JP2008070206A - Object recognizizing apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To allow an object recognizing apparatus for determining that an object does not exist on a road in the recognition range, if an object whose recognition makes it possible to consider the object conditions on a road for its recognition as being proper cannot be recognized in the recognition range. <P>SOLUTION: In the recognition range of the road to be set based on a position determined, by positioning of own vehicle, the position of the object on the road stored in a map database and supposed to be recognized by the vehicle, and the accuracy of positioning for the vehicle position, a processing for recognizing the object is performed. Information representing mark blurring on roads of the object is stored in the map database; and if the extent of mark blurring of the object by the map database is proper for recognizing the object, when the object is not recognized in the set recognition range, it is determined that the object did not exist in the set recognition range. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to an object recognition apparatus, and more particularly, to an object recognition apparatus suitable for recognizing an object in a road recognition range based on a vehicle position, an object position on a road, and positioning accuracy.

2. Description of the Related Art Conventionally, an object recognition apparatus that recognizes a stationary object such as a road sign in a vehicle that is a moving body is known (see, for example, Patent Document 1). This object recognition apparatus includes a camera that captures a road around a vehicle, and recognizes a stationary object on the road by processing a captured image captured by the camera.
JP 2000-346653 A

  In general, stationary objects recognized from camera-captured images as described above are road signs and road markings that are installed or drawn on a road, and are scattered on the road. In addition, the position of the stationary object is often determined absolutely or relative to a fixed object on a map such as an intersection. In this respect, on the vehicle side having a database that stores in advance the position information of the stationary object to be recognized by the vehicle, when the object on the road is recognized, the road around the vehicle is always photographed and the captured image thereof As a result, the recognition processing load increases. As a result, it takes time to recognize an object, or an inconvenience is required in which an expensive device capable of high-speed processing is required.

  In order to avoid such inconvenience, it is conceivable to set a recognition range in which the object is to be recognized from the captured image in advance, and to perform processing for recognizing the object only within the recognition range. According to this method, it is possible to recognize an object on the road while reducing its processing load.

  By the way, the recognition range for recognizing the object is set based on the position of the object stored in the database of the vehicle, but the position information of the object stored in the database is caused by road repair or the like. It is different from the real thing, that is, the object whose position information is stored in the database may be erased from the real road and may not be drawn. In this case, it is assumed that the object exists in the setting recognition range Even if the object recognition process is performed, it is impossible to recognize the object from the captured image of the camera. On the other hand, if the object is a road marking drawn on a road, the marking of the object may be faded. When such fading has occurred, it may not be possible to recognize the object from the captured image of the camera. In this regard, if the object cannot be recognized when the object is recognized within the recognition range as described above, is the recognition failure caused by the object itself not being drawn on the road? Alternatively, although the object is drawn on the road, there is a possibility that it cannot be distinguished whether the object is caused by the marking blur of the object.

  The present invention has been made in view of the above points, and an object of the present invention is to provide an object recognition device capable of determining the absence of an object in a recognition range.

  The above-mentioned purpose is that the position of the own vehicle measured by the positioning means, the position of the object on the road to be recognized by the own vehicle stored in the storage means, and the positioning accuracy of the own vehicle position by the positioning means, An object recognition device comprising object recognition means for executing processing for recognizing the object in a road recognition range set based on the information, and stores information indicating a state of the object on the road When the object is not recognized in the recognition range by the object information storage means and the object recognition means, the recognition is performed based on the state information of the object stored in the object information storage means. It is achieved by an object recognition device comprising: an object presence / absence determination unit that determines whether or not the object exists in a range.

  In the invention of this aspect, when the object is not recognized in the road recognition range based on the position of the own vehicle, the position of the object, and the positioning accuracy of the own vehicle position, the object is stored in the object information storage means. Based on the information indicating the state of the target object on the road, the presence / absence of the target object in the recognition range is determined. If the recognition is not performed when the state of the object on the road is good for recognition, it can be determined that the object has been erased from the actual road. Therefore, according to the present invention, when an object that is in good condition on the road is recognized in the recognition range, the object exists on the road in the recognition range. It can be determined that it was not.

  In this regard, in the above-described object recognition device, the object presence / absence determination unit is stored in the object information storage unit when the object recognition unit does not recognize the object in the recognition range. When the state of the object is good, it may be determined that the object does not exist in the recognition range.

  On the other hand, if the object is not recognized when the condition on the road is not good for recognition, the object has been deleted from the actual road or the object is It is impossible to distinguish whether a recognition failure has occurred in a situation that exists on an actual road.

  Therefore, in the above-described object recognition device, the object presence / absence determination means is stored in the object information storage means when the object recognition means does not recognize the object in the recognition range. When the state of the object is not good, it may be determined that the presence or absence of the object in the recognition range is unknown.

  In the above-described object recognition device, the object information storage means may store information indicating a blurring state on the road of the object drawn on the road.

  In the above-described object recognition device, the object recognition means may recognize the object in the recognition range by processing a captured image of a camera that captures a road mounted on the host vehicle. Good.

  According to the present invention, when an object that is in good condition on the road is not recognized in the recognition range, the object exists on the road in the recognition range. It can be determined that there was not.

  FIG. 1 shows a configuration diagram of a system mounted on a vehicle according to an embodiment of the present invention. As shown in FIG. 1, the system of the present embodiment includes a positioning unit 12 for positioning the position of the host vehicle, and a support control unit 14 for controlling the traveling of the host vehicle. Is determined by the positioning unit 12, and the support control unit 14 executes predetermined support control for driving the host vehicle according to the position of the host vehicle.

  The positioning unit 12 receives a GPS signal transmitted from a GPS (Global Positioning System) satellite and detects the latitude and longitude of the position where the host vehicle exists, and uses the turning angle and the geomagnetism of the host vehicle. An azimuth sensor 18 for detecting the yaw angle (azimuth), a G sensor 20 for detecting acceleration / deceleration, a vehicle speed sensor 22 for detecting vehicle speed, and outputs of these receivers and sensors 16 to 22 are mainly connected. It has the dead reckoning part 24 comprised with the computer. Output signals from the receivers and the sensors 16 to 22 are respectively supplied to the dead reckoning navigation unit 24. The dead reckoning navigation unit 24 detects the latitude and longitude (initial coordinates) of the position of the host vehicle based on information from the GPS receiver 16, and the traveling direction and vehicle speed of the host vehicle based on information from the sensors 18 to 22. And a traveling state of acceleration / deceleration is detected, and a traveling locus (estimated locus) of the vehicle from the initial coordinates of the own vehicle position is created.

  The positioning unit 12 also has a map matching unit 28 mainly composed of a microcomputer connected to the dead reckoning navigation unit 24 described above. A map database 30 is also connected to the map matching unit 28. The map database 30 is composed of a hard disk (HDD), DVD, CD, or the like that is mounted on a vehicle or provided at the center, and is drawn on the road or link data of the road itself necessary for route guidance and map display. Stores information about stationary objects to be installed and lane lane position information. The map matching unit 28 appropriately accesses the map database 30 and reads data stored in the map database.

  Specifically, the map database 30 includes data on lane shapes and road types such as latitude / longitude, curvature, gradient, number of lanes, lane width, corner presence / absence representing roads, information on intersections and node points, Stores information related to buildings for map display, etc., as well as pedestrian crossings and temporary stop lines drawn on road surfaces, direction arrows, rhombus markings with “crosswalks”, maximum speed markings, and rotation prohibition markings For each stationary object such as shape data, paint data, position data, feature size, distance data with other objects before and after, data indicating the degree of marking rubbing at the time of measurement, marking rubbing ease Data indicating the degree, distance data with a target object in the vehicle traveling direction, and the like are stored. In addition, the map database 30 can update the stored map data to the latest one by exchanging disks or establishing update conditions.

  The map matching unit 28 is supplied with the initial coordinates of the vehicle position detected and created by the dead reckoning unit 24 and information on the estimated trajectory from the initial coordinates. The map matching unit 28 uses the link information of the road itself stored in the map database 30 based on the initial coordinates and the estimated trajectory based on the GPS of the host vehicle supplied with information from the dead reckoning unit 24. And performing a map matching (first map matching) to be corrected on the road link, and calculating a position accuracy level indicating the accuracy of the position of the host vehicle detected as a result of the first map matching. ing.

  The map matching unit 28 reads the object data on the road surface within the predetermined range from the current position of the own vehicle obtained as a result of the first map matching as described above from the map database 30 and determines the predetermined data from the own vehicle position. An object within the road range is set as an object to be recognized by the host vehicle. Then, after the setting, based on the position of the host vehicle whose position is detected, it is determined whether or not the set object should be recognized using an image captured by the back camera 32 described later, and the object is recognized. When necessary, the recognition process is performed as described later.

  The positioning unit 12 also has a back camera 32 connected to the map matching unit 28. The back camera 32 is disposed in a vehicle rear bumper or the like, and can capture the outside of a predetermined area including the road surface behind the vehicle from the disposed position. The captured image of the back camera 32 is supplied to the map matching unit 28.

  When it is determined that the object should be recognized from the image captured by the back camera 32, the map matching unit 28 performs image processing such as edge extraction on the captured image when the captured image is supplied from the back camera 32. Thus, the object and the traveling lane drawn on the road surface are detected, and the relative positional relationship between the object and the traveling lane and the host vehicle is grasped.

  The map matching unit 28 calculates the position of the own lane relative to the own vehicle on the road on which the own vehicle actually travels based on the detection result of the travel lane from the captured image of the back camera 32. In addition, based on the detection result of the object, while measuring the relative relationship between the own vehicle and the recognized object existing behind the road of the own vehicle (specifically, the distance from the own vehicle to the recognized object) Based on the measurement result and the position data of the recognition object stored in the map database 30, the position of the host vehicle (particularly the position in the vehicle front-rear direction) is in a relative positional relationship with the recognition object. Map matching (second map matching) for correcting the position is performed.

  As described above, the map matching unit 28 corrects the current position of the host vehicle on the road link stored in the map database 30 every time information on the estimated trajectory is supplied from the dead reckoning unit 24 as described above. While performing 1 map matching, when the target object which should be recognized is recognized from the captured image of the back camera 32, the 2nd map matching which corrects the position of the own vehicle to the position based on the recognition target object by the recognition result is performed. The map matching unit 28 calculates the accuracy indicating the accuracy of the current position of the host vehicle measured as a result of the map matching.

  In addition, the map matching unit 28 executes the support control ahead of a predetermined range in the traveling direction of the own vehicle as a result of collating the position of the own vehicle measured by the map matching with the map data stored in the map database 30. When there is a target object (for example, a stop line, an intersection, a curve entrance, etc.) that is a necessary control target, thereafter, the position of the subject vehicle and the map database 30 are stored for each positioning. The distance along the center line of the driving lane from the host vehicle to the target object ahead in the direction of travel (hereinafter referred to as the road remaining distance) is calculated based on the relationship with the stored position of the target object. To do.

  The positioning unit 12 and the support control unit 14 are connected to each other. Information on the position of the host vehicle and the remaining road distance detected by the positioning unit 12 is supplied to, for example, a display that can be seen by an occupant provided in the vehicle, and is schematically shown on a road map displayed on the screen. Are superimposed and displayed, and supplied to the support control unit 14.

  The support control unit 14 includes an electronic control unit (hereinafter referred to as a support ECU) 40 mainly composed of a microcomputer. The support ECU 40 controls the operation of the driver when the host vehicle travels on the road. Execute support control to support. This assistance control is executed according to the position of the host vehicle, specifically according to the remaining distance along the road from the host vehicle to the target object. For example, when the driver does not perform a brake operation in particular. When the vehicle is behind or behind the road, the target object on the road is paused control, which is driving support control for stopping the vehicle in front of the target line on the road, such as the stop line or railroad crossing line. Intersection control, which is driving support control to prevent crossing with other vehicles that are predicted to cross at the intersection, speed control for driving the vehicle at an appropriate speed with respect to the target curve corner, target object Guidance control for performing voice route guidance on the relative distance to

  The assist ECU 40 includes a brake actuator 42 for generating an appropriate braking force for the host vehicle, a throttle actuator 44 for applying an appropriate driving force to the host vehicle, and a gear stage for switching the automatic transmission of the host vehicle. A shift actuator 46, a steer actuator 48 for imparting an appropriate steering angle to the host vehicle, and a buzzer alarm 50 for buzzer sounding, alarm output, and speaker output toward the vehicle interior are connected. The support ECU 40 appropriately drives the actuators 42 to 50 based on the position of the host vehicle measured by the positioning unit 12 and the relative relationship between the host vehicle and the target object in order to execute the above-described support control. Make a command. Each actuator 42-50 is driven according to the drive command supplied from support ECU40.

  Next, a specific operation in the system of this embodiment will be described. When the vehicle driver desires to display the vehicle position on the display or to execute the assist control by the assist control unit 14, the vehicle driver activates the system according to the present embodiment to an operable state by operating a predetermined switch.

  In the system of the present embodiment, after starting, the positioning unit 12 causes the dead reckoning unit 24 to run from the initial coordinates of the own vehicle based on the output signals of the receivers and the sensors 16 to 22 every predetermined time. Create Then, the map matching unit 28 compares the initial coordinate position and the estimated trajectory of the host vehicle by the dead reckoning unit 24 with the link information of the road stored as map data in the map database 30, thereby determining the position of the host vehicle. First map matching to be corrected on the road link is performed.

  When the map matching unit 28 detects the position of the host vehicle, the map matching unit 28 displays the position of the host vehicle over the road map on the display screen of a display display that can be viewed by the occupant. Also, the road range from the vehicle position to the position when the vehicle travels for a predetermined time or a predetermined distance in the future or to the position of the target object that is the control target of the support control (all lanes when there are multiple lanes) Data such as the position and shape of the target object is read from the map database 30 and the target object to be recognized by the back camera 32 is extracted from the target object located within the predetermined road range. Then, based on the extracted position of the object to be recognized and the position of the vehicle that is constantly updated, it is determined whether or not the vehicle position has reached the vicinity of the position of the object to be recognized. Then, the captured image of the back camera 32 is processed to determine whether or not it is time to recognize the set object to be recognized.

  When the object to be recognized is to be recognized as a result of the determination of the timing, the map matching unit 28 receives the captured image of the rear of the vehicle imaged by the back camera 32 and extracts the edge of the captured image. Perform image processing. Then, the result of the image processing is compared with feature data such as shape data and position data of the object to be recognized to determine whether the object to be recognized is recognized.

  When the map matching unit 28 recognizes an object to be recognized as a result of the above determination, the map matching unit 28 determines from the own vehicle to the rear of the vehicle based on the relative relationship between the own vehicle specified by the image processing and the recognition object. The relative position and the distance to the recognition target existing in the map are detected, and the position data of the recognition target stored in the map database 30 is read out. Then, based on the relative distance to the recognition object and the position data of the recognition object, the position of the host vehicle (especially the position in the vehicle front-rear direction) is corrected to a position having a detection relative relationship with the position of the recognition object. Second map matching is performed.

  By the way, the object on the road to be recognized by processing the captured image of the back camera 32 to correct the vehicle position by the second map matching is a stop line, a pedestrian crossing, an arrow, a rotation prohibition, a rhombus sign, a character string, Since it is a deceleration zone or the like and is scattered on the road, it is not always necessary to perform the object recognition processing. Further, the position information of the object to be recognized is stored in advance in the map database 30 and the position thereof has a certain degree of accuracy. On the other hand, the position of the host vehicle is measured according to the method as described above. It has a corresponding error. If there is no positioning error in the vehicle position, the object using the camera-captured image when the measured vehicle position coincides with or near the position of the object stored in the map database 30. However, in reality, a positioning error occurs at the vehicle position as described above. Therefore, if the object recognition process is performed only at the time point described above, the vehicle Has not yet passed through the object or has already passed through the object, and a situation in which the object cannot be recognized may occur.

  In order to cope with this situation, based on the relative relationship between the vehicle position and the position of the object to be recognized and the positioning accuracy of the vehicle position, the extension direction of the road to be recognized to recognize the object It is conceivable to set a range (hereinafter referred to as a recognition range). However, since the positioning of the vehicle position is performed based on the estimated trajectory using the output of the GPS receiver and the various sensors 16 to 22, detection parameters of the sensors 16 to 22 and the camera 32 are used for this positioning calculation. An error (for example, a timing rounding error) or the like is superimposed, and an error occurs in the positioning result. Since this positioning error includes what is integrated with the movement of the vehicle, the positioning error in positioning the vehicle position increases as the moving distance of the vehicle increases, and the positioning accuracy decreases. For this reason, assuming that the recognition range in which the recognition processing of the object should be performed without considering the positioning accuracy of the own vehicle position is set uniformly, after the setting, the own vehicle whose position is positioned has entered the recognition range. Even when the determination is made, there may be a situation where the actual position is not within the recognition range due to a positioning error that increases as the vehicle moves. As a result, the object may not be recognized.

  In general, when the positioning accuracy of the own vehicle position is extremely high, the position of the own vehicle to be positioned has accuracy, and therefore, from the relationship between the positioning position and the position of the object to be recognized stored in the map database 30. In addition, it is possible to accurately grasp the timing until the target object can be recognized by the host vehicle on the real road. In this respect, the target object to be recognized can be recognized using the captured image of the back camera 32. Then, it is enough even if the recognition range for the recognition is relatively narrow (distance is short). On the other hand, the lower the positioning accuracy of the vehicle position, the larger the error in the position of the vehicle that is positioned. Therefore, from that position, the timing until the object that the vehicle should recognize can be recognized on the actual road. In this respect, when recognizing an object to be recognized using the captured image of the back camera 32, the recognition range for recognition is relatively wide (longer in distance). It is necessary to do.

  In addition, since the positioning error of the own vehicle position becomes larger as the moving distance of the vehicle becomes longer, even if the positioning accuracy of the own vehicle position at a certain time is the same, the next position on the traveling lane from the current vehicle position at that time is the same. The longer the distance to the position of the target object, the larger the positioning error until the host vehicle approaches the target object. In this regard, when recognizing an object to be recognized using the captured image of the back camera 32, a road recognition range for the recognition is defined as a relative distance on the traveling lane between the vehicle position and the object position. It is appropriate to make it narrower as the (traveling distance along the road) is shorter and wider as the relative distance is longer.

  FIG. 2 is a diagram for explaining a method for setting a road recognition range in which the object should be recognized according to the positioning accuracy of the vehicle position in the system of the present embodiment. FIG. 3 shows a flowchart of an example of a control routine executed by the map matching unit 28 in the system of this embodiment.

  In the present embodiment, the map matching unit 28 determines and acquires the position of the vehicle based on the estimated trajectory of the vehicle using GPS or sensor output or based on the recognition result of the object using the camera captured image ( Step 100) and first, as described above, from the measured own vehicle position to the position where the own vehicle will travel for a predetermined time or a predetermined distance in the future, or the nearest target that is the target of the support control that can be executed in the own vehicle Data of all object candidates on the road range up to the position of the object is read from the map database 30, and an object to be recognized by the back camera 32 is extracted from the objects in the road area and set (step 102). ).

  At this time, the objects to be recognized using the back camera 32 extracted from the object candidates in the road range are all objects in the road range whose position data is stored in the map database 30. It is not necessary to be a thing, for example, it is good also as limiting to the thing within the range of the predetermined distance (for example, 1 km or 700 m) before that from the target object of support control, and it is easy to specify a shape from a camera image Or those that are less likely to cause rubbing, those that follow an arrangement pattern according to the type of road that runs from the vehicle position to the target object for assistance control, or that the object is recognized by a camera image. If the position of the host vehicle is corrected, the positioning accuracy required for the support control that can be executed in the host vehicle is maintained, and the host vehicle reaches the target object by executing the support control. It may be only those that can Rukoto. In this case, for example, in order to satisfy the positioning accuracy when reaching the temporary stop line required for stopping the own vehicle on the temporary stop line as the target object by the temporary stop control, the correction of the own vehicle position is performed. The positioning accuracy required to properly execute the support control for the object existing on the traveling lane before the temporary stop line to be implemented, the current positioning accuracy, the relative distance from the own vehicle, and Extraction is based on the relative distance to the target object.

  When the map matching unit 28 extracts the object to be recognized by the back camera 32 from the predetermined road range as described above, the position of the object to be recognized (when there are a plurality of objects to be recognized, at least The data of the position of the object closest to the own vehicle position is acquired from the map database 30 (step 104), and the road along the travel lane between the acquired position of the object to be recognized and the measured own vehicle position The relative distance is calculated (step 106).

  Further, the map matching unit 28 calculates the current positioning accuracy for positioning the vehicle position (step 108). This calculation may be performed by substituting parameters into predetermined formulas experimentally determined in advance. For example, an initial value corresponding to a GPS accuracy error that is not accumulated according to the moving distance of the vehicle is used. In addition, the highest accuracy is obtained each time the vehicle position correction (second map matching) based on the object recognition in the camera captured image is performed, and in advance according to the vehicle movement distance after the second map matching. What is necessary is just to acquire the precision which becomes low, so that it falls with the predetermined | prescribed predetermined gradient and the movement distance becomes long.

  Then, the map matching unit 28 recognizes the object on the road using the back camera 32 based on the relative distance between the vehicle and the object to be recognized and the positioning accuracy of the vehicle position calculated as described above. A road recognition range (that is, a region range in which an object can be reliably recognized if recognition processing is performed within that range even if a positioning error that is allowed to the maximum in the vehicle position to be measured occurs) is set ( Step 110).

  Specifically, the map matching unit 28 stores in advance a map to be used for setting the recognition range based on the relative distance between the host vehicle and the object to be recognized and the positioning accuracy. As shown in FIG. 2, this map is such that the smaller the positioning error of the vehicle position, that is, the higher the positioning accuracy, the narrower the recognition range, and the larger the positioning error, ie, the lower the positioning accuracy, the wider the recognition range. The recognition range is narrowed as the relative distance on the travel lane between the subject vehicle and the object to be recognized is shorter, and the recognition range is wider as the relative distance is longer. Note that the road recognition range that is set may correspond to the maximum positioning error that may occur when the vehicle travels a relative distance from the object to be recognized. The map matching unit 28 sets the road recognition range as described above with reference to this map.

  When the map matching unit 28 sets the recognition range of the road on which the object is to be recognized as described above, thereafter, the map matching unit 28 compares the own vehicle position to be updated with the set recognition range, and the own vehicle becomes the recognition range. It is determined whether or not the vehicle has entered, and the determination is repeated until an affirmative determination is made. If it is determined that the host vehicle has entered the set road recognition range, it is determined that the object to be recognized is in a situation to be recognized by processing the captured image of the back camera 32, and thereafter the image is captured from the back camera 32. In response to the supply of the captured image behind the vehicle, image processing such as edge extraction is performed on the captured image (step 112), and the image processing result is compared with the feature data of the object to be recognized. A process for recognizing an object to be performed is performed (step 114).

  As a result of the processing, when an object to be recognized is recognized in a situation where the host vehicle is located within the set road recognition range, it is determined that there is an object to be recognized within the road recognition range (step 116). ), Second map matching for grasping the relative relationship between the vehicle identified by the image processing and the recognition object, and correcting the vehicle position to a position having the relative relationship with respect to the position of the recognition object. carry out. On the other hand, when the subject vehicle is not recognized in the situation where the subject vehicle is located within the set road recognition range, that is, the subject vehicle position to be updated and its set are recognized without recognizing the subject matter to be recognized. When the own vehicle has advanced from the set road recognition range based on the comparison result with the road recognition range, the process is terminated without performing the second map matching.

  Thereafter, the map matching unit 28 acquires position data for each object to be recognized until the host vehicle reaches the target object for assist control or the vicinity thereof, and determines the recognition range of the road on which the recognition process is to be performed. Set and execute the same process as described above.

  When the map matching unit 28 performs the second map matching and corrects the position of the vehicle, the map matching unit 28 accesses the map database 30 and travels the road from the recognition target object to the target target object for support control. And the initial value of the road remaining distance from the vehicle to the target object is calculated based on the position of the vehicle and the distance from the recognition object to the target object.

  Further, when the map matching unit 28 recognizes an object to be recognized within a predetermined road range, the map matching unit 28 acquires and recognizes information on a driving lane on the road by performing image processing of a captured image from the back camera 32. To grasp the relative relationship of the lane to the host vehicle. Then, the lane width, the number of lanes, the shape, etc. of the traveling lane in the vicinity of the own vehicle position are acquired by accessing the map database 30, and the relative relationship of the traveling lane with respect to the own vehicle, the number of lanes acquired from the map database 30, etc. Based on this, the position of the own lane on the road on which the own vehicle is traveling at the present time is specified. The target object may be different for each driving lane, but when the position of the own lane is specified as described above, the target object ahead of the traveling direction to which the host vehicle existing on the own lane should pass is specifically Since it is specified, it is possible to calculate the above-mentioned road calculation distance based on the target object on the own lane.

  The dead reckoning unit 24 creates an estimated trajectory of the vehicle position using the GPS receiver 16 and the various sensors 18 to 22 every predetermined time, and transmits the trajectory information to the map matching unit 28. When the map matching unit 28 performs the second map matching accompanying the object recognition as described above, each time it receives the estimated trajectory information from the dead reckoning navigation unit 24, first, the map matching unit 28 first estimates from the second map matching time point. Based on the locus and the position of the own lane, the position of the own vehicle (particularly the distance in the front-rear direction) with respect to the position coordinates of the recognition target on the center line of the own lane is calculated. Based on the distance in the front-rear direction and the distance between the recognition object and the target object on the own lane, the road remaining distance from the current position of the own vehicle to the target object is calculated.

  The information on the own vehicle position and the information on the remaining road distance detected by the positioning unit 12 is supplied to the display display and to the support control unit 14. In accordance with the information from the positioning unit 12, the display display schematically displays the vehicle position and the remaining distance along the road superimposed on the road map displayed on the display screen.

  The assistance ECU 40 of the assistance control unit 14 determines, for each assistance control, on the basis of the distance on the road traveling to the target object such as a stop line or an intersection that is the control object of the assistance control supplied from the positioning unit 12. It is determined whether or not a control start condition defined for the control is satisfied. When the control start condition is satisfied, the support control is started.

  For example, in the pause control, the brake is applied when the distance from the measured vehicle to the pause line that is the target object is, for example, 30 meters (may be a distance that changes according to the vehicle speed). Automatic braking by the actuator 42 is started, and the host vehicle is stopped at the temporary stop line. At this time, before starting the automatic braking brake by the brake actuator 42, voice guidance for notifying the driver that the automatic braking brake is performed may be performed. In the route guidance control by voice, when the distance from the vehicle to be measured to the target object such as an intersection reaches 100 meters, for example, the speaker outputs the buzzer alarm 50 to the driver in front of the vehicle. Guidance that the object exists is given.

  Thus, in the system of the present embodiment, the position of the host vehicle is stored in the map database 30 based on the GPS signal received by the GPS receiver 16 and the estimated trajectory using the outputs of the various sensors 18-22. The first map matching for positioning on the road link of the existing map data can be performed. Further, an object drawn or installed on a road is recognized by processing of an image captured by the back camera 32, and the position of the recognition object stored in advance in the map database 30 and the vehicle at the time of recognition. The second map matching for measuring the position of the host vehicle can be performed based on the relative relationship between the vehicle and its recognition object.

  As described above, as the moving distance of the vehicle increases, the positioning error of the own vehicle position increases, and the positioning accuracy decreases. On the other hand, the position information of the object stored in the map database 30 is generally measured and has high accuracy. Therefore, if the position data of the object is used, the position information for positioning the vehicle is determined. The error becomes smaller and the positioning accuracy becomes higher. Therefore, according to the system of the present embodiment, the first map matching is normally performed based on the estimated trajectory of the vehicle using GPS or sensor output, while the recognition is performed when the object is recognized using the camera captured image. By performing the second map matching based on the result, the positioning accuracy of the own vehicle position becomes longer as the moving distance of the own vehicle becomes longer before the object drawn or installed on the road is recognized by the captured image by the back camera 32. However, every time the object is recognized, the positioning accuracy of the vehicle position can be improved.

  Further, in the system of the present embodiment, the support control is executed according to the own vehicle position measured by the positioning unit 12 (specifically, the distance from the own vehicle to the target object that is the target of the support control). Can do. Although the assist control is not performed before the vehicle reaches a predetermined relative positional relationship with the target object as a positioning result, the assist control can be performed when the relative positional relationship is reached. For this reason, according to the system of the present embodiment, the driver who runs the vehicle on the road by performing the pause control, the intersection control, the speed control, and the guidance control according to the positioning result of the vehicle. It is possible to drive the host vehicle safely and appropriately by supporting the operation.

  In the system of the present embodiment, in order to recognize an object on the road to be recognized using the captured image of the back camera 32, the positioning accuracy of the own vehicle position and the actual vehicle position to be actually measured are determined. And the recognition process is performed based on the position of the object to be recognized that is stored in advance in the map database 30 (specifically, the relative distance on the travel lane between the subject vehicle and the object to be recognized). Set the road recognition range. Specifically, the narrower road recognition range is set as the positioning accuracy is higher, and the wider road recognition range is set as the positioning accuracy is lower. Also, the narrower road recognition range is set as the relative distance on the travel lane between the subject vehicle and the object to be recognized is shorter, and the wider road recognition range is set as the relative distance is longer. And the recognition process of the target object using the captured image of the back camera 32 is performed in the road recognition range.

  That is, the object recognition process using the captured image of the back camera 32 is performed based on a predetermined road recognition set based on the relative distance between the vehicle and the object to be recognized and the positioning accuracy of the vehicle position. This is limited to the case where the vehicle travels within the range, and the object recognition process is not performed outside the range. Therefore, according to the system of the present embodiment, the object to be recognized by the host vehicle is prevented from becoming unrecognizable, and the object is compared with the system that always performs the object recognition process using the camera captured image. It is possible to reduce the processing load required for object recognition.

  By the way, it is not only when the object itself is not drawn on the road in the recognition range that the object cannot be recognized as a result of the image processing of the camera captured image within the set road recognition range. This also occurs when an object is drawn on the road in the recognition range, but the sign of the object is faint or the shape of the object is difficult to recognize. For this reason, the recognition of the object within the recognition range is caused by the fact that the object itself is not drawn on the road, or the object is drawn on the road, but the object There is a possibility that it cannot be distinguished whether it is caused by fading or the like.

  Therefore, in the system of the present embodiment, first, in the map database 30 storing various types of map data, for each stationary object drawn on the road surface, data indicating the degree of rubbing at the time of the latest road measurement is displayed. Stored. The rubbing degree is a numerical value indicating the degree to which an object cannot be recognized from a camera-captured image obtained by copying the object on the road, and the larger the numerical value, the more difficult the object is recognized. The smaller the numerical value, the easier it is to recognize the object.

  In general, when the degree of rubbing of an object to be recognized is large and it is recognized that the state is not good for recognizing the object from a camera image, the object cannot be recognized within the recognition range. If the object is unrecognizable due to the fact that the object itself is not drawn on the road, or the object is drawn on the road but the sign of the object is faint, etc. It is still difficult to distinguish whether it is due. However, if the degree of rubbing of the target object to be recognized is small and it is recognized that the state is good for recognizing the target object from the camera captured image, the target object exists within the recognition range. For example, if the recognition process is performed within the recognition range, the object must be recognized, and conversely, if the recognition process within the recognition range fails to recognize the object, It can be determined that there is a very high possibility that the object was not present within the recognition range by being erased from the actual road.

  In the system of the present embodiment, when the map matching unit 28 determines in step 114 above that the subject vehicle should not be recognized in the situation where the host vehicle is located within the set road recognition range, The database 30 is accessed, information indicating the rubbing degree state on the road of the object to be recognized stored in the map database 30 is acquired (step 118), and the rubbing degree is set to a predetermined threshold. By comparing with the value, it is determined whether or not the state of the object is favorable for recognition from the camera-captured image (step 120). The predetermined threshold is a rubbing degree that represents a boundary indicating whether or not the object can be recognized from a camera-captured image obtained by copying the object, and is a value experimentally given in advance. The predetermined threshold value may be different for each type of object.

  As a result of the determination, if the rubbing degree of the object that should have been recognized is good, it is assumed that the object to be recognized actually did not exist within the set road recognition range. Judgment is made (step 122), and the process is terminated without performing the second map matching. On the other hand, if the marking rubbing degree of the object that should have been recognized is not good, it is determined that the presence or absence of the object within the recognition range is unknown, and the second map matching is performed. The process is finished without.

  Even though it is recognized that the marking rubbing state of the object to be recognized within the recognition range is good for recognizing the object from the camera image, the recognition process recognizes the object. When it was not possible, the data of the object stored in the map database 30 became different from the actual data, that is, the data was retained after the most recent data was stored in the map database 30. It can be determined that the object has been erased from the actual road. Therefore, according to the system of the present embodiment, when an object that is in good condition on the road is not recognized in the recognition range, the object is displayed on the road in the recognition range. It is possible to determine that an object did not exist, and it is possible to accurately determine the absence of an object in the recognition range in a predetermined situation.

  In this regard, according to the present embodiment, when it is determined that the object does not exist on the road in the recognition range, it is determined that the data of the object stored in the map database 30 is not accurate at the present time. Since the update can be performed, the accuracy of the object data stored in the map database 30 can be improved as appropriate. Further, when it is determined that the object does not exist on the road in the recognition range, the own vehicle position determined by matching is different from the actual own vehicle position. For example, when two roads are parallel, the own vehicle It is possible to determine that the driving road is not a currently matching road but a road parallel to that road, and the vehicle position can be re-matched to that road, thus improving the positioning accuracy. It is possible.

  In the above-described embodiment, the map database 30 for storing the position data of the object is stored in the “object recognition device” described in the claims by the positioning unit 12. "And" object information storage means ", the back camera 32 corresponds to the" camera "described in the claims, and the map matching unit 28 performs the process of step 100 in the routine shown in FIG. The “positioning means” described in the claims by executing the process of recognizing the object to be recognized in the set road recognition range causes the “object” described in the claims. The “recognition means” executes the processing of steps 118 to 124, thereby realizing the “object presence / absence determination means” recited in the claims.

  By the way, in the above embodiment, the object on the road is recognized using the captured image of the back camera 32 disposed in the rear part of the vehicle, but the present invention is not limited to this. Such recognition of the object may be performed based on a captured image of a camera disposed in the front portion of the vehicle or information sent from an external infrastructure.

  Further, in the above-described embodiment, the second map matching for correcting the object on the road recognized using the camera captured image from the vehicle position based on the traveling locus by the GPS, the vehicle speed, the steering angle, or the like. However, the present invention is not limited to this. For example, when the object is provided as a road sign indicating an area where entry of the vehicle is prohibited, the vehicle is in the area. It may be used for purposes other than map matching, such as a system that issues a vehicle warning when entering the vehicle.

  Further, in the above embodiment, the sign rubbing degree used to determine whether or not the state of the object is good for recognition from the camera-captured image is the latest stored in the map database 30. However, the present invention is not limited to this, and the present invention is not limited to this. The elapsed time from the measurement and the vehicle running frequency based on the degree of rubbing at the time of the most recent road measurement. , It may be the current rubbing degree calculated in consideration of the degree of ease of rubbing.

It is a block diagram of the system mounted in the vehicle which is one Example of this invention. It is a figure for demonstrating the method of setting the recognition range of the road which should recognize a target object according to the positioning accuracy of the own vehicle position in the system of a present Example. It is a flowchart of an example of the control routine performed in the system of a present Example.

Explanation of symbols

12 Positioning unit 16 GPS receiver 18 Direction sensor 20 G sensor 22 Vehicle speed sensor 24 Dead reckoning unit 28 Map matching unit 30 Map database 32 Back camera

Claims (5)

It is set based on the position of the own vehicle measured by the positioning means, the position of the object on the road to be recognized by the own vehicle stored in the storage means, and the positioning accuracy of the own vehicle position by the positioning means. An object recognition device comprising object recognition means for executing processing for recognizing the object in a road recognition range,
Object information storage means for storing information indicating the state of the object on the road;
When the object is not recognized in the recognition range by the object recognition means, the presence of the object in the recognition range based on the state information of the object stored in the object information storage means An object presence / absence determining means for determining presence / absence;
An object recognition apparatus comprising:   The object presence / absence determination means has a good state of the object stored in the object information storage means when the object recognition means does not recognize the object in the recognition range. 2. The object recognition apparatus according to claim 1, wherein it is determined that the object does not exist in the recognition range.   The object presence / absence determining means is in a case where the state of the object stored in the object information storing means is not good when the object is not recognized in the recognition range by the object recognizing means. The object recognition apparatus according to claim 1 or 2, wherein it is determined that the presence or absence of the object in the recognition range is unknown.   4. The object according to claim 1, wherein the object information storage unit stores information indicating a blurring state on the road of the object drawn on the road. 5. Recognition device.   5. The object recognition unit according to claim 1, wherein the object recognition unit recognizes the object in the recognition range by processing a captured image of a camera that captures a road mounted on the host vehicle. The object recognition apparatus according to one item.

Images