JP2007164671A - Device for deciding approaching obstacle and system for warning collision with obstacle - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a device for deciding approaching obstacle and a system for warning collision with obstacle by which an obstacle on a traveling route is recognized in an early stage by reducing the processing amount of image recognition. <P>SOLUTION: The device has: a recognition processing part 102 which recognizes the obstacle from image data; a road shape analysis part 106 which analyzes the road shape on which a vehicle 400 travels; a traveling state acquisition part 107 which acquires the traveling state of the vehicle 400; a moving destination prediction part 105 which predicts the moving destination of the vehicle 400 on the basis of an analysis result by the road shape analysis part 106 and an acquisition result by the traveling state acquisition part 107; a recognition area specification part 104 which specifies an area to be recognized by the recognition processing part 102 on the basis of a prediction result by the moving destination prediction part 105; and an approach deciding part 103 which decides approach of the obstacle on the basis of the recognition result by the recognition processing part 102 in the area specified by the recognition area specification part 104, and recognizes the obstacle on the traveling route on the basis of moving destination prediction information on the own vehicle. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to an obstacle approach determination device and an obstacle collision warning system that determine the approach of a vehicle and an obstacle by recognizing acquired image data.

  Conventionally, various obstacle approach determination devices and obstacle collision warning systems that support safe driving have been proposed in order to prevent a collision accident during driving of an automobile. One of such obstacle approach determination devices and obstacle collision warning systems is a vehicle for preventing a collision accident between the own vehicle and an obstacle such as another vehicle or a pedestrian (hereinafter referred to as an obstacle). There is a camera that recognizes the surroundings of the vehicle by recognizing the image data obtained by photographing the state outside the vehicle with a camera mounted on the vehicle, and judges the approach between the vehicle and an obstacle and warns the driver.

  However, the recognition processing for recognizing an obstacle from the acquired image data increases the amount of processing dramatically according to the size of the area where the obstacle is searched, and determines the approach between the vehicle and the obstacle in real time. It is difficult. Therefore, how to identify an area for searching for obstacles is a major issue for real-time recognition processing.

  A conventional obstacle approach determination device that has tackled the above-described problems is mounted on a vehicle and inputs image data obtained by photographing a state around the vehicle as image information, and a single or a plurality of devices existing outside the vehicle. Distance measuring means for obtaining distance and direction information to an object, judgment means for judging and setting an image processing method based on distance and direction information obtained by the distance measuring means, and image information input by an image sensor Image processing means for performing image processing based on the image processing method set by the determination means. The determination means determines, for example, a process for determining a region for performing image processing and a target for performing image processing. An image processing method including at least one of the process and the process of enlarging, reducing, or changing the direction of a specific object in the image is determined and set. Furthermore, a measuring means for detecting the traveling state of the host vehicle is provided, and the judging means determines an object to be subjected to image processing according to the detection result of the measuring means, and the distance and direction to an object outside the vehicle. The area which processes the image acquired based on this information and the information of the vehicle speed sensor and the rudder angle sensor is specified.

For example, when an area for processing an image acquired based on the information of the rudder angle sensor is specified, as shown in FIG. 9, when there is no rudder angle change detected by the rudder angle sensor, the acquired image area 9a Among them, the region 9b in front of the vehicle is specified as the image processing region, and when there is a steering angle change detected by the steering angle sensor, the region 9c in the direction corresponding to the detected amount is specified as the image processing region (for example, , See Patent Document 1).
JP-A-6-124340

  However, in the conventional obstacle approach determination device, the image processing area is specified based on the information detected by the vehicle speed sensor and the rudder angle sensor acquired as the traveling state of the host vehicle, and the image processing area is limited. For example, if you are turning right or left at the next intersection where the road you are currently driving is approaching on a straight road, or you are approaching a curve, Since there is no steering angle change detected by the steering angle sensor because the current travel path is straight, the front of the vehicle is specified as the image processing area.

For this reason, even if there is an obstacle at the left or right turn or corner of the intersection where the host vehicle will run, it will not be the target of the image processing area, so the obstacle cannot be recognized. There was a problem that the approach judgment and the warning to the driver were delayed.

  The present invention has been made to solve such a conventional problem, and reduces the processing amount of image recognition by specifying the destination in the image processing area based on the destination prediction information of the host vehicle. In addition, an object of the present invention is to provide an obstacle approach determination device and an obstacle collision warning system that can recognize obstacles existing on a travel route at an early stage.

  The obstacle approach determination device according to the present invention is an apparatus mounted on a vehicle, and includes obstacle recognition means for recognizing an obstacle from image data, road shape analysis means for analyzing a road shape on which the vehicle travels, and vehicle A travel state acquisition unit that acquires the travel state of the vehicle, a destination prediction unit that predicts a destination of the vehicle based on an analysis result in the road shape analysis unit and an acquisition result in the travel state acquisition unit, and a destination prediction unit Based on the prediction result, the recognition unit for identifying the area where the obstacle recognition unit performs the recognition process, and the approach of the obstacle based on the recognition result by the obstacle recognition unit in the area specified by the recognition area specification unit. It has a configuration provided with an obstacle approach determining means for determining.

  With this configuration, the travel destination of the host vehicle is predicted based on the road shape on which the vehicle travels and the travel state of the vehicle, the travel destination of the host vehicle is specified as a recognition area, and obstacles existing in the specified recognition area are identified. It is possible to recognize and determine the approach of the vehicle and the obstacle.

  The obstacle approach determination device of the present invention further includes a traveling lane recognition unit that detects a lane in which the vehicle is traveling and calculates a curvature of the lane, and the road shape analysis unit is based on a calculation result of the traveling lane recognition unit. The road shape is analyzed.

  With this configuration, the curvature of the travel lane in which the vehicle travels can be calculated, and the road shape can be analyzed from the curvature of the travel lane.

  Also, the obstacle approach determining device of the present invention is a speed indicator that detects the speed of a vehicle, a steering angle detector that detects a steering angle of a vehicle, or a direction indicator that detects a presentation state of a vehicle direction indicator. At least one of the state detection means is further provided, and the traveling state acquisition means is a traveling state of the vehicle based on at least one detection result of the speed detection means, the steering angle detection means, or the direction indicator state detection means. It has the structure which acquires.

  With this configuration, the traveling state of the vehicle can be acquired from at least one of the vehicle speed, the steering angle of the vehicle, or the presentation state of the direction indicator.

  Further, the obstacle collision warning system of the present invention is a system mounted on a vehicle, an image acquisition device that acquires video information outside the vehicle as image data, and a recognition process based on the image data acquired by the image acquisition device. An obstacle approach judging device for judging the approach of the obstacle to the vehicle and a warning device for warning the vehicle driver of the collision with the obstacle based on the judgment result of the obstacle approach judging device. The obstacle approach determination device includes obstacle recognition means for recognizing an obstacle from image data, road shape analysis means for analyzing a road shape on which the vehicle travels, traveling state acquisition means for acquiring a traveling state of the vehicle, Destination prediction means for predicting the destination of the vehicle based on the analysis result in the road shape analysis means and the acquisition result of the traveling state acquisition means, and the obstacle recognition means on the basis of the prediction result of the destination prediction means Recognizing area specifying means for specifying an area for performing recognition processing, and obstacle approach determining means for determining the approach of an obstacle based on the recognition result of the obstacle recognizing means in the area specified by the recognition area specifying means It has a configuration.

  With this configuration, video information outside the vehicle is acquired as image data, the destination of the host vehicle is predicted based on the shape of the road on which the host vehicle travels and the traveling state of the host vehicle, and the movement of the host vehicle in the acquired image data Identify the destination in the recognition area, recognize obstacles in the identified recognition area, determine the approach of the vehicle and the obstacle, and drive the vehicle against the collision between the vehicle and the obstacle based on the determination result Can be warned.

  The obstacle approach determination device according to the present invention is an apparatus mounted on a vehicle, and includes obstacle recognition means for recognizing an obstacle from image data, road shape analysis means for analyzing a road shape on which the vehicle travels, and vehicle A travel state acquisition unit that acquires the travel state of the vehicle, a destination prediction unit that predicts a destination of the vehicle based on an analysis result in the road shape analysis unit and an acquisition result in the travel state acquisition unit, and a destination prediction unit Based on the prediction result, the recognition unit for identifying the area where the obstacle recognition unit performs the recognition process, and the approach of the obstacle based on the recognition result by the obstacle recognition unit in the area specified by the recognition area specification unit. By having a configuration with an obstacle approach judging means for judging, the destination of the own vehicle is predicted based on the road shape on which the own vehicle runs and the running state of the own vehicle, and the destination of the own vehicle is recognized. As well as reduce the processing amount of image recognition by identifying the area, early recognize an obstacle that exists on the travel route has the effect that it is possible to determine the approach of the own vehicle and the obstacle.

  Furthermore, the obstacle collision warning system of the present invention is a system mounted on a vehicle, which is a system mounted on a vehicle, and acquires an image information outside the vehicle as image data, and an image acquisition device The obstacle approach determination device that determines the approach of the obstacle to the vehicle based on the image data acquired by the vehicle, and a warning that warns the vehicle driver of the collision with the obstacle based on the determination result of the obstacle approach determination device The obstacle approach determination device includes an obstacle recognition unit that recognizes an obstacle from image data, a road shape analysis unit that analyzes a road shape on which the vehicle travels, and a travel that acquires a travel state of the vehicle. Based on the state acquisition means, the destination prediction means for predicting the destination of the vehicle based on the analysis result of the road shape analysis means and the acquisition result of the travel state acquisition means, and the prediction result of the destination prediction means A recognition area specifying means for specifying an area where the obstacle recognition means performs recognition processing, and an obstacle approach for determining the approach of the obstacle based on the recognition result of the obstacle recognition means in the area specified by the recognition area specifying means By having a configuration including a determination unit, image information outside the vehicle is acquired as image data, and the destination of the host vehicle is predicted and acquired based on the road shape on which the host vehicle travels and the traveling state of the host vehicle. In addition to reducing the amount of image recognition processing by identifying the destination of the host vehicle in the recognition area of the image data that has been processed, it is possible to recognize obstacles existing on the travel route at an early stage, and to approach the host vehicle and the obstacle. It has the effect that it can judge and warn the driver | operator of the collision with an obstruction based on the judgment result.

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

(Embodiment 1)
FIG. 1 is a block diagram showing a configuration of an obstacle approach determination device according to Embodiment 1 of the present invention.

  In FIG. 1, an obstacle approach determination device according to the present embodiment is mounted on a vehicle, and includes an image data input unit 101, a recognition processing unit 102 as an obstacle recognition unit, and an approach determination unit as an obstacle approach determination unit. 103, a recognition area specifying unit 104, a destination prediction unit 105, a road shape analysis unit 106, a travel state acquisition unit 107, a warning data output unit 108, a travel lane recognition unit 109, and a speed detection unit 110 The steering angle detection unit 111 and the direction indicator state detection unit 112 are configured.

  For example, the image data input unit 101 acquires image data in front of the vehicle from an external device such as a television camera mounted on the vehicle.

  The recognition processing unit 102 recognizes an obstacle present in the image data acquired by the image data input unit 101. This recognition process is not performed on the entire acquired image data but limited to the area specified by the recognition area specifying unit 104. Therefore, the specific area to be recognized is inquired of the recognition area specifying unit 104, and the area information of the specific area to be recognized is obtained as a response.

  The method for recognizing the obstacle in the recognition processing unit 102 is not particularly limited. For example, characteristic information such as vehicle-likeness or pedestrian-likeness is registered in advance as a database, and the specified area in the acquired image data This is possible by searching for an object having the same characteristics as the registered database.

  As feature information representing the vehicle-likeness and the pedestrian-likeness, object shape information can be used. For example, the vehicle-like shape information is a shape in which a brake lamp circle is arranged symmetrically on a quadrilateral, and the pedestrian-like shape information is a shape with a rectangular body under a circular head. These may be registered in the database, and the presence of the same object as the shape information may be searched from the specified area in the acquired image data.

  Moreover, a brightness | luminance change can be utilized as characteristic information of a vehicle or a pedestrian. By learning in advance the characteristic luminance changes of vehicles and pedestrians using a plurality of images, each of them is registered in a database. For example, in the case of a vehicle, the body portion has a high luminance and the tire portion has a luminance. Becomes lower. This characteristic luminance change may be determined using a plurality of images.

  Further, the recognition processing unit 102 estimates the distance from the own vehicle by recognizing the size of other vehicles and pedestrians in the image because the size is generally almost constant. The distance from the own vehicle to the obstacle is estimated, and the result is output to the approach determination unit 103 as a recognition result.

  The approach determination unit 103 determines the approach of the obstacle based on the recognition result of the recognition processing unit 102, that is, the obstacle and the distance between the host vehicle and the obstacle that change from moment to moment. Warning data is output from the warning data output unit 108 to an external alarm device or the like.

  There is no particular limitation on the method for determining the approach of the obstacle in the approach determining unit 103. For example, the distance from the own vehicle to the obstacle obtained by the recognition processing unit 102 and the speed detection mounted on the own vehicle are detected. Based on the speed information from the unit 109, it is possible to calculate the time until the host vehicle reaches the obstacle.

  The warning data output from the approach determination unit 103 may output information indicating how many meters from the present time the distance to the obstacle is. For example, the number of distances after 1 second, 2 seconds, and 3 seconds from the present time may be output as many as the number of obstacles that detected the distance.

  The recognition area specifying unit 104 specifies an area where recognition processing is to be performed, and notifies the recognition processing unit 102 of the area. For identifying the recognition area, the destination prediction information of the host vehicle predicted by the destination prediction unit 105 is used. In order to obtain the destination prediction information of the host vehicle, the destination prediction unit 105 is inquired and the destination prediction information is obtained as a response.

  The destination prediction unit 105 predicts the destination of the host vehicle based on the road shape information and the running state information, and notifies the recognition area specifying unit 104 of the result. In order to predict the destination of the host vehicle, the road shape analysis unit 106 and the travel state acquisition unit 107 are inquired to obtain road shape information and travel state information as responses.

  The road shape analysis unit 106 uses the image data acquired by the image data input unit 101 based on the calculation result of the travel lane recognition unit 109 that calculates the curvature of the lane on which the host vehicle travels, Road shape information relating to the shape of the road on which the vehicle travels, such as the shape of an intersection, is generated.

  The traveling state acquisition unit 107 includes speed information from the speed detection unit 110 that detects the speed of the host vehicle, steering angle information from the steering angle detection unit 111 that detects the steering angle of the host vehicle, and a direction indicator of the host vehicle. Based on the direction indicator presenting state information from the direction indicator state detecting unit 112 that detects the present state of the vehicle, traveling state information relating to the traveling state of the host vehicle is generated.

  Here, an example in which the destination prediction unit 105 predicts the destination of the host vehicle will be described with reference to FIG.

  FIG. 2 shows the relationship between the analysis result of the road shape analysis unit 106, the acquisition result of the traveling state acquisition unit 107, and the recognition area specified by the recognition area specification unit 104.

  As a result of the analysis by the road shape analysis unit 106, when the road shape is a straight road, the center front in the traveling direction is specified as the recognition area.

  When the road shape ahead is a curve, the previous point that traveled along the curve is specified as a recognition area.

  When the road shape in front is an intersection, it is not known which road will enter the intersection road alone, so the destination of the host vehicle is predicted with reference to the traveling state information. That is, the content presented by the direction indicator is confirmed, and when the direction indicator is put out to the right, the right turn destination is specified as the recognition area.

  For example, as shown in FIG. 3, when the direction indicator presents the right turn 3a before the intersection, the right turn destination is specified as the recognition area 3c in the acquired image area 3b.

  Even if the direction indicator is not presented, the future destination of the host vehicle is predicted from the history of the speed information and the steering angle information of the host vehicle so far. For example, in the case of lane change or decelerating traveling so as to approach the right direction in the traveling lane, the right turn destination is specified as the recognition area. Similarly, the left turn destination is specified as the recognition area in the case of lane change or decelerating traveling so as to move leftward in the traveling lane.

  Note that the method of predicting the destination of the host vehicle in the destination prediction unit 105 is not limited to the above, and the road shape information and the driving state information are comprehensively determined, and the probability of right turn and left turn is calculated. In such a configuration, the recognition area may be specified according to the calculated probability, or another method may be used.

  Next, a specific example of calculating the road curvature in the travel lane recognition unit 109 will be described.

  In the calculation of the curvature of the road, a straight line is conventionally detected by using a Hough transform, which is a method for converting between a graphic space and a parameter space.

  That is, for an arbitrary point in the image data (graphic space), a line segment in the image data can be extracted by listing a line segment candidate that can be a constituent element (voting to the parameter space).

  However, the conventional method for calculating the curvature of the road using the Hough transform is to use the straight line 10a detected by using the Hough transform from the road plane of the driving lane 1001 as shown in FIG. Although it is an effective means for specifying, since the recognition area 10b is set in the central portion of the image, it is not suitable as a means for specifying the area of the destination of the travel route targeted by the present invention as the recognition area.

In the present embodiment, when the traveling lane 401 is curved as shown in FIG. 4, in order to identify the recognition area, a circular component is detected by the Hough transform and the curvature of the road is calculated. Specifically, a plurality of circular equations are applied to the edge points constituting the left and right white lines of the road (the driving lane 401) while changing the coefficients. Identify the circle passing through the more points for the right and left white lines, the identified circle 4a, the radius r _a of _4b, calculates r _b, for these resulting plurality of circular, and its center coordinates and the radius By obtaining the relationship, the road shape can be estimated.

  For example, in the case of the obtained circle, if the radius of the circle with the center coordinates relatively to the right is larger than the circle with the center coordinates relatively to the left, the road shape curves to the left. Can be judged. Conversely, if the radius of the circle with the center coordinates relatively to the left is larger than the circle with the center coordinates relatively to the right, the road shape is judged to curve to the right. The curve destination can be specified in the recognition area 4c.

  At the intersection shown in FIG. 5, in order to recognize the branching situation of the intersection, the number of road branches and the angle of the branch are calculated by detecting the elliptical shape of the quadratic curve using the Hough transform. can do.

  Specifically, as in the case of calculating the curvature of the curve, each point in the image is voted for a quadratic curve equation. At this time, the points in the image are voted while rotating and translating the quadratic curve and changing the shape of the quadratic curve. Thus, when a plurality of elliptical shapes 5a, 5b, 5c, 5d, and 5e of a quadratic curve are obtained, the number of branched paths (4) is changed from the number of elliptical shapes (5) to the rotational direction of the elliptical shape. The branching direction can be calculated from the elliptical shape, and the branching angle can be calculated from the elliptical shape. Based on the intersection information and the traveling state information obtained in this way, the destination of the host vehicle can be predicted.

  Note that the method for calculating the curvature of the road in the travel lane recognition unit 109 is not limited to the above, and other calculation methods may be used.

  The operation of the obstacle collision determination device configured as described above will be described with reference to the flowchart of FIG.

  First, the image data input unit 101 acquires image data obtained by photographing the front of the host vehicle from an external device (S601).

  Next, when receiving the image data from the image data input unit 101, the recognition processing unit 102 inquires of the recognition area specifying unit 104 about an area to be recognized (S602).

  The recognition area specifying unit 104 inquires about the future destination of the host vehicle to the destination prediction unit 105 in order to specify the area to be recognized (S603).

  The destination prediction unit 105 starts collecting various information in order to predict the destination of the host vehicle (S604).

  First, the road shape analysis unit 106 is inquired to collect road shape information related to the shape of the road on which the vehicle travels (S605). In addition, the traveling state acquisition unit 107 is also inquired to collect traveling state information relating to the traveling state of the host vehicle (S606).

  Next, based on the collected road shape information and traveling state information, the destination prediction unit 105 predicts the future destination of the host vehicle, and sends the predicted destination prediction information to the recognition area specifying unit 104. A response is made (S607).

  The recognition area specifying unit 104 specifies an area to be recognized based on the destination prediction information from the destination prediction unit 105. The identification result of this area is returned to the recognition processing unit 102 (S608).

The recognition processing unit 102 performs a process of recognizing an obstacle present in the area specified by the recognition area specifying unit 104 in the image data. The information on the recognized obstacle is sent to the approach determination unit 103 (S609).
The approach determination unit 103 determines the approach between the obstacle recognized by the recognition processing unit 102 and the host vehicle (S610).

  If there is a possibility of collision between the obstacle and the host vehicle as a result of the determination by the approach determination unit 103 (YES in S611), warning data is output from the warning data output unit 108 to another device (S612). After outputting the warning data or when there is no possibility of collision between the obstacle and the host vehicle (NO in S611), it is determined whether or not to continue the above operation (S613), and when the operation is continued (in S613) (YES) returns to step S601. If the operation is not continued (NO in S613), the operation is terminated.

  The above-described operation may be coordinated so as to start at the same time as starting the engine of the host vehicle and end at the same time as stopping the engine, for example.

  As described above, according to the obstacle collision determination device of the present embodiment, the movement destination of the host vehicle is predicted based on the road shape on which the host vehicle travels and the traveling state of the host vehicle, and among the acquired image data, By identifying the destination of the vehicle in the recognition area, the amount of image recognition processing is reduced, and obstacles existing on the travel route are recognized at an early stage. In some cases, warning data can be output.

  In the above embodiment, the destination of the host vehicle is predicted based on the shape of the road on which the host vehicle travels and the driving state of the host vehicle, and the destination of the host vehicle in the acquired image data is used as the recognition area. Although described as being specified, the present invention is not limited to this, and may be linked to a car navigation device mounted on the host vehicle.

  For example, a map database in a car navigation device may be used, or a destination is comprehensively predicted using a combination of the image data recognition process and the map database described in this embodiment, and an approach determination is made. You may make it do.

  Further, instead of acquiring the driving state information in the driving state acquisition unit 107, route information set in the car navigation device may be used.

  Further, in addition to the vehicle speed information, the steering angle information, and the direction indicator presenting information that are the current driving state information, the route information of the car navigation device may be used for comprehensive determination.

(Embodiment 2)
FIG. 7 is a block diagram showing the configuration of the obstacle collision warning system according to Embodiment 2 of the present invention.

  In FIG. 7, the obstacle collision warning system according to the present embodiment has a configuration in which an image acquisition device 200, an obstacle approach determination device 100, and a warning device 300 are mounted on a vehicle 400.

  The image acquisition device 200 includes a camera unit 201 and an image data output unit 202 that outputs image data captured by the camera unit 201 to the obstacle approach determination device 100.

  Since the configuration of the obstacle approach determination device 100 is the same as that in FIG. 1 described in the first embodiment, the image data input unit 101 that receives image data from the image acquisition device 200 and the warning that outputs warning data to the warning device 300. Only the data output unit 108 is described, and the others are omitted.

  The warning device 300 includes a warning data input unit 301 that receives warning data from the obstacle approach determination device 100 and a speaker unit 302 that issues a warning based on the warning data received by the warning data input unit 301.

  The camera unit 201 of the image acquisition apparatus 200 captures the outside of the vehicle 400 in the traveling direction with a CCD camera or the like that can capture the day and night.

  The image data output unit 202 of the image acquisition device 200 outputs an image captured by the camera unit 201 to the obstacle approach determination device 100 as image data.

  The image data input unit 101 of the obstacle approach determination device 100 receives the image data sent from the image data output unit 202 of the image acquisition device 200.

  The warning data output unit 108 of the obstacle approach determining device 100 outputs warning data to the warning device 300 when there is a possibility of collision between the host vehicle and the obstacle based on the result processed in the obstacle approach determining device 100. Output to.

  The warning data input unit 301 of the warning device 300 receives the warning data sent from the warning data output unit 108 of the obstacle approach determination device 100.

  The speaker unit 302 of the warning device 300 outputs a warning to the driver of the vehicle 400 by voice based on the received warning data.

  Data transmission / reception among the image acquisition device 200, the obstacle approach determination device 100, and the warning device 300 is performed wirelessly or by wire.

  About the obstacle collision warning system comprised as mentioned above, the operation | movement is demonstrated using the flowchart of FIG.

  First, the camera unit 201 of the image acquisition device 200 captures the outside of the vehicle ahead in the traveling direction (S801), outputs the captured image as image data from the image data output unit 202 to the obstacle approach determination device 100, and approaches the obstacle. Obtained by the image data input unit 101 of the determination apparatus 100 (S802).

  Hereinafter, the operation from the recognition specific area inquiry (S803) to the warning data output (S813) is the same as the recognition specific area inquiry (S602) to the warning data output (S612) shown in FIG.

  Next, based on the warning data output in step S813, the speaker unit 302 of the warning device 300 notifies the driver of the vehicle 400 of a warning that there is a possibility of a collision between the obstacle and the vehicle 400 (S814). ).

  After the warning notification or when there is no possibility of collision between the obstacle and the vehicle 400 (NO in S812), it is determined whether or not to continue the above operation (S815), and the operation is continued (YES in S815). Returns to step S801. If the operation is not continued (NO in S815), the operation is terminated.

  The above-described operation may be coordinated so as to start at the same time as starting the engine of the host vehicle and end at the same time as stopping the engine, for example.

  As described above, according to the obstacle collision warning system of the present embodiment, based on the image acquisition device that acquires video information outside the vehicle as image data, the road shape on which the host vehicle travels, and the traveling state of the host vehicle. An obstacle approach determination device that predicts the destination of the host vehicle, identifies the destination of the host vehicle as a recognition area, determines an approach between the obstacle and the host vehicle in the recognition area, and outputs warning data; By providing a warning device that warns the driver of a collision with an obstacle, the amount of image recognition processing is reduced, and obstacles existing on the travel route are recognized at an early stage so that the vehicle approaches the obstacle. If there is a possibility of a collision, a warning can be promptly notified to the driver.

  In the above-described embodiment, the camera unit 201 of the image acquisition device 200 has been described as photographing the front in the traveling direction of the vehicle. However, the present invention is not limited to this, and the side and rear of the vehicle can be viewed. Even if the object is to be photographed, it is possible to determine an obstacle present on the side or rear of the vehicle and notify the driver of a warning.

  Further, although the warning device 300 has been described as a notification of a collision warning between a vehicle and an obstacle by voice, the warning of a collision may be displayed on a display or the like with characters or images. You may make it use together. Also, other methods such as ringing and vibration can be similarly implemented.

  As described above, the obstacle approach determination device and the obstacle collision warning system according to the present invention reduce the processing amount of image recognition by specifying the movement destination of the host vehicle in the recognition area among the image data acquired by the camera. Obstacle approach determination device having the effect of recognizing obstacles existing on the traveling route at an early stage and promptly warning the vehicle driver of a collision between the host vehicle and the obstacles. It is also useful as an obstacle collision warning system.

The block diagram which shows the structure of the obstruction approach judgment apparatus in Embodiment 1 of this invention. The figure explaining the destination prediction of the own vehicle in Embodiment 1 of this invention The figure explaining the movement prediction destination in the intersection in Embodiment 1 of this invention The figure explaining the calculation of the curvature in the curve in Embodiment 1 of this invention The figure explaining the driving | running | working lane recognition operation | movement in the intersection in Embodiment 1 of this invention. The flowchart which shows operation | movement of the obstruction approach judgment apparatus in Embodiment 1 of this invention. The block diagram which shows the structure of the obstacle collision warning system in Embodiment 2 of this invention. The flowchart which shows operation | movement of the obstacle collision warning system in Embodiment 2 of this invention. The figure explaining the conventional image recognition area The figure explaining the driving lane recognition with the conventional curve

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 Obstacle approach judgment apparatus 101 Image data input part 102 Recognition processing part 103 Approach judgment part 104 Recognition area specific | specification part 105 Destination prediction part 106 Road shape analysis part 107 Travel condition acquisition part 108 Warning data output part 109 Travel lane recognition part 110 Speed detection unit 111 Steering angle detection unit 112 Direction indicator state detection unit 200 Image acquisition device 201 Camera unit 202 Image data output unit 300 Warning device 301 Warning data input unit 302 Speaker unit 400 Vehicle 401 Traveling lane

Claims (4)

An apparatus mounted on a vehicle, comprising obstacle recognition means for recognizing an obstacle from image data, road shape analysis means for analyzing a road shape on which the vehicle travels, and a travel state for acquiring the travel state of the vehicle Based on the acquisition means, the destination prediction means for predicting the destination of the vehicle based on the analysis result in the road shape analysis means and the acquisition result in the traveling state acquisition means, and the prediction result of the destination prediction means Recognition area specifying means for specifying an area where the obstacle recognition means performs recognition processing, and approaching the obstacle based on the recognition result of the obstacle recognition means in the area specified by the recognition area specifying means. An obstacle approach judgment device comprising obstacle approach judgment means for judging. The vehicle further comprises traveling lane recognition means for detecting a lane in which the vehicle is traveling and calculating a curvature of the lane, and the road shape analysis means analyzes the road shape based on the calculation result of the traveling lane recognition means. The obstacle approach determination device according to claim 1, wherein At least one of speed detection means for detecting the speed of the vehicle, steering angle detection means for detecting the steering angle of the vehicle, and direction indicator state detection means for detecting a presentation state of the direction indicator of the vehicle. In addition, the traveling state acquisition unit acquires the traveling state of the vehicle based on at least one detection result of the speed detection unit, the steering angle detection unit, or the direction indicator state detection unit. The obstacle approach determination device according to claim 1, wherein: An image acquisition device for acquiring video information outside the vehicle as image data, and an obstacle for determining the approach of the obstacle to the vehicle based on the image data acquired by the image acquisition device. An approach determination device, and a warning device that warns a driver of the vehicle of a collision with the obstacle based on a determination result in the obstacle approach determination device. In the obstacle recognition means for recognizing the obstacle from the data, the road shape analysis means for analyzing the road shape on which the vehicle travels, the travel state acquisition means for acquiring the travel state of the vehicle, and the road shape analysis means A destination prediction unit that predicts a destination of the vehicle based on an analysis result and an acquisition result of the travel state acquisition unit, and the obstacle recognition unit based on a prediction result of the destination prediction unit Recognition area specifying means for specifying an area for performing recognition processing, and obstacle approach determining means for determining the approach of the obstacle based on the recognition result of the obstacle recognition means in the area specified by the recognition area specifying means An obstacle collision warning system characterized by comprising:

Images