by the vehicle type identification devices 103 to 3103 of the above-described first to fourth embodiments may be provided or distributed as a computer program product via a network such as the Internet.

The vehicle type identification program executed by the vehicle type identification devices 103 to 3103 according to the first to fourth embodiments has a module configuration including the above-described units, and as actual hardware, a CPU (processor) determines the vehicle type from the ROM. By reading and executing the program, the above-mentioned respective parts are loaded on the RAM, and the respective parts are generated on the RAM.

While some embodiments of the present invention have been described, these embodiments are presented as examples and are not intended to limit the scope of the invention. These novel embodiments can be implemented in various other forms, and various omissions, replacements, and changes can be made without departing from the spirit of the invention. These embodiments and modifications thereof are included in the scope and the gist of the invention, and are also included in the invention described in the claims and the scope equivalent thereto.

1a, 1b, 1c, 1d Vehicle type identification system 10 Vehicle 20 Vehicle type identification information 90, 1090 Storage section 91, 1091, 2091 Acquisition section 92, 1092, 2092 Image generation section 93 Panoramic image estimation section 94 Interpolation section 95, 1095 Axle axis detection section 96,1096 Discriminating unit 97 Counter 98 Axle weight measuring unit 99 Overload determining unit 101 Camera 102 Laser scanner 103, 1103, 2103, 3103 Vehicle type discriminating device 104 Speedometer 105, 105a, 105b, 105c Weight measuring device T Predetermined position

Claims (7)

The detection point distance detected by irradiating the vehicle passing the predetermined position with the laser detected by the scanner unit and the reflection intensity of the laser reflected at the detection point are acquired, and the predetermined position is continuously obtained from the imaging unit. an acquisition unit that acquires a plurality of still images obtained by imaging and,
The image of the predetermined position including the detection point, the density of the detection point to generate a reflectance image is an image made different depending on the reflection intensity of laser reflected in the detection point, the reflexology An image generation unit that performs a normalization process for adjusting the aspect ratio of the reflectance image to the actual aspect ratio of the vehicle with respect to the closet image ,
A panoramic image estimation unit that estimates a panoramic image, which is an image showing the overall shape of the vehicle, based on the plurality of still images continuously captured by the imaging unit;
A matching process is performed to match the feature points of the panoramic image with the detection points of the normalized reflectance image, and the panoramic image is based on the result of the matching process and the detection point distance. An interpolation unit for interpolating the feature points of
An axle detecting unit that detects an axle from the panoramic image in which the feature points are interpolated, and measures the number of the axles and the distance between the axles,
A determination unit that determines the vehicle type of the vehicle based on the number of axles and the distance between the axles;
Vehicle type identification device equipped with. If the determination unit determines that the vehicle has three or more axles and that the distance between the axles is two or more, the vehicle is an oversized vehicle. Determine that there is,
The vehicle type identification device according to claim 1. The acquisition unit further acquires the speed of the vehicle from a speed measurement unit,
The image generation unit performs the normalization processing of the reflectance image based on the speed of the vehicle acquired by the acquisition unit,
The vehicle type identification device according to claim 1. The acquisition unit further acquires the weight of the vehicle from a weight measurement unit that measures the weight of the vehicle,
An overload determination unit that determines that the vehicle is overloaded when it is determined that the weight of the vehicle acquired by the acquisition unit exceeds an overload reference value set for the vehicle type determined by the determination unit Further comprising,
The vehicle type identification device according to claim 1. The acquisition unit further acquires the time when the weight of the vehicle is measured from the weight measurement unit,
The number of the axles detected by the axle detection unit, the distance between the axles, the time when the axle is scanned by the scanner unit, the weight of the vehicle acquired by the acquisition unit, and the time when the weight of the vehicle is measured. By further comparing, by further comprising an axle weight measuring unit for measuring the weight of each axle,
The overload determination unit determines whether the weight of each axle of the vehicle measured by the axle weight measurement unit has reached a limit weight that is an upper limit value of the weight of each axle,
The vehicle type identification device according to claim 4 . The axle-by-axle weight measurement unit determines the number of the axles and the number of the axles excluding the axles that are floating from the road surface when determining from the measurement result of the weight of each axle that the axle is floating from the road surface. Measure the distance between
The determining unit determines the vehicle type of the vehicle based on the number of the axles and the distance between the axles measured by the axle weight measuring unit,
The vehicle type identification device according to claim 5 . From the scanner unit, the detection point distance detected by irradiating the vehicle passing a predetermined position with a laser and the reflection intensity of the laser reflected at the detection point are acquired, and from the image pickup unit, the predetermined position is continuously obtained. An acquisition step of acquiring a plurality of still images obtained by imaging,
The image of the predetermined position including the detection point, the density of the detection point to generate a reflectance image is an image made different depending on the reflection intensity of laser reflected in the detection point, the reflexology An image generation step of performing a normalization process on the closet image so that the aspect ratio of the reflectance image matches the actual aspect ratio of the vehicle ,
A panoramic image estimation step of estimating a panoramic image, which is an image showing the shape of the entire vehicle, based on the plurality of still images continuously captured by the imaging unit ;
A matching process is performed to match the feature points of the panoramic image with the detection points of the normalized reflectance image, and the panoramic image is based on the result of the matching process and the detection point distance. An interpolation step of interpolating the feature points of
An axle detection step of detecting an axle from the panoramic image in which the feature points are interpolated, and measuring the number of axles and the distance between the axles,
A determining step of determining the vehicle type of the vehicle based on the number of the axles and the distance between the axles;
Vehicle type identification method including.