JP2007279805A - Obstacle collision determination device and obstacle collision alarm system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an obstacle collision determination device and an obstacle collision alarm system, which determine a collision with an object at an intersection with poor visibility. <P>SOLUTION: The device to be mounted on a vehicle comprises: an object detection means for detecting an object and a road reflector from image data; a road reflector state recognition means for recognizing the state of the road reflector detected by the object detection means; an obstacle state recognition means for recognizing the state of the obstacle detected by the object detection means; and an access determination means for determining the access of the obstacle, based on the recognition result of the obstacle state recognition means. The obstacle state recognition means recognizes the state of the obstacle, based on the recognition result of the road reflector state recognition means, and collision of the vehicle with the object at the intersection with poor visibility is thereby determined. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a technique for acquiring an image using an in-vehicle camera, measuring a distance from a vehicle to an object to be an obstacle, and determining the possibility of a collision between the vehicle and the obstacle, particularly for a road reflector. The present invention also relates to an obstacle collision determination device that measures the distance from a vehicle to determine the possibility of a collision between the vehicle and an obstacle, and an obstacle collision warning system equipped with the obstacle collision determination device.

  In order to prevent accidents in driving a car, various in-vehicle systems that support safe driving have been devised. As one of such systems, in order to prevent collision accidents between the host vehicle and obstacles such as other vehicles and pedestrians, it recognizes the surrounding conditions of the vehicle by recognizing and processing image data taken outside the vehicle. However, there are some which make an obstacle approach judgment and a driver warning. This system mainly includes an image acquisition device for acquiring image data of the surrounding situation of the own vehicle, an obstacle collision determination device for recognizing the acquired image data and determining a collision with an obstacle, and a determination It consists of a warning device that warns the driver based on the result.

As a method for the obstacle collision determination device to measure the relative distance of the obstacle with respect to the host vehicle, a method for calculating the parallax of an image acquired by a stereo camera is known. In the obstacle collision determination device that calculates the parallax of an image acquired by a stereo camera, a pair of images formed on an image sensor by a pair of optical systems are compared, and the images from the optical axes of both images are compared. The amount of deviation is detected, the calculation is performed based on the principle of triangulation, the distance to the object is measured, and the object up to the object is recognized without misrecognizing the background other than the object even when traveling on a curve etc. An improvement in which the distance is accurately measured has also been proposed (see, for example, Patent Document 1).
Japanese Patent Laid-Open No. 11-51646 (3rd page, FIG. 6)

  At intersections with poor visibility, road reflectors are installed so that the driver can easily recognize vehicles appearing behind shades such as walls. However, the conventional obstacle collision determination device measures the distance between the vehicle and the object from the image acquired using the in-vehicle camera and determines the possibility of the collision, but the road in the acquired image There is no special treatment for the object reflected on the reflector. Therefore, even if a road reflector is installed, the conventional obstacle collision determination device determines that the road reflector itself is an obstacle, or the obstacle reflected on the road reflector is left as it is. It will be recognized in the state.

  For example, when the road reflector is installed so that the road in the crossing direction can be seen from the own vehicle, even if the obstacle is traveling in the crossing direction with respect to the own vehicle, An obstacle cannot be recognized at all, or it is erroneously recognized so as to approach the host vehicle from the front. FIG. 4 shows the positional relationship among a road reflector, an obstacle, and the host vehicle installed at an intersection with poor visibility. In FIG. 4, the pair of cameras 41 and 42 installed in the host vehicle 40 measure the distance between the obstacle 60 and the host vehicle 40 from the image 61 of the obstacle 60 reflected on the road reflector 50. In FIG. 4, the obstacle 60 actually moves from the right to the left side of the wall 70, but the obstacle collision determination device has an image 61 of the obstacle 60 in front of the cameras 41 and 42. Therefore, the distance (L2) is measured as the distance between the obstacle 60 and the host vehicle 40, assuming that the obstacle 60 is ahead. In this state, the driver is given inappropriate information that the obstacle 60 is located at the front distance (L2). Thus, although the road reflector 50 shows an image of the obstacle 60 having a high possibility of collision, the conventional obstacle collision determination device ignores the information or is inappropriate inappropriate from the actual situation. There was a problem of providing information.

  The present invention has been made in order to solve the above-described conventional problems, and an obstacle that can measure the actual distance from the vehicle and determine the possibility of a collision for an object reflected on a road reflector. An object of the present invention is to enable collision judgment of an obstacle at an intersection with poor visibility by supplying a collision judgment device.

  An obstacle collision determination device of the present invention is an apparatus mounted on a vehicle, and includes an object detection unit that detects an obstacle and a road reflector from image data, and a state of the road reflector detected by the object detection unit. A road reflector state recognition unit that recognizes, an obstacle state recognition unit that recognizes the state of an obstacle detected by the object detection unit, and an approach that determines the approach of an obstacle based on the recognition result of the obstacle state recognition unit The obstacle state recognition means has a configuration for recognizing the state of the obstacle based on the recognition result of the road reflector state recognition means.

  With this configuration, even in an intersection with poor visibility, it is possible to measure the state of obstacles reflected on the road reflector by recognizing the state of the road reflector installed at the intersection, Judgment of collision with objects is possible.

  In the obstacle collision determination device of the present invention, the state of the obstacle recognized by the obstacle state recognition means is a relative position with respect to the vehicle, and the state of the road reflector recognized by the road reflector state recognition means is the vehicle It has the structure which consists of relative position and angle with respect to.

  With this configuration, even at an intersection with poor visibility, the relative position of the obstacle reflected in the road reflector can be determined by recognizing the relative position of the road reflector installed at the intersection with the host vehicle and the angle with respect to the host vehicle. It is possible to measure the collision of the obstacle.

  In the obstacle collision determination device of the present invention, the obstacle detected by the object detection means is configured to be a four-wheeled vehicle, a two-wheeled vehicle, or a pedestrian.

  With this configuration, even in an intersection with poor visibility, the state of a four-wheeled vehicle, two-wheeled vehicle, or pedestrian reflected on the road reflector can be measured by recognizing the state of the road reflector installed at the intersection. It is possible to judge the collision.

  In the obstacle collision determination device of the present invention, the object detection means has a configuration for detecting the presence of a road reflector based on background continuity or background distortion in image data.

  With this configuration, the road reflector in the image data can be detected with high accuracy from the background information even in an environment in which the road reflector is difficult to identify.

  In the obstacle collision determination device of the present invention, the road reflector state recognition means has a configuration for analyzing the shape of the road reflector and recognizing the state of the road reflector based on the analysis result of the shape. ing.

  With this configuration, the state of the road reflector necessary for measuring the state of the obstacle can be recognized with high accuracy based on the shape of the road reflector.

  The obstacle collision determination device of the present invention further includes an intersection shape information storage unit that stores intersection shape information of a road on which the vehicle is traveling, and the road reflector state recognition unit is stored in the intersection shape information storage unit. It has a configuration for recognizing the state of the road reflector based on the intersection shape information.

  With this configuration, the state of the road reflector necessary for measuring the state of the obstacle can be recognized with high accuracy based on the intersection shape information.

  The obstacle collision determination device of the present invention includes a means for detecting a road reflector and an obstacle from image data, a means for recognizing an angle between the relative position of the road reflector and the host vehicle, and the obstacle. Means for measuring the relative position between the vehicle and the host vehicle, and means for judging the approach of the obstacle based on the measurement result of the relative position of the obstacle. In addition, the relative position of the obstacle reflected on the road reflector with the host vehicle can be measured. As a result, it has an excellent effect that it is possible to determine the possibility of collision with an obstacle reflected on a road reflector installed at an intersection with poor visibility.

  Hereinafter, an obstacle collision determination device according to an embodiment of the present invention will be described with reference to the drawings.

(First embodiment)
FIG. 1 is a block diagram of the obstacle collision determination device according to the first embodiment of the present invention. In FIG. 1, in addition to the obstacle collision determination device 1 in the vehicle 4, an image acquisition device 2 and a warning device 3 are mounted and connected to each other to acquire an image and warn of a collision. Shows the configured figure. In the present invention, the image acquisition device 2 is equipped with a pair of cameras, a first camera 201 and a second camera 202. The first camera 201 and the second camera 202 are installed in parallel as a stereo configuration. The pair of cameras is a camera device that captures the front of the vehicle. However, even if it is a camera device that captures the side and the rear of the vehicle, other methods can be used.

  As shown in FIG. 1, the obstacle collision determination device 1 includes an object detection unit 11, a distance measurement unit 12, an obstacle state recognition unit 13, a collision determination unit 14, a road reflector state recognition unit 15, And an intersection shape storage unit 16. The object detection unit 11 receives image data obtained by capturing an image in front of the vehicle from the image acquisition device 2, and performs an obstacle and road reflector detection process. The distance measuring unit 12 receives image data obtained by capturing an image in front of the vehicle from the image acquisition device 2, and obtains a relative distance from the own vehicle for an arbitrary position in the image data. The obstacle state recognition unit 13 operates when a four-wheeled vehicle, a two-wheeled vehicle, or a pedestrian that is an obstacle is detected in the detection result of the object detection unit 11, and recognizes the state of the detected obstacle. The state of the obstacle is a relative position of the obstacle with the host vehicle.

  The road reflector state recognition unit 15 operates when a road reflector is detected in the detection result of the object detection unit 11, and recognizes the state of the detected road reflector. The state of the road reflector is the relative position and angle with the host vehicle. The intersection shape information storage unit 16 stores in advance intersection shape information of the road on which the vehicle 4 travels. The intersection shape information storage unit 16 stores the intersection shape information including information on the position and angle of the road reflectors arranged at the intersection. The collision determination unit 14 determines a collision with an obstacle based on the recognition result of the obstacle state recognition unit 13.

  When the collision determination unit 14 determines that there is a risk of collision with an obstacle, the warning device 3 warns the driver of the risk of collision. The warning device 3 is a speaker device that notifies the warning of the collision by voice, but can be implemented similarly as a monitor device that notifies the warning of the collision by characters or images, or by another method.

  As described above, in the obstacle collision warning system configured by the obstacle collision determination device 1, the image acquisition device 2, and the warning device 3, the image data in the image acquisition device 2 is one by one from each of the cameras 201 and 202. Two pieces are acquired as a set at the same time. The distance measuring unit 12 obtains the relative distance from the host vehicle at an arbitrary position of the image data based on the set of image data. The calculation of the relative distance is performed based on the deviation (parallax) of the object in a set of images. It is calculated that the larger the parallax in the image, the closer to the host vehicle.

  In order to recognize the obstacle state in the obstacle state recognition unit 13, the recognition result of the road reflector is used. When a road reflector is present, in order to measure the relative position of the obstacle, the recognition process is switched between when the obstacle image is inside and outside the road reflector. If the obstacle image is outside the road reflector, the conventional vehicle recognition method may be used.However, if the obstacle image is inside the road reflector, the vehicle will be reflected by the road reflector. It is necessary to obtain the relative distance (LX) on the assumption that it exists.

  A method for obtaining the relative distance (LX) will be described. The relative distance (LX) is the distance (L1) between the road reflector and the host vehicle, the angle (θ) of the road reflector with respect to the host vehicle, and the distance between the obstacle and the host vehicle as a mirror image reflected on the road reflector ( L2). In order to facilitate understanding, FIG. 2 shows a state where a road reflector is in front of the host vehicle and an obstacle is reflected on the road reflector as a mathematical model. In FIG. 2, the host vehicle is the point “A”, the obstacle is the point “B”, the road reflector is “MM”, the position of the road reflector is the point “C”, and the mirror image of the obstacle on the road reflector is shown. Is designated as point “B ′”. Note that the point “D” is a virtual point for deriving a calculation formula for obtaining the relative distance (LX). In FIG. 2, the road reflecting mirror is described as a plane mirror for ease of explanation.

In FIG. 2, the distance (L2) between the mirror image “B ′” of the obstacle reflected on the road reflector and the vehicle “A” (L2) is the distance between the road reflector “MM” and the vehicle “A” ( The line segment B′C is longer than L1). Since the line segment B′C is equal to the line segment BC, the line segment BC = (L2−L1). ∠BCD has a size obtained by subtracting ∠B′CB, that is, 2θ from 180 °, and ∠BCD = (π−2θ). Here, of the three sides of the right triangle ΔADB, the lengths of the two sides BD and AD are BD = (L2−L1) sin (π−2θ) and AD = {L1− (L2−L1) cos ( Therefore, from the Pythagorean theorem, that is, the relationship LX ² = BD ² + AD ^2, the following equation for obtaining the relative distance (LX) that is another side is derived.

  As is clear from equation (1), the relative distance (LX) is the distance (L1) to the road reflector, the angle (θ) of the road reflector with respect to the vehicle, and the mirror image reflected on the road reflector. It can be determined from the distance (L2) between the obstacle and the vehicle.

  Since Japanese road reflecting mirrors often use convex mirrors, a calculation formula for correcting the calculation formula of relative distance (LX) according to the curved surface state of the convex mirror may be used. Further, the method for obtaining the relative position of the obstacle based on the state of the road reflector is not limited to this, and other methods may be used.

  As a procedure, the distance (L1) to the road reflector and the distance (L2) to the obstacle as a mirror image reflected on the road reflector are first obtained from the output result of the distance measuring unit 12. Then, it can be seen that the road reflector mirror recognition unit 15 obtains the angle (θ) of the road reflector with respect to the host vehicle and obtains the distance (LX) from these relationships.

  A method for detecting the presence of the road reflector in the object detection unit 11 will be described. The image data is searched for a place where background continuity is interrupted. For example, when a road reflector is installed in front of a building, the background should be interrupted at the boundary line portion of the road reflector, and a hole should appear in the building. Search for such places where the background breaks. If a road reflector cannot be found due to the feature of background continuity, this time, the image is searched for distortion. When there is a road reflector in the image, the projected image is distorted. A road reflector is detected by searching for a distorted portion from the entire image. In this embodiment, the presence of the road reflector is detected based on both the information on the background continuity and the background distortion. However, the present invention is not limited to this, and only the background continuity is detected. Or it may be recognized only by the distortion of the background, or another method may be used.

  A method of recognizing an angle which is one of the states of the road reflector in the road reflector state recognition unit 15 will be described. First, the shape of the road reflector detected by the object detection unit is recognized. The road reflector itself is circular, but when viewed from the host vehicle traveling on the road, the road reflector looks elliptical. By obtaining the ratio of the length of the major axis to the minor axis of the ellipse, the angle of the road reflector with respect to the host vehicle can be determined. Further, by obtaining the angle formed by the major axis of the ellipse and the vertical direction, the angle of the road reflector with respect to the road surface can be known (this is referred to as angle E). Next, intersection shape information on the installation position of the road reflector is retrieved from the intersection shape storage unit 16. The absolute position of the host vehicle is specified using GPS. The relative position of the road reflector with respect to the host vehicle is obtained by the distance measuring unit 12. The absolute position of the road reflector can be determined from the absolute position of the host vehicle and the relative position of the road reflector. The intersection shape information is searched from the intersection shape storage unit 16 using the information on the absolute position of the road reflector. Based on the intersection shape information obtained by the search, the angle of the road reflector is obtained (this is defined as angle F). Based on the information (angle E, angle F) of the angle of the road reflector obtained above, the angle of the road reflector is calculated. For example, the average of the angle E and the angle F is taken, and this average value is determined as the angle of the road reflector. In the present embodiment, the angle of the road reflector is recognized based on the information of both the shape of the road reflector and the intersection shape. However, the present invention is not limited to this, and only the shape of the road reflector is recognized. Alternatively, it may be recognized only by the shape of the intersection, or another method may be used.

  As described above, the object detection unit 11 detects the presence of the road reflector, the road reflector state recognition unit 15 recognizes the angle (θ) of the road reflector, and the distance measurement unit 12 determines the distance to the road reflector. By obtaining (L1) and the distance (L2) to the obstacle as a mirror image reflected on the road reflector, the distance (LX) between the host vehicle and the obstacle can be obtained.

  In the present embodiment, parallax information from a stereo camera is used as a method of measuring the relative positions of the vehicle and the road reflector, but other methods may be used. The camera of the image acquisition device 2 may be a single eye, and the position may be obtained by estimating the distance from the host vehicle based on the size information of each object detected by the object detection unit 11.

  In the present embodiment, the obstacle detected by the object detection unit 11 is described as a vehicle. However, this vehicle may be either a four-wheeled vehicle or a two-wheeled vehicle, and other forms. It may be a vehicle. Further, even if the obstacle is a pedestrian, it can be similarly implemented. When an obstacle other than a vehicle is assumed, the obstacle state recognition unit 13 is replaced with a recognition unit that matches the assumed obstacle, and the state of the obstacle is recognized using the state of the road reflector.

  Next, the operation of the obstacle collision determination device in the embodiment of the present invention will be described with reference to FIG. First, the obstacle collision determination device 1 receives image data obtained by photographing the front of the host vehicle from the image acquisition device 2 (S101). When the object detection unit 11 of the obstacle collision determination device 1 receives the image data from the image acquisition device 2, the object detection unit 11 performs a process of detecting a preceding vehicle and a road reflector from the image data (S102). If no road reflector is present in the image data in the processing result of the object detection unit 11 (No in S103), it is confirmed whether a vehicle ahead is present. If there is no preceding vehicle in the image data (No in S104), the process returns to the process of acquiring the image data from the image acquisition device 2 again. When the vehicle ahead is present in the image data (Yes in S104), the obstacle state recognition unit 13 performs normal obstacle state recognition without considering the presence and state of the road reflector (S105). If a road reflector is present in the image data in the processing result of the object detection unit 11 (Yes in S103), the road reflector state recognition unit 15 recognizes the state of the road reflector (S106). At this time, information stored in the intersection shape storage unit 16 is also used. When the processing of the road reflector state recognition unit 15 is completed, it is confirmed in the output result of the object detection unit 11 whether there is a vehicle ahead including the inside and outside of the road reflector. If there is no preceding vehicle in the image data (No in S107), the process returns to the process of acquiring the image data from the image acquisition device 2 again. When a forward vehicle exists in the image data (Yes in S107), it is confirmed whether the forward vehicle exists inside the road reflector, that is, whether it is an object reflected on the road reflector. When the vehicle ahead does not exist inside the road reflector (No in S108), the obstacle state recognition unit 13 performs normal obstacle state recognition without considering the presence and state of the road reflector (S105). When the preceding vehicle is present inside the road reflector (Yes in S108), the obstacle state recognition unit 13 performs obstacle state recognition in consideration of the state of the road reflector (S109). When the collision determination unit 14 receives the result of the obstacle state recognition from the obstacle state recognition unit 13, the collision determination unit 14 determines the possibility of a collision with the obstacle based on the information. If it is determined that there is a possibility of an obstacle collision, the warning device 3 is used to warn the driver of the danger of the collision (S110). The above operation is repeated until the processing is completed (S111). This process may be started at the same time as starting the engine of the vehicle, and may be ended at the same time as stopping the engine, or may be another method.

  The obstacle collision judgment device of the present invention is useful in a system that is mounted on a vehicle and warns of the danger of collision with an obstacle by recognizing an image acquired by an in-vehicle camera.

The block diagram containing the obstacle collision judgment apparatus in the 1st Embodiment of this invention The figure which shows the mathematical model which calculates | requires the distance of the obstacle and the own vehicle which are reflected in the road reflector in the 1st Embodiment of this invention The flowchart for operation | movement description of the obstacle collision judgment apparatus in the 1st Embodiment of this invention Conceptual diagram for explaining the operation of a conventional obstacle collision determination device

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Obstacle collision determination apparatus 2 Image acquisition apparatus 3 Warning apparatus 4 Vehicle 11 Object detection part 12 Distance measurement part 13 Obstacle state recognition part 14 Collision judgment part 15 Road reflector state recognition part 16 Intersection shape memory | storage part

Claims (7)

An apparatus mounted on a vehicle, comprising: an object detection unit that detects an obstacle and a road reflector from image data; and a road reflector state recognition unit that recognizes a state of the road reflector detected by the object detection unit. An obstacle state recognition unit that recognizes the state of the obstacle detected by the object detection unit, and an approach determination unit that determines the approach of the obstacle based on the recognition result of the obstacle state recognition unit, The obstacle collision determination device, wherein the obstacle state recognition means recognizes the state of the obstacle based on the recognition result of the road reflector state recognition means. The state of the obstacle recognized by the obstacle state recognition means is a relative position with respect to the vehicle, and the state of the road reflector recognized by the road reflector state recognition means is based on a relative position and an angle with respect to the vehicle. The obstacle collision determination device according to claim 1, wherein The obstacle collision determination device according to claim 1, wherein the obstacle detected by the object detection means is a four-wheeled vehicle, a two-wheeled vehicle, or a pedestrian. The said object detection means detects presence of the said road reflector based on the continuity of the background in the said image data, or distortion of a background, The any one of Claim 1 thru | or 3 characterized by the above-mentioned. The obstacle collision determination device described. The road reflector state recognition means analyzes the shape of the road reflector and recognizes the state of the road reflector based on the analysis result of the shape. The obstacle collision determination device according to any one of the preceding claims. Intersection shape information storage means for storing intersection shape information of the road on which the vehicle travels is further provided, and the road reflector state recognition means is based on the intersection shape information stored in the intersection shape information storage means. The obstacle collision determination device according to any one of claims 1 to 3, wherein the state of the road reflector is recognized. A system mounted on a vehicle, wherein an image acquisition device that acquires video information outside the vehicle as image data, and a collision of an obstacle with the vehicle is determined by recognizing the image data acquired by the image acquisition device An obstacle collision determination device, and a warning device that warns the vehicle driver of a collision with an obstacle based on the determination result of the obstacle collision determination device.
The obstacle collision determination device includes an object detection unit that detects an obstacle and a road reflector from image data, a road reflector state recognition unit that recognizes a state of the road reflector detected by the object detection unit, An obstacle state recognition means for recognizing the state of the obstacle detected by the object detection means; and an approach determination means for determining the approach of the obstacle based on the recognition result of the obstacle state recognition means, The obstacle state recognition unit recognizes the state of the obstacle based on the recognition result of the road reflector state recognition unit.

Images