JP2010072947A - Obstacle detection device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an obstacle detection device that accurately detects an obstacle on a road on which a vehicle is running without erroneous detection. <P>SOLUTION: When an obstacle is detected on the road on which the vehicle is running on the basis of detection information of a millimeter wave radar 3, which is an environment recognition sensor, a camera unit 4 and a navigation system 5, the position information and the image information of the obstacle identified by use of the detection information are collated with the position information and the image information of an unnecessary operation point stored on an unnecessary operation point database 2. When both are similar to each other, it is determined that no obstacle exists. When both are not similar to each other, it is determined that the obstacle exists. As a result, the obstacle on the road on which the vehicle is running can be accurately detected without false detection. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to an obstacle detection device that detects an obstacle on a road on which a host vehicle travels based on detection information of an environment recognition sensor.

  As devices that detect obstacles on the road on which the host vehicle travels based on detection information of an environment recognition sensor, for example, devices described in Patent Literature 1 and Patent Literature 2 are generally known. Here, in Patent Document 1, in order to accurately recognize the vehicle ahead as an obstacle, image processing is performed on the image region of the camera predicted based on the detection result of the forward search radar, and the vertical direction of the processed image is determined. A technique for detecting a feature amount related to the vehicle width of a preceding vehicle from an edge is disclosed.

On the other hand, in Patent Document 2, in order to accurately detect an object such as a pedestrian as an obstacle from a stationary object such as a guardrail or a utility pole, the position of the stationary object and its camera image are stored in association with each other. When the detected object position and camera image match the stationary object position and camera image stored in the map database by referring to the map database, the detected object is detected as an object such as a pedestrian. Techniques that are excluded from the above are disclosed.
JP 2005-141517 A JP 2007-310572 A

  By the way, in the technology for detecting obstacles on the road on which the host vehicle is traveling with an image processing camera, when the road on which the host vehicle is traveling is a curved road with a guard rail, a plurality of poles supporting the guard rail are provided. By performing image processing as a vertical edge, the pole may be erroneously detected as an obstacle having a predetermined width. Similarly, when an iron plate is laid on the road surface of the road on which the host vehicle is running and there is a road marking in the vicinity, image processing is performed on both sides of the iron plate as vertical edges. May be erroneously detected as an obstacle having a predetermined width.

  Accordingly, an object of the present invention is to provide an obstacle detection device that can accurately detect an obstacle on a road on which the host vehicle travels without erroneously detecting the obstacle.

  An obstacle detection device according to the present invention is provided in a host vehicle equipped with a pre-crash compatible device that operates in accordance with detection information of an environment recognition sensor, thereby detecting an obstacle on a road on which the host vehicle travels. Is an obstacle detection device for detecting based on the detection information of the unnecessary operation point information storage means for storing information on an unnecessary operation point where the pre-crash compatible device operates unnecessarily, detection information of the environment recognition sensor and unnecessary operation It is characterized by comprising obstacle judging means for judging the presence or absence of an obstacle according to the collation result with the stored information relating to the point.

  In the obstacle detection device according to the present invention, when an obstacle on the road on which the vehicle travels is detected based on the detection information of the environment recognition sensor, the detection information of the environment recognition sensor and unnecessary operation point information storage means are stored. The stored information on the unnecessary operating point stored is collated by the obstacle judging means, and if the result of the collation is, for example, the two are close to each other, it is determined that the obstacle does not exist, and the two are close to each other. If not, it is determined that there is an obstacle.

  Here, when there is no obstacle on the road on which the host vehicle is running and the pre-crash compatible device is operating even though the host vehicle is in a normal driving state without sudden deceleration or sudden steering, the operating point is It can be said that it is an unnecessary operating point where the pre-crash compatible device operates unnecessarily. Therefore, in the obstacle detection device of the present invention, it is preferable that the unnecessary operation point is a point where the pre-crash compatible device operates in the normal driving state of the host vehicle.

  Further, since the information related to the unnecessary operation point is information for determining whether or not an obstacle actually exists on the road on which the host vehicle is traveling, the information related to the unnecessary operation point in the obstacle detection device of the present invention is , Position information for specifying the unnecessary operation point and image information around the unnecessary operation point, or position information for specifying the unnecessary operation point and shape information of the obstacle detected at the unnecessary operation point Is preferred.

  Here, it is preferable to provide travel simulation means for obtaining information related to the unnecessary operation point by a travel simulation of the host vehicle offline, since information related to the unnecessary operation point can be efficiently obtained.

  In the obstacle detection device according to the present invention, when an obstacle on the road on which the vehicle travels is detected based on the detection information of the environment recognition sensor, the detection information of the environment recognition sensor and unnecessary operation point information storage means are stored. The stored information on the unnecessary operating point stored is collated by the obstacle judging means, and if the result of the collation is, for example, the two are close to each other, it is determined that the obstacle does not exist, and the two are close to each other. If not, it is determined that there is an obstacle. Therefore, according to the present invention, it is possible to accurately detect an obstacle on the road on which the host vehicle travels without erroneous detection.

  Hereinafter, the best embodiment of an obstacle detection device according to the present invention will be described with reference to the drawings. Here, in the drawings to be referred to, FIG. 1 is a functional block diagram showing the configuration of an obstacle detection apparatus according to an embodiment, and FIG. 2 is a series of processes executed by the obstacle detection ECU shown in FIG. FIG. 3 is a flowchart showing a series of processing steps executed by the obstacle detection ECU as an obstacle determination means.

  An obstacle detection device according to an embodiment is provided on a host vehicle including various pre-crash compatible devices that operate according to detection information of various environment recognition sensors, so that an obstacle on a road on which the host vehicle travels is provided. It is an apparatus that detects an object based on detection information of an environment recognition sensor.

  This obstacle detection device stores travel simulation means for obtaining information on an unnecessary operation point where the pre-crash compatible device operates unnecessarily by an offline travel simulation of the host vehicle, and information on the unnecessary operation point obtained by the travel simulation means. Unnecessary operation point information storage means, and obstacle determination means for determining the presence or absence of an obstacle according to the collation result between the detection information of the environment recognition sensor and the stored information regarding the unnecessary operation point.

  Here, the traveling simulation means and the obstacle determination means are configured by using hardware and software of an obstacle detection ECU (Electronic Control Unit) 1 shown in FIG. An unnecessary operating point database 2 as an operating point information storage means is connected so that information can be exchanged.

  The unnecessary operation point database 2 stores information on unnecessary operation points obtained by the traveling simulation means as described later, that is, position information for specifying the unnecessary operation points and image information in the vicinity of the unnecessary operation points.

  As the various environment recognition sensors for detecting obstacles on the road on which the vehicle travels, the millimeter wave radar 3, the camera unit 4, and the navigation system 5 each output detection information to the obstacle detection ECU 1. It is connected.

  The millimeter wave radar 3 emits a millimeter wave toward the front of the road on which the host vehicle travels, receives the reflected wave, detects the distance, direction and relative speed of the obstacle with respect to the host vehicle, and detects these. An information signal is output to the obstacle detection ECU 1.

  The camera unit 4 captures the front of the road on which the host vehicle travels, and outputs a signal of the image information to the obstacle detection ECU 1. The camera unit 4 has, for example, a structure having a monocular CCD camera, but may have a structure having a night vision camera such as an infrared camera.

  The navigation system 5 includes a GPS (Global Positioning System) receiver and an autonomous navigation device for acquiring information on the current absolute position of the host vehicle, a DVD drive for reading map data, and overlaying at least the current position of the host vehicle on the map. It has a display that displays images.

  Such an obstacle detection ECU 1 is configured to be able to output an obstacle detection signal to a pre-crash safety ECU (Electronic Control Unit) 6. The pre-crash safety ECU 6 includes a pre-crash alarm and a pre-crash brake assist. In addition, it is configured to be able to output predetermined command signals for operating various pre-crash compatible devices 7 generally called pre-crash seat belts, pre-crash brakes and the like.

  The obstacle detection ECU 1 and the pre-crash safety ECU 6 include an input / output interface I / O, an A / D converter, a ROM (Read Only Memory) that stores programs and data, and a RAM (Random Access Memory) that temporarily stores input data and the like. ), A CPU (Central Processing Unit) for executing the program, and the like.

  Here, when there is no obstacle on the road on which the host vehicle travels and the pre-crash handling device 7 operates even though the host vehicle is in a normal driving state without sudden deceleration or sudden steering, the operating point Is an unnecessary operating point where the pre-crash compatible device 7 operates unnecessarily. Therefore, in order to determine whether or not the host vehicle is in a normal driving state, the obstacle detection ECU 1 includes a front-rear G sensor 8 for detecting the sudden deceleration behavior and a steering angle for detecting the sudden steering behavior. Sensors 9 are connected to output detection signals.

  Next, a series of processing procedures executed by the obstacle detection ECU 1 will be described in order along the flowcharts shown in FIGS. Here, the flowchart of FIG. 2 shows a procedure of a series of processes executed by the obstacle detection ECU 1 as the travel simulation unit, and the flowchart of FIG. 3 shows a series of processes executed by the obstacle detection ECU 1 as the obstacle determination unit. The procedure is shown.

  The obstacle detection ECU 1 executes a series of processes shown in the flowchart of FIG. 2 as a traveling simulation unit in order to accumulate information on unnecessary operating points where the pre-crash handling device 7 operates unnecessarily by a traveling simulation of the host vehicle offline. .

  First, in step S10, the obstacle detection ECU 1 acquires environmental data stored in the cache memory as detection information of the millimeter wave radar 3, the camera unit 4, and the navigation system 5.

  In the next step S11, when an obstacle is detected on the road on which the host vehicle travels, the obstacle detection ECU 1 may collide with the obstacle (the collision probability). ) And the time to collision (TTC (Time To Collision)).

  In the subsequent step S12, it is determined whether or not the collision possibility of the obstacle is high based on the calculated collision probability or the collision margin time (TTC). If the collision possibility is high, the pre-crash alarm and the pre-crash The pre-crash compatible device 7 to be operated is selected from the brake assist, the pre-crash seat belt, the pre-crash brake, etc., and the device flag is turned ON.

  In the next step S13, it is determined whether or not the device flag is ON. If the determination result is NO and the device flag is not ON, the process proceeds to return. On the other hand, if the determination result is YES and the device flag is ON, the obstacle detection ECU 1 in step S14 obtains the front and rear G signal detected by the front and rear G sensor 8 and the steering angle signal detected by the steering angle sensor 9. Get from cache memory.

  In the next step S15, the obstacle detection ECU 1 determines whether or not the host vehicle is in a normal driving state without sudden deceleration or sudden steering based on the front / rear G signal and the steering angle signal acquired in step S14. If the determination result is NO and the host vehicle is in a collision avoidance behavior involving sudden deceleration or sudden steering, the pre-crash handling device 7 selected in step S12 operates according to the necessity of collision avoidance. Proceed to return as if to be.

  On the other hand, if the determination result in step S15 is YES and the host vehicle is in a normal driving state, the pre-crash handling device 7 selected in step S12 operates unnecessarily although there is no need for collision avoidance. Proceeding to step S16, it is assumed.

  In step S16, the obstacle detection ECU 1 recognizes the traveling point of the host vehicle when the device flag is turned on in step S12 as an unnecessary operating point where the pre-crash compatible device 7 operates unnecessarily, and the navigation system 5 at the unnecessary operating point. The information on the absolute position of the own vehicle detected by the camera unit and the image information around the unnecessary operation point captured by the camera unit 4 at the unnecessary operation point are acquired from the cache memory.

  In subsequent step S17, the obstacle detection ECU 1 registers the information on the absolute position of the host vehicle acquired as an unnecessary operation point and image information in the vicinity of the unnecessary operation point in the unnecessary operation point database 2 as information on the unnecessary operation point.

  Thereafter, the obstacle detection ECU 1 sequentially registers and accumulates information on unnecessary operating points in the unnecessary operating point database 2 by repeating a series of processes of steps S10 to S17 as a running simulation of the host vehicle.

  On the other hand, the obstacle detection ECU 1 executes a series of processes shown in the flowchart of FIG. 3 as obstacle determination means in order to accurately detect an obstacle on the road while the host vehicle is traveling.

  First, in step S20, the obstacle detection ECU 1 calculates the relative position of the obstacle with respect to the host vehicle based on the detection signal of the millimeter wave radar 3. This relative position is the observed value (l (t), θ (t)) detected by the millimeter wave radar 3 when the position of the obstacle at time t is Pc (xc (t), yc (t)). Xc (t) = l * cos (θ (t)) and yc (t) = l * sin (θ (t)).

  In the next step S21, the obstacle detection ECU 1 obtains an image position corresponding to the relative position of the obstacle calculated in step S20 on the image taken by the camera unit 4, and a predetermined margin (fixed frame) around the position. Alternatively, a rectangular area having a variable frame according to the distance is set on the image. Note that the positional relationship of the image coordinate system with respect to the vehicle coordinate system is obtained in advance by calibration or the like so that calculation for performing such processing can be easily performed.

  In step S22, for example, a vertical edge component is extracted as a feature amount for specifying the width of an obstacle in a rectangular area set on the image. In step S23, the width of the obstacle is estimated by detecting the peak value of the histogram of the edge image based on the extracted vertical edge component.

  In subsequent step S24, an obstacle region is identified based on the relative position of the obstacle with respect to the host vehicle calculated in step S20 and the width of the obstacle estimated in step S22, and the host vehicle may collide with the obstacle. After calculating the property (collision probability) and the time to collision (TTC (Time To Collision)), it is determined whether or not there is a high possibility of an obstacle colliding.

  If the determination result in step S24 is NO and the possibility of obstacle collision is low, the process proceeds to the end and the series of processes is terminated. If the determination result is YES and the possibility of obstacle collision is high, In step S25, the absolute position of the obstacle is calculated.

  In step S25, the obstacle relative position Pc (xc (t), yc (t)) calculated in step S20 is converted into an absolute position Pa (xa (t), ya (t)) as seen in the ground fixed coordinate system. By doing so, the absolute position of the obstacle is calculated. Note that such coordinate conversion processing can be easily performed by estimating the relative positional relationship of the vehicle coordinate system of the host vehicle with respect to the ground fixed coordinate system using the absolute position information detected by the GPS receiver of the navigation system 5. It can be carried out.

  In the next step S26, the obstacle detection ECU 1 detects the absolute position information of the obstacle calculated in step S25 and the absolute position of the host vehicle accumulated as an unnecessary operation point in the unnecessary operation point database 2 (see FIG. 1). By collating with the information, it is determined whether or not the distance between the absolute position of the obstacle and the absolute position of the unnecessary operating point is within a predetermined threshold. That is, it is determined whether or not the absolute position of the obstacle is in the vicinity of the unnecessary operating point.

  If the determination result in step S26 is NO and the absolute position of the obstacle is not in the vicinity of the unnecessary operating point, it is estimated that an obstacle with a high possibility of collision exists on the road on which the host vehicle is traveling. Then, the process proceeds to step S27. Then, the obstacle detection ECU 1 outputs a predetermined command signal to the pre-crash safety ECU 6 in step S27 to appropriately operate the pre-crash handling device 7, and then proceeds to the end and ends the series of processes. Thereby, the collision with an obstacle is avoided beforehand.

  On the other hand, if the determination result in step S26 is YES and the absolute position of the obstacle is in the vicinity of the unnecessary operating point, it means that there is no actual obstacle with high possibility of collision on the road on which the host vehicle is traveling. I can't conclude. Therefore, in the subsequent step S28, the obstacle detection ECU 1 collates the image information of the rectangular area set in step S21 with the image information in the vicinity of the unnecessary operation point stored in the unnecessary operation point database 2, thereby It is determined whether the image information is close to each other.

  In step S28, the degree of coincidence between both pieces of image information is calculated by a method such as correlation calculation or normalized correlation, which is generally known as an image processing technique, and the degree of coincidence is equal to or greater than a predetermined threshold value. It can be done depending on whether or not there is. In this case, if the degree of coincidence of both image information is greater than or equal to the threshold value, it is determined that the two image information items are close to each other. Judge that it is not.

  If the determination result in step S28 is NO and the two pieces of image information are not close to each other, it is estimated that another obstacle with high possibility of collision exists in the vicinity of the unnecessary operation point, and the process proceeds to step S27. As a result, the pre-crash handling device 7 operates appropriately and collision with an obstacle is avoided in advance.

  On the other hand, if the determination result in step S28 is YES and both pieces of image information are close to each other, it is determined that an obstacle with a high possibility of collision does not exist on the road on which the host vehicle travels, and the process ends. Proceed to to end the series of processing.

  As described above, in the obstacle detection device according to the embodiment, when an obstacle on the road on which the vehicle travels is detected based on the detection information of the millimeter wave radar 3 and the camera unit 4, the obstacle is detected. The position information and the image information are collated with the position information of the unnecessary operation point stored in the unnecessary operation point database 2 and the image information near the unnecessary operation point. If the two do not approximate each other, it is determined that an obstacle exists on the road on which the host vehicle travels. If the two approximate each other, it is determined that no obstacle exists. Is done.

  Therefore, an object that is likely to be erroneously detected as an obstacle on the road on which the host vehicle travels, for example, a pillar Pl of a guardrail GR erected along the road side of a right curve road as shown in FIG. 4, or as shown in FIG. Even if there is a laid iron plate SP laid near the road marking Pn, these can be excluded as obstacles that do not exist on the road, and only obstacles that exist on the road can be accurately identified. Can be detected.

  The obstacle detection device according to the present invention is not limited to the above-described embodiment. For example, as another embodiment, the millimeter wave radar 3 shown in FIG. 1 is changed to a laser radar having a higher angular resolution, and a series of processes executed by the obstacle detection ECU 1 as an obstacle determination means are shown in FIG. You may change into the series of processes shown in the flowchart of FIG. 6 slightly different from a flowchart.

  Here, each process of steps S30 to S38 in the flowchart of FIG. 6 as another embodiment corresponds to each process of steps S20 to S28 in the flowchart of FIG. 3 as an embodiment, and steps S31 to S31. Each process of steps S30, S34, S35, S36, and S37 excluding S33 and S38 is substantially the same as each process of steps S20, S24, S25, S26, and S27 in the flowchart of FIG. Therefore, in the flowchart of FIG. 6 as another embodiment, only the processes of steps S31 to S33 and S38 will be described.

  In step S31, a clustering process is performed for collecting point groups that are close in distance with respect to the point group at the relative position of the obstacle calculated in step S30. By this clustering process, a continuous portion of the point cloud data is regarded as a block of obstacles, and when there are a plurality of obstacles separated from each other, the obstacles are separated and extracted.

  In the next step S32, the points located on the outermost side are connected to one obstacle block extracted by the clustering in step S31, and the point sequence data of these points is extracted as the obstacle shape information. At that time, the coordinates of the center of gravity are calculated as a representative value representing the position of the obstacle. In the subsequent step S33, the distance in the vertical direction with respect to the host vehicle is estimated as the width of the obstacle from the point sequence data extracted as the obstacle shape information in step S32.

  In step S38, the shape information of the obstacle extracted in step S32 and the shape information of the object near the unnecessary operation point stored in the unnecessary operation point database 2 are collated, so that the shape information of both is mutually It is determined whether or not they are approximate.

  In such another embodiment of the present invention, when an obstacle is detected based on the detection information of the laser radar, the position information and shape information of the obstacle are accumulated in the unnecessary operation point database 2. If the location information of the point and the shape information of the object near the unnecessary operating point are collated and the shape information of both is not close to each other, it is determined that an obstacle exists on the road on which the host vehicle is traveling. If the two pieces of shape information are close to each other, it is determined that the obstacle does not exist.

  Therefore, it becomes possible to eliminate the pillars Pl of the guard rail GR as shown in FIG. 4 and the iron plate SP near the road marking Pn as shown in FIG. 5 as obstacles that do not exist on the road. Only obstacles existing on the road can be accurately detected.

  Here, in the obstacle detection device according to one embodiment and the other embodiments, the image information in the vicinity of the unnecessary operating point photographed by the camera unit 4 includes daytime image information using natural light as a light source, a headlamp and the like as a light source. This is very different from the nighttime image information. Therefore, in order to enable accurate verification of image information in the vicinity of unnecessary operating points regardless of day and night, image information of two patterns of day and night is stored in the unnecessary operating point database 2 according to the time at the time of verification. Thus, the image information of two patterns of day and night may be switched and referred to.

It is a functional block diagram which shows the structure of the obstruction detection apparatus which concerns on one Embodiment of this invention. It is a flowchart which shows a series of processing procedures which the obstruction detection ECU shown in FIG. 1 performs as a travel simulation means. It is a flowchart which shows a series of processing procedures which the obstacle detection ECU performs as an obstacle determination means. It is a sketch of the road which shows the support | pillar of the guardrail which is an example which is easy to be erroneously detected as an obstacle on a road. It is a sketch of the road which shows the laying iron plate which is another example which is misdetected as an obstacle on a road. It is a flowchart which shows a series of processing procedures which the obstacle detection ECU performs as an obstacle determination means of other embodiment.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Obstacle detection ECU, 2 ... Unnecessary operation point database, 3 ... Millimeter wave radar, 4 ... Camera unit, 5 ... Navigation system, 6 ... Pre-crash safety ECU, 7 ... Pre-crash correspondence device, 8 ... Front-rear G sensor, 9: Steering angle sensor.

Claims (5)

An obstacle for detecting an obstacle on the road on which the vehicle travels based on the detection information of the environment recognition sensor by being provided in the host vehicle equipped with a pre-crash compatible device that operates according to the detection information of the environment recognition sensor An object detection device comprising:
Unnecessary operation point information storage means for storing information on an unnecessary operation point where the pre-crash compatible device operates unnecessarily;
An obstacle detection device comprising: obstacle determination means for determining presence or absence of the obstacle according to a result of collation between detection information of the environment recognition sensor and stored information regarding the unnecessary operation point.   The obstacle detection apparatus according to claim 1, wherein the unnecessary operation point is a point where the pre-crash handling device operates in a normal driving state of the host vehicle.   The obstacle detection device according to claim 1, wherein the information related to the unnecessary operation point is position information for specifying the unnecessary operation point and image information around the unnecessary operation point.   The obstacle detection apparatus according to claim 1, wherein the information related to the unnecessary operation point is position information for specifying the unnecessary operation point and shape information of an obstacle detected at the unnecessary operation point. The obstacle detection device according to any one of claims 1 to 4, further comprising travel simulation means for obtaining information on the unnecessary operation point by a travel simulation of the host vehicle offline.

Images