JP2004013592A - Vehicle detection device and method - Google Patents

Abstract

An object of the present invention is to reliably detect another vehicle even when the illuminance around the own vehicle is low and the other vehicle to be detected moves.
When detecting an object such as another vehicle around a host vehicle, first, a high brightness area included in a brightness image, which is obtained by continuously imaging the host vehicle over a plurality of frames, is provided. Extract 12A and 13A. In this vehicle detection device, the high-brightness areas 11A, 12A, and 13A are extracted, and the high-brightness areas in the brightness image between a plurality of temporally preceding and succeeding frames are identified, and the frames of the high-brightness areas 11A, 12A, and 13A are identified. The integrated image 21 is created by integrating the brightness values of the brightness images so as to overlap the high brightness regions 11A, 12A, and 13A identified over a plurality of frames using the positional information between the frames, and an integrated vehicle 21 is created from the created integrated image 21. Is detected.
[Selection] Fig. 2

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a vehicle detection device and method for capturing an image in front of a vehicle or behind a vehicle and performing image processing to detect other vehicles around the own vehicle, and particularly to other vehicles when the brightness around the vehicle is low such as at night. The present invention relates to a device and a method for detecting a vehicle.
[0002]
[Prior art]
2. Description of the Related Art Conventionally, there has been known a vehicle detection device that captures an image of the surroundings of a host vehicle using an in-vehicle camera and detects a nearby other vehicle using the captured image. In such a vehicle detection device, particularly when a camera is used, it is important to be able to reliably detect another vehicle even in an environment where ambient illuminance is low and the sensitivity of the camera is low.
[0003]
There are many known methods for reliably detecting a vehicle even when the illuminance around the own vehicle is low. As one of the methods, image integration processing for improving image sensitivity by superimposing images continuously input from a camera is proposed. There is something to do. This image integration process has many advantages, such as versatility and cost, as compared to exposure control that increases the amount of incident light on the camera.
[0004]
As a conventional technique for reliably detecting a vehicle even when the illuminance around the own vehicle is low by using the image integration processing as described above, there is a technique disclosed in Japanese Patent Application Laid-Open No. H11-96367. This conventional technique is effective as a method for reliably detecting a vehicle even when the illuminance around the own vehicle is low. However, simply integrating a continuously input image allows another vehicle to move with respect to the own vehicle. In such a case, there is a problem that the contour portion of the other vehicle image becomes trailing, and the contour portion becomes an unclear image, making it difficult to detect the other vehicle.
[0005]
On the other hand, in the related art, the difference image processing used for moving object detection and the like is used, and the difference in the image luminance value over time is small, and the weight of the point at which the other vehicle is stationary with respect to the own vehicle is increased. Then, image integration is performed, and the weight of the point at which the other vehicle is moving relative to the own vehicle is small, since the difference in the image luminance value changes with time, and the image integration is performed. As a result, the image sensitivity can be improved, and an integrated image can be calculated in which the moving object or the like hardly causes a blur due to trailing of the moving object. Therefore, even when the illuminance around the own vehicle is low, the other vehicle can be reliably detected. Was.
[0006]
[Problems to be solved by the invention]
However, in the above-described related art, since the weight of integration is changed according to the movement of the other vehicle on the captured image, the sensitivity of the point at which the movement of the other vehicle relative to the own vehicle is large cannot be improved by image integration. That is, in the related art, when integrating the image area in which another vehicle to be detected is moving, the weight of the luminance value is reduced, so that the sensitivity improvement processing by the image integration is not performed. I will. Therefore, in the related art, since the sensitivity of another moving vehicle cannot be improved, it has been difficult to detect in a situation where the surrounding illuminance is low.
[0007]
Therefore, the present invention has been proposed in view of the above-described circumstances, and reliably detects another vehicle even when the illuminance around the own vehicle is low and the other vehicle to be detected moves. The present invention provides a vehicle detection device and a vehicle detection method capable of detecting a vehicle.
[0008]
[Means for Solving the Problems]
According to the present invention, when detecting an object such as another vehicle around the own vehicle, first, a high-brightness area included in a brightness image continuously imaged over a plurality of frames around the own vehicle is extracted. The vehicle detection device identifies a high-luminance area in a luminance image between a plurality of temporally preceding and succeeding frames from which the high-luminance area is extracted, and uses the position information between the frames of the high-luminance area to determine a high-luminance area. In order to solve the above-described problem, the integrated image is created by integrating the brightness values of the brightness image so as to overlap the high brightness region identified in the above, and another vehicle is detected from the created integrated image.
[0009]
【The invention's effect】
According to the present invention, a high-luminance area is identified in a luminance image between a plurality of temporally preceding and succeeding frames from which a high-luminance area is extracted, and position information between frames of the high-luminance area is used to identify the high-luminance area. Since the integrated image is created by integrating the luminance values of the luminance image so as to overlap the identified high-luminance area, even when another vehicle moves and the position in the luminance image changes, the image area of the other vehicle is changed. , A sufficient luminance gradation can be obtained.
[0010]
Further, according to the vehicle detection device, even when another vehicle moves when the illuminance around the own vehicle is low, a sufficient luminance gradation can be obtained in the image area of the other vehicle.
[0011]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, a first embodiment and a second embodiment of the present invention will be described with reference to the drawings.
[0012]
[First Embodiment]
The present invention is applied to, for example, a vehicle detection device according to the first embodiment configured as shown in FIG.
[0013]
[Configuration of the vehicle detection device according to the first embodiment]
As shown in FIG. 1, a vehicle detection device according to the first embodiment is provided with a camera 1 provided so that light from the surroundings of the own vehicle is incident with the front of the vehicle as an imaging direction, and provided inside the own vehicle. Input image memory 2, control operation unit 3, integral image memory 4, display 5 provided at a position visible to the vehicle driver, and alarm device 6 including, for example, a sound emission mechanism for the vehicle driver. .
[0014]
The camera 1 is provided in the front part of the vehicle in the present example, and captures an image in front of the vehicle to generate a captured image. A captured image generated by the camera 1 is continuously input to the input image memory 2 at predetermined time intervals, converted into a digital value by the input image memory 2, and a luminance image of a plurality of temporally continuous frames, for example. Is held as
[0015]
The control calculation unit 3 performs other vehicle detection processing using the luminance images over a plurality of frames held in the input image memory 2, displays the detection result on the display 5, and displays the distance between the own vehicle and the other vehicle. When the distance is shorter than a predetermined distance, the warning device 6 is driven to issue a warning.
[0016]
When performing the other vehicle detection process, the control operation unit 3 reads out the luminance image from the input image memory 2, extracts the high luminance points in the luminance image, and associates the extracted high luminance points between the frames. Then, the luminance images of each frame are integrated so that the high luminance points between the frames overlap from the result of the association, and the integrated images are stored in the integrated image memory 4. At this time, the control calculation unit 3 causes the integrated image memory 4 to store the integrated image for each high luminance point. Then, the control calculation unit 3 reads out the integral image from the integral image memory 4 and performs other vehicle detection. When another vehicle is detected from the processing result, the control calculation unit 3 calculates vehicle information such as a vehicle position and a vehicle size, and controls operations of the display 5 and the alarm device 6.
[0017]
FIG. 2 shows a process of creating an integrated image by performing an integration process for each high-luminance point extracted by the control calculation unit 3. In the example shown in FIG. 2, a case is shown in which the other vehicle is a preceding vehicle and the left tail lamp when the vehicle is running at night is extracted as a high luminance point.
[0018]
In the vehicle detection device, when the luminance images 11, 12, and 13 are successively input by the camera 1 in the order of frames F1, F2, and F3, the control arithmetic unit 3 turns on the left tail lamp each time the luminance images 11 to 13 are input. It is extracted as high brightness points 11A, 12A, and 13A. At this time, even when the positions of the left tail lamps of the other vehicles in the images are different in the luminance images 11 to 13, the control calculation unit 3 associates the preceding and succeeding frames by performing, for example, a matching process or the like. High brightness points 11A, 12A and 13A are extracted.
[0019]
Then, the integral image 21 is obtained by integrating the luminance images such that the high luminance points 11A to 13A overlap. As a result, the left tail lamp is matched as a high brightness point as a detection portion of another vehicle, and an integrated image 21 in which the other vehicle including the left tail lamp to be detected is included in the structure is obtained. That is, since the control calculation unit 3 focuses on the high-brightness points and integrates the same high-brightness points in each of the brightness images 11 to 13 so as to overlap, the integrated image 21 having a sufficient gradation can be calculated. Thus, the vehicle can be reliably detected.
[0020]
FIG. 3 is a diagram illustrating a method for extracting a high-luminance area including the above-described high-luminance points. FIG. 3 also shows a case where the detection target is the left tail lamp of another vehicle, as in the case shown in FIG.
[0021]
For example, when detecting a high luminance point in the frame F1, the control calculation unit 3 sets a high luminance area (initial high luminance area) 31 as shown in FIG. The initial high luminance area 31 sets an image area of a preset image size centered on the point of the highest luminance in the luminance image.
[0022]
Then, when the luminance images 12 and 13 are continuously input, a change in luminance value (luminance distribution) of the luminance images 12 and 13 is observed. Here, as shown in FIG. 3B, the luminance distribution of the image area where the object (the left tail lamp of another vehicle) exists is unchanged (the shaded portion in FIG. 3B), and Utilizing the fact that the luminance distribution of the image area changes (open portion in FIG. 3B), the control calculation unit 3 changes the high luminance area in the image area direction where the luminance distribution does not change. . That is, by changing the initial high-brightness area 31 in FIG. 3B in the direction of the arrow, the high-brightness area is enlarged in the other vehicle direction.
[0023]
As a result, as shown in FIG. 3 (c), the control arithmetic unit 3 can set the initial high-luminance area 31 and the high-luminance area 32 including the entire other vehicle.
[0024]
[Other vehicle detection processing by the vehicle detection device according to the first embodiment]
Next, a processing procedure of another vehicle detection processing for detecting another vehicle ahead in the above-described vehicle detection device will be described with reference to a flowchart of FIG.
[0025]
For example, when the vehicle starts to run, the vehicle detection device is activated, and firstly, the camera 1 starts capturing an image in an image capturing direction and continuously stores a luminance image in the input image memory 2.
[0026]
Then, in step S1, the control arithmetic unit 3 captures the luminance image held in the input image memory 2, and proceeds to step S2.
[0027]
In step S2, the control calculation unit 3 refers to the luminance information of the luminance image captured in step S1, compares the luminance value with a predetermined threshold value set in advance, and binarizes the luminance value. A candidate pixel is extracted, and the process proceeds to step S3.
[0028]
In step S3, the control arithmetic unit 3 groups adjacent pixels of the high-brightness point candidates among the high-brightness point candidates extracted in step S2 as being the same high-brightness point, thereby obtaining a high-brightness point. The luminance region is extracted, and the process proceeds to step S4.
[0029]
At this time, the control calculation unit 3 performs the processing of step S2 on the luminance images in a plurality of continuous frames to extract a high luminance point, and further extracts a high luminance area in step S3.
[0030]
In step S4, the control operation unit 3 calculates a variance value of a change in the brightness value of each pixel by calculating a variance value of a position in the image with respect to a high brightness area of each brightness image that is continuously input. The process proceeds to S5. Here, the variance value is increased when the luminance value of each pixel constituting the luminance image changes greatly between frames, and the variance value is decreased when the change is small.
[0031]
In step S5, the control operation unit 3 refers to the variance value of each pixel calculated in step S4, expands the boundary of the high-luminance region to include the pixel with the small variance value, and increases the pixel with the large variance value. Is set so that the image size of the high-brightness area is changed from the initial high-brightness area 31 by reducing the boundary of the high-brightness area, and the process proceeds to step S6.
[0032]
In step S6, the control calculation unit 3 calculates the barycentric position of the binarized pixel in the high-luminance area for which the image size has been set in step S5, and proceeds to step S7. At this time, the control calculation unit 3 calculates the position of the center of gravity in the high-luminance area with reference to the luminance values of the pixels constituting the high-luminance area.
[0033]
In step S7, the control calculation unit 3 obtains the shape and position information of each high-luminance region extracted over a plurality of frames, identifies the high-luminance region over a plurality of frames, and proceeds to step S8.
[0034]
In step S8, the control calculation unit 3 determines whether or not the number of times of inter-frame identification of each high-luminance area performed in step S7 is equal to or greater than a specified value set in advance. As a result of the identification processing in step S7 by the control operation unit 3, when it is determined that the number of times of inter-frame identification has reached the specified value, the process proceeds to step S9, and when it is determined that the number of times of inter-frame identification has not reached the specified value, the process proceeds to step S1. Back. That is, the processing of steps S1 to S7 is repeatedly performed for a plurality of frames by the control calculation unit 3, and the process proceeds to step S9 when the number of times of inter-frame identification for each high-luminance area reaches the specified value.
[0035]
In step S9, based on the center-of-gravity position information of the high-brightness area calculated in steps S1 to S8, the high-brightness calculated in step S5 by the control calculation unit 3 so that the center of gravity of the high-brightness area in each frame overlaps. Image integration is performed on the area size, and the process proceeds to step S10.
[0036]
In step S10, edge detection processing of the integrated image calculated in step S9 is performed, and as a result of the edge detection processing, the shape of an object (other vehicle) that is detected as an edge and includes a high luminance point to be processed is extracted. Then, the process proceeds to step S11.
[0037]
In step S11, the control calculation unit 3 performs pattern matching between the shape of the target object extracted in step S10 and a vehicle model prepared and stored in advance, and a matching degree with a detected vehicle (hereinafter, referred to as a correlation degree). Is calculated, and the process proceeds to step S12.
[0038]
In step S12, the control calculation unit 3 determines whether the degree of correlation calculated in step S11 is equal to or greater than a specified value set in advance, and if it is equal to or greater than the specified value, the process proceeds to step S13. It is determined that the vehicle exists, and the process ends. Here, the preset designated value is a threshold value at which the object shape extracted in step S10 is recognized as a vehicle. On the other hand, if it is determined in step S12 that the degree of correlation is not equal to or greater than the specified value, it is determined that the object existing ahead is not a vehicle, and the process ends.
[0039]
[Effects of First Embodiment]
As described in detail above, according to the vehicle detection device according to the first embodiment, since another vehicle is detected by creating an integral image by superimposing the center of gravity of the high-luminance area extracted in each frame, Even if the position in the luminance image changes due to the movement of the vehicle, or if the position of the camera 1 moves due to pitching of the own vehicle, etc., a sufficient luminance gradation is obtained in the image area of another vehicle. Can be. Further, according to the vehicle detection device, even if the other vehicle moves when the illuminance around the own vehicle is low, for example, when the vehicle is running at night, sufficient luminance gradation can be obtained in the image area of the other vehicle. Can be obtained.
[0040]
Further, according to this vehicle detection device, the high-luminance points are extracted, and the image size of the high-luminance area is changed according to the presence or absence of a temporal change in the distribution of the high-luminance points. The entire area of the other vehicle whose structure is invariable can be a high brightness area. Therefore, according to this vehicle detection device, even when the surrounding illuminance is low, pattern matching can be performed using not only the light portion of the other vehicle but also the entire other vehicle, and the other vehicle can be reliably detected. Can be.
[0041]
[Second embodiment]
Next, a vehicle detection device according to a second embodiment will be described. In the following description, the same parts as those in the above-described first embodiment are denoted by the same reference numerals, and detailed description thereof will be omitted.
[0042]
[Operation of Vehicle Detecting Device According to Second Embodiment]
FIG. 5 is a diagram illustrating a process when the control calculation unit 3 groups high-luminance points.
[0043]
In this example, when two other vehicles, a first vehicle 41 and a second vehicle 42, are present in front of the host vehicle, each of the first vehicle 41 and the second vehicle 42 has two tail lamps. In step 3, a first high-luminance area 51 and a second high-luminance area 52 of the first vehicle 41 and a third high-luminance area 53 and a fourth high-luminance area 54 of the second vehicle 42 are extracted.
[0044]
Here, the first high brightness area 51 and the second high brightness area 52 are based on the tail lamp of the first vehicle 41, and the third high brightness area 53 and the fourth high brightness area 54 are based on the tail lamp of the second vehicle 42. Therefore, when an image is taken from the own vehicle, the image is left-right symmetric. In addition, the first high-luminance area 51 and the second high-luminance area 52 and the third high-luminance area 53 and the fourth high-luminance area 54 are formed by tail lamps of different vehicles, and thus have no relationship in shape.
[0045]
Due to the nature of each of the high-brightness areas, when a plurality of other vehicles are present in front of the host vehicle, the control calculation unit 3 sets the vehicle assuming that the high-brightness areas 51 to 54 are bilaterally symmetric. By grouping the lights, the first vehicle 41 and the second vehicle 42 can be divided and grouped to set a high luminance area.
[0046]
FIG. 6 is a diagram illustrating a process of changing the size of the luminance image captured by the camera 1 and generating an integrated image according to the distance between the own vehicle and another vehicle.
[0047]
The control calculation unit 3 captures the luminance images 61 and 62 in the frames F1 and F2, respectively, extracts a high luminance region of another vehicle, and obtains the symmetric tail lamp intervals W1 and W2. Here, when the distance between the host vehicle and the other vehicle is different from the frame F1 to the frame F2, the image size of the other vehicle is different, and the tail lamp interval W1 and the tail lamp interval W2 are different.
[0048]
In this case, the control calculation unit 3 calculates the change rate (W2 / W1) of the image size from the sizes W1 and W2 between the high brightness points (tail lamps) of the same object, and enlarges the brightness image 61 of the frame F1. Thereby, the control calculation unit 3 makes the image size of the other vehicle of the enlarged luminance image 61 ′ of the frame F1 equal to the image size of the other vehicle of the luminance image 62 of the frame F2, and makes the luminance image 61 ′ and the luminance image 62 to create an integral image.
[0049]
FIG. 7 is a diagram illustrating a process of creating an integral image when image rotation due to vehicle behavior (roll) occurs.
[0050]
When the luminance images 61 and 62 are captured in the frames F1 and F2, respectively, and the high luminance area of the other vehicle is extracted, the control calculation unit 3 tilts the line 71 connecting the symmetric tail lamps and the image horizontal line 72. Obtain the angle θ. Here, when the inclination angle θ1 is changed from the inclination angle θ1 in the frame F2 to the inclination angle θ2 in the frame F2, a difference angle (θ2−θ1) due to a rotation change is calculated from each of the inclination angles θ1 and θ2, and the luminance is calculated by the difference angle. The inclination angle between a line 71 connecting the tail lamps by rotating the image 61 and the image horizontal line 72 is defined as θ2. Then, the control calculation unit 3 performs image integration using the rotated luminance image 61 ′ and the luminance image 62.
[0051]
[Other vehicle detection processing by the vehicle detection device according to the second embodiment]
Next, in the vehicle detection device according to the second embodiment described above, a processing procedure of another vehicle detection processing for detecting another vehicle in front will be described with reference to a flowchart of FIG. Note that the same processes as those of the other vehicle detection process in the first embodiment are denoted by the same step numbers, and detailed description thereof is omitted.
[0052]
In the other vehicle detection process by the vehicle detection device according to the second embodiment, after step S2, the process is the same as step S3 described above, but in step S21, primary grouping is performed as the first-stage grouping. After performing the processing of steps S4 and S5, the secondary grouping of step S22 is performed.
[0053]
In the secondary grouping, the control calculation unit 3 detects a symmetric high luminance area from the high luminance areas extracted in step S4. If the control calculation unit 3 detects a symmetric high-luminance area in the binarized image, the control operation unit 3 groups the high-luminance areas as high-luminance areas of the same vehicle and proceeds to step S6. At this time, when a plurality of different vehicles exist in the binarized image, the control calculation unit 3 compares the shapes of the high-brightness areas to identify and group the high-brightness areas by different vehicles. . That is, the control calculation unit 3 groups the left and right symmetric tail lamps as shown in FIG. 5, for example.
[0054]
In step S23 after detecting the center of gravity in step S6, the control calculation unit 3 calculates the image distance between the high-brightness points of the high-brightness areas grouped in step S22, and proceeds to step S24.
[0055]
In step S24, the control calculation unit 3 calculates the inclination angle θ of the line (line 71) connecting the high luminance points of the high luminance areas grouped in step S22 with respect to the road surface (image horizontal line 72), and then proceeds to step S7. When the number of times of identification in step S7 is equal to or more than the specified value, the process proceeds to step S9.
[0056]
When creating the integral image in step S9, the control calculation unit 3 sets the image based on the image size calculated in step S23 such that the high-luminance areas in the binary images of the preceding and succeeding frames are the same. The size is changed, and based on the tilt angle calculated in step S24, the image is rotated so that the high-luminance areas in the binarized images of the preceding and succeeding frames have the same tilt angle. As a result, an integrated image is created as shown in FIGS. 6 and 7, and the processing after step S10 is performed.
[0057]
[Effect of Second Embodiment]
As described in detail above, according to the vehicle detection device according to the second embodiment, as described in FIG. 5 and step S22, the shapes of the nearby high-luminance regions are compared, and the left-right symmetric high-luminance regions are compared. Are grouped, so that even when a plurality of high-luminance points are observed in a situation where the surrounding illuminance is low, the other vehicles are reliably detected by grouping for each object (vehicle) having the high-luminance points. be able to.
[0058]
Further, according to the vehicle detection device, an interval between a plurality of grouped high-brightness areas is detected, and the image size is enlarged (or reduced) according to a change in the interval between successively input frames. Therefore, an integrated image with a sufficient gradation can be obtained even when the distance between other vehicles is different. Therefore, according to the vehicle detection device, in a situation where the surrounding illuminance is low, the other vehicle can be reliably detected even when the other vehicle approaches or separates.
[0059]
Further, according to this vehicle detection device, the inclination angle with respect to the image horizontal line of the plurality of grouped high-luminance areas is detected, and the binarized image is rotated according to the inclination angle change between the continuously input frames. Therefore, even when another vehicle rolls or the camera 1 is tilted, an image having a sufficient gradation can be obtained. Therefore, according to the vehicle detection device, even when another vehicle rolls in a situation where the surrounding illuminance is low, it can be reliably detected.
[0060]
Note that the above embodiment is an example of the present invention. For this reason, the present invention is not limited to the above-described embodiment, and other than the present embodiment, various modifications may be made according to the design and the like within a range not departing from the technical idea according to the present invention. Can be changed.
[Brief description of the drawings]
FIG. 1 is a block diagram illustrating a configuration of a vehicle detection device according to a first embodiment to which the present invention has been applied.
FIG. 2 is a diagram for explaining a process in which a control operation unit performs an integration process for each extracted high-luminance point to create an integrated image.
3A and 3B are diagrams illustrating a method of extracting a high-brightness area including a high-brightness point, wherein FIG. 3A illustrates a case where an initial high-brightness area is set, and FIG. Is changed, and (c) shows a case where a high-luminance area of the entire other vehicle is set.
FIG. 4 is a flowchart illustrating a processing procedure of another vehicle detection processing for detecting another vehicle in front in the vehicle detection device according to the first embodiment to which the present invention is applied.
FIG. 5 is a diagram illustrating a process when a high-intensity point is grouped by a control calculation unit in a vehicle detection device according to a second embodiment to which the present invention is applied.
FIG. 6 is a diagram illustrating a process of changing the size of a luminance image captured by a camera and generating an integrated image according to the distance between the host vehicle and another vehicle.
FIG. 7 is a diagram illustrating a process of creating an integral image when image rotation due to vehicle behavior (roll) occurs.
FIG. 8 is a flowchart showing a procedure of another vehicle detection processing for detecting another vehicle ahead in the vehicle detection device according to the second embodiment to which the present invention is applied.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 Camera 2 Input image memory 3 Control operation part 4 Integral image memory 5 Display 6 Alarm device 21 Integral image 31 Initial high luminance area 32 High luminance area 41 First vehicle 42 Second vehicle 51 First high luminance area 52 Second high luminance Area 53 Third high brightness area 54 Fourth high brightness area 61, 62 Brightness image 71 Line 72 Image horizontal line

Claims (14)

Imaging means for imaging the surroundings of the vehicle,
A high-luminance area extraction unit that extracts a high-luminance region included in the luminance image captured by the imaging unit,
Region identification means for identifying a high-luminance area in a plurality of luminance images from which the high-luminance area has been extracted by the high-luminance extraction means,
Using the position information between the frames of the high-luminance area extracted by the high-luminance area extraction means, the image-capturing means continuously overlaps the high-luminance areas identified between a plurality of frames by the area identification means. Image integration means for integrating the luminance value of the luminance image input to
A vehicle detection device comprising: a vehicle detection unit that detects another vehicle from the integrated image created by the image integration unit. 2. The vehicle according to claim 1, wherein the image integration means obtains a center of gravity of a high-brightness area in the brightness image in each frame, and creates an integrated image so as to overlap the center of gravity of each high-brightness area. Detection device. The high-luminance area extracting means extracts a high-luminance area of the luminance image as a high-luminance area with a preset image size, and extracts the high-luminance area based on a temporal change in the luminance value of the luminance image. The vehicle detection device according to claim 1, wherein an image size is changed. The high-brightness area extracting means detects a positional relationship between the plurality of high-brightness areas, and, when the two high-brightness areas are symmetric, the two high-brightness areas are grouped as a high-brightness area by the same other vehicle. The vehicle detection device according to claim 1, wherein the vehicle detection device is used. The high-brightness area extracting means compares the shapes of the plurality of high-brightness areas, and groups the plurality of high-brightness areas as high-brightness areas by the same other vehicle based on a result of the comparison. The vehicle detection device according to claim 1. The high-brightness region extracting means detects distances in the image of the plurality of high-brightness regions, and sets the image sizes of the high-brightness regions in the luminance images that are temporally different from each other based on the temporal change of the distances in the image. The vehicle detection device according to claim 1, wherein the image size of the luminance image is changed so as to satisfy the following condition. When the plurality of high-luminance regions are grouped, the high-luminance region extracting means detects a tilt in the image of the plurality of high-luminance regions, and determines a luminance which changes in time based on a temporal change of the tilt in the image. The vehicle detection device according to claim 1, wherein the luminance image is rotated such that a high luminance area in the image has the same inclination in the image. A high-luminance area included in a luminance image obtained by continuously capturing the surroundings of the vehicle over a plurality of frames is extracted,
The high-luminance area is extracted and the high-luminance area in the luminance image between a plurality of temporally preceding and succeeding frames is identified,
Using the position information between the frames of the high-luminance area, to create an integrated image by integrating the luminance value of the luminance image so as to overlap the high-luminance area identified between a plurality of frames,
A vehicle detection method for detecting another vehicle from the created integral image. When creating the integral image,
Find the center of gravity of the high-luminance area in the luminance image in each frame,
9. The vehicle detecting method according to claim 8, wherein an integral image is created by superimposing the center of gravity of each high-luminance area. When extracting the high brightness area,
In the luminance image, a region having a high luminance value is extracted as a high luminance region with a preset image size,
9. The vehicle detection method according to claim 8, wherein an image size of the preset high luminance area is changed based on a temporal change of a luminance value of the luminance image. When extracting the high brightness area,
Detects the positional relationship between multiple high-luminance areas,
9. The vehicle detection method according to claim 8, wherein when the two high-luminance regions are symmetric, the two high-luminance regions are grouped as high-luminance regions of the same other vehicle. When extracting the high brightness area,
Compare the shapes of multiple high brightness areas,
9. The vehicle detection method according to claim 8, wherein a plurality of high-luminance areas are grouped as high-luminance areas by the same other vehicle based on a result of the shape comparison. When extracting the high brightness area,
Detect distances in the image of multiple high brightness areas,
The image size of a brightness image is changed so that the image size of a high brightness area in a brightness image which temporally changes before and after may be the same based on a temporal change of the intra-image distance. Vehicle detection method. When extracting the high brightness area,
When multiple high brightness areas are grouped,
Detects tilt in the image of multiple high brightness areas,
The vehicle according to claim 8, wherein the brightness image is rotated based on the temporal change of the tilt in the image such that the tilt in the image of the high brightness area in the brightness image that precedes and succeeds in time is the same. Detection method.

Images