JP2012155660A - Map data generation device and travel support device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To enable a three-dimensional map data to be generated without including a moving body.SOLUTION: A storage medium of a map data generation device 10 accumulates and stores probe data including image data, ranging data and positioning data collected along with a travel of a probe vehicle 2. Then the device: extracts a plurality of the probe data, including a photographic image photographing a same subject point from a same camera position, from the storage medium; specifies the photographic image without including a moving body based on a dispersion value of pixel data on a specific pixel (X, Y) in the photographic image photographing the same subject point; and generates a three-dimensional map data without including the moving body based on the probe data including the photographic image without including the moving body (S202).

Description

  The present invention relates to a map data generation device and a driving support device.

  2. Description of the Related Art Conventionally, there is a map data generation device that mounts a measurement device such as a stereo camera or a laser radar on a vehicle and generates three-dimensional map data based on image information and distance information collected using the measurement device as the vehicle travels. Are known.

  In addition, the outer shape of the building is acquired from the two-dimensional map data, and the height is obtained using a photograph of the building included in the two-dimensional map data, so that the surrounding area of the base station of the wireless communication system There is also a map data generation system that generates the three-dimensional map data (see, for example, Patent Document 1).

Republished Patent 2008-062819

  However, in the configuration in which the 3D map data is generated using the image captured using the camera as in each of the above-described devices, the captured image of the camera, for example, a pedestrian or a parked vehicle parked on the shoulder, If moving objects that can move are included, map data including these moving objects is generated.

  Such map data including a moving object has a problem that accuracy as a feature map is lowered, which is not preferable.

  The present invention has been made in view of the above problems, and a first object thereof is to enable generation of 3D map data that does not include a moving object.

  In addition, a second object of the present invention is to provide driving support for an on-board vehicle so as to avoid a moving object using 3D map data that does not include the moving object.

  In order to achieve the first object, according to the first aspect of the present invention, information collected by a data collection device is obtained from a plurality of vehicles equipped with the data collection device, and map data is obtained using the obtained information. The data collection device generates image data representing a photographed image obtained by photographing the periphery of the mounted vehicle, and distance measurement data for specifying the distance between each part of the subject of the photographed image, The probe data associated with the positioning data for specifying the position and orientation of the mounted vehicle is collected, and when the probe data collected by the data collection device is received, the received probe data is stored in the storage medium. Including a captured image obtained by photographing the same place from the same point from the storage processing means for storing and storing the probe data and the probe data stored and stored in the storage medium. A plurality of probe data is extracted, a captured image that does not include a movable object that can be moved is identified based on a variance value of pixel data of specific pixels of the captured image that is captured at the same location, and the moving object Map data generation data registration means for registering probe data including a captured image not including the image data as map data generation data, and each area of the subject included in the captured image based on the probe data registered in the map data generation data And a feature map data generating means for generating three-dimensional feature map data that does not include a moving object by calculating the absolute position.

  According to such a configuration, a plurality of probe data including a photographed image obtained by photographing the same place from the same spot is extracted from the probe data accumulated and stored in the storage medium, and the photographed image obtained by photographing the same place is extracted. Based on the dispersion value of the pixel data of the specific pixel, a captured image that does not include a movable object that can move is specified, and probe data that includes the captured image that does not include the moving object is registered as map data generation data. Then, based on the probe data registered in the map data generation data, the absolute position of each region of the subject included in the captured image is calculated, and three-dimensional feature map data not including a moving object is generated. That is, it is possible to generate 3D map data that does not include a moving object.

  The map data generation data registration means may be configured to perform a plurality of photographic images included in the probe data stored and stored in the storage medium and photograph the same location from the same location. It is characterized by extracting probe data including the photographed images.

  According to such a configuration, probe data including a plurality of photographed images obtained by photographing the same place from the same spot by moving the viewpoint of the photographed image contained in the probe data accumulated and stored in the storage medium is extracted. Even if the photographed position is slightly deviated, it can be extracted as probe data including a photographed image obtained by photographing the same place from the same spot.

  Further, as in the invention described in claim 3, the pixel data can be at least one of color data and luminance data.

  In order to achieve the second object, the invention described in claim 4 is a three-dimensional feature generated by a photographing means for photographing the periphery of the vehicle and the map data generating device described in claim 1. Feature map data acquisition means for acquiring feature map data at the same place as the photographed image taken by the photographing means from the storage medium storing the figure data, and photographed image and feature map data acquisition means obtained by the photography means The moving object detecting means for detecting a moving object capable of moving based on the variance value of the pixel data of the specific pixel of the feature map based on the feature map data acquired by the And travel support means for supporting travel of the mounted vehicle so as to avoid the moving object.

  According to such a configuration, the feature in the same place as the photographed image photographed by the photographing means from the storage medium storing the three-dimensional feature map data generated by the map data generating device according to claim 1 Obtaining the figure data, detecting a moving object that can move based on the captured image taken by the photographing means and the dispersion value of the pixel data of the specific pixel of the feature map based on the feature map data, Driving assistance of the mounted vehicle is performed so as to avoid the detected moving object. That is, it is possible to perform the travel assistance of the mounted vehicle so as to avoid the moving object using the three-dimensional map data not including the moving object.

  According to a fifth aspect of the present invention, there is provided a current position specifying means for specifying the current position of the mounted vehicle, and a three-dimensional ground around the mounted vehicle based on the current position of the mounted vehicle specified by the current position specifying means. Object map data is acquired and photographed by the photographing means so that the error amount between the feature map around the mounted vehicle specified by the three-dimensional feature map data and the photographed image photographed by the photographing means is less than a threshold value. Viewpoint moving means for moving the viewpoint of the captured image, and current position correcting means for correcting the current position of the mounted vehicle so that the position of the viewpoint moved by the viewpoint moving means is the current position. It is a feature.

  As described above, the photographed image photographed by the photographing means so that the error amount between the feature map around the mounted vehicle specified by the three-dimensional feature map data and the photographed image photographed by the photographing means is less than the threshold value. The current position of the mounted vehicle can be accurately identified by correcting the current position of the mounted vehicle specified by the current position specifying means so that the position of this viewpoint is the current position.

It is a figure which shows a mode that the map data generation apparatus which concerns on one Embodiment of this invention performs data communication with several probe vehicles. 1 is a diagram illustrating an overall configuration of a map data generation device and an in-vehicle device according to an embodiment of the present invention. It is a figure for demonstrating the ranging data output from a distance image sensor. It is a flowchart of the probe data collection process by the control part of a vehicle-mounted apparatus. It is a flowchart of the map data generation process by the control part of a map data generation apparatus. It is a flowchart of DB creation processing for map data generation. It is a figure for demonstrating feature map data. It is a figure for demonstrating the determination of the dispersion value of the pixel data of a picked-up image. It is a flowchart of the map data update process by the control part of a vehicle-mounted apparatus. Flow chart of driving support processing by control unit of in-vehicle device It is a flowchart of the mobile body detection process by the control part of a vehicle-mounted apparatus. It is a figure for demonstrating calculation of the planned driving | running | working position of the mounted vehicle which avoids a moving object.

  FIG. 1 shows the overall configuration of a map data generation apparatus according to an embodiment of the present invention. The map data generation device 10 is configured by a computer arranged in the information center 1 and performs data communication with an unspecified number of probe vehicles 2. Each probe vehicle 2 is equipped with an in-vehicle device 20 having a function as an information collection device for collecting probe information necessary for the map data generation device 10 to generate map data.

  The map data generation device 10 generates three-dimensional feature map data based on probe information transmitted from the in-vehicle device 20 mounted on each probe vehicle 2. Further, the in-vehicle device 20 uses the three-dimensional feature map data generated by the map data generation device 1 to assist in driving the vehicle while avoiding moving objects such as a pedestrian and a parked vehicle parked on the shoulder. It also has a function as a driving support device that performs

  FIG. 2 shows a block configuration of the map data generation device 10 and the in-vehicle device 20. The map data generation device 10 includes a map database 11, a communication unit 12, and a control unit 13.

  The communication unit 12 is for data communication with each probe vehicle 2 via a public communication network.

  The control unit 13 is configured as a computer including a CPU, a memory, a hard disk driver, an I / O, and the like, and the CPU performs various processes according to programs stored in the memory. The processing of the control unit 13 includes processing for generating three-dimensional feature map data based on probe information transmitted from the in-vehicle device 20 and storing the generated three-dimensional feature map data in the map database 11. .

  On the other hand, the in-vehicle device 20 includes a distance image sensor 21, a position detection device 22, a map data storage unit 23, a communication unit 24, and a control unit 25. A vehicle drive control unit 30 is connected to the control unit 25.

  The distance image sensor 21 captures a subject using a stereo camera, and uses distance measurement data for specifying the distance between the distance image sensor 21 and each part of the subject of the captured image and a captured image for each pixel of the captured image. Output the image data to represent. The distance image sensor 21 is attached to each probe vehicle 2 so that the viewing angle of the stereo camera coincides with the traveling direction of the vehicle and the height from the road surface is the same.

  As shown in FIG. 3, the distance measurement data includes the XY coordinate position in the photographed image as XY, the color information at this coordinate position as RGB, and the relative coordinates of the subject relative to the position of the distance image sensor 21 as xyz. Then, (XY, RGB, xyz) is output. Such distance measurement data (XY, RGB, xyz) is output for all pixels of the captured image. That is, such distance measurement data (XY, RGB, xyz) is output by the number of pixels for one captured image.

  The position detection device 22 includes a geomagnetic sensor, a gyroscope, a distance sensor, a GPS receiver, and the like (all not shown), and information for specifying the current position input from these is sent to the control unit 25. Output.

  The map data storage unit 23 is a storage device for storing map data. The map data storage unit 23 stores three-dimensional feature map data generated by the map data generation device 10.

  The communication unit 24 is provided to communicate with the map data generation device 10 of the information center 1 via the public wireless communication network.

  The control unit 25 is configured as a computer including a CPU, a memory, an I / O, and the like, and the CPU performs various processes according to a program stored in the memory.

  For the processing of the control unit 25, the current position of the vehicle is specified based on the information for specifying the current position input from the position detection device 22, and the probe data based on the information input from the distance image sensor 21 is used. There are probe data processing to be collected, travel support processing for supporting travel of a vehicle while avoiding a moving object such as a parked vehicle parked on the shoulder of a road.

  The vehicle drive control unit 30 includes a brake ECU that performs vehicle brake control, a steering ECU that controls the steering angle of the vehicle, an engine ECU that controls the engine of the vehicle (none of which are shown), and the like.

  The brake ECU in the present embodiment can perform vehicle brake control in accordance with an instruction from the travel control unit 25, and the steering ECU can control the vehicle in response to an instruction from the travel control unit 25. It is possible to control the corner. Further, the engine ECU can adjust the fuel injection amount and the fuel injection timing in accordance with an instruction from the travel control unit 25 to control acceleration / deceleration of the mounted vehicle.

  The in-vehicle device 20 is configured to perform processing for collecting probe data as the mounted vehicle 2 travels. FIG. 4 shows a flowchart of probe data collection processing by the control unit 25 of the in-vehicle device 20. The control part 25 will start the process shown to a figure, if the ignition switch of a vehicle will be in an ON state.

  First, information for specifying the current position input from the position detection device 22 is acquired, and the current position (absolute coordinates) of the vehicle and the vehicle traveling direction are specified (S100). The vehicle traveling direction can be obtained from the positional relationship between the current position (absolute coordinates) of the vehicle specified last time and the current position (absolute coordinates) of the vehicle specified this time.

  Next, ranging data is acquired from the distance image sensor 21 (S102). Specifically, distance measurement data (XY, RGB, xyz) and image data are acquired.

  Next, absolute position data of each part of the subject of the photographed image is calculated (S104). Specifically, based on the current position (absolute coordinates) and the vehicle traveling direction of the vehicle specified in S100 and the distance measurement data (XY, RGB, xyz) acquired in S102, each part of the subject of the captured image Calculate absolute coordinates of. Here, if the absolute coordinates of the current position of the vehicle are (X0, Y0, Z0) and the relative coordinates of each part of the subject are (x, y, z), the absolute coordinates of each part of the subject are (X0 + x, Y0 + y, Z0 + z). Can be calculated as

  Next, a captured image obtained by capturing the periphery of the mounted vehicle, distance measurement data for specifying a distance from the subject in the captured image for each region included in the captured image, and a position and orientation of the mounted vehicle are specified. The probe data associated with the positioning data is transmitted to the information center 2 (S106), the process returns to S100, and the above processing is repeatedly performed. In this way, probe data is periodically transmitted from the in-vehicle device 20 mounted on the probe wheel 2 to the information center 1.

  A large amount of probe data is transmitted to the map data generation device 10 from the in-vehicle device 20 mounted on the unspecified number of probe cars 2 in this way. In addition, since the photographing position of the photographed image is narrowed down to a point on the road where the probe car 2 is located, a plurality of probe data including a photographed image obtained by photographing the same place from the same point can be efficiently collected.

  The map data generation device 10 stores and stores the received probe data in a storage medium, and performs map data generation processing for generating feature map data based on the probe data stored and stored in the storage medium.

  In FIG. 5, the flowchart of the map data generation process by the control part 13 of the map data generation apparatus 10 is shown. The control part 13 starts the process shown in FIG. 5 according to an operator's operation.

  First, it is determined whether probe data has been received (S200). Here, when the probe data is not received, the determination of S200 is repeated. When the probe data is received, the map data generation DB creation process of S300 is performed.

  FIG. 6 shows a flowchart of this map data generation DB creation processing. In this map data generation DB creation process, first, received data is fetched (S302). Specifically, the received probe data is accumulated and stored in a hard disk driver as a storage medium.

  Next, the position of the subject in the captured image is specified (S304). The position (coordinates) of the subject in the captured image can be specified based on distance measurement data and positioning data included in the received probe data.

  Next, a photographed image obtained by photographing the same place from the same spot is read out (S306). Specifically, the shooting position of the shot image is specified based on the positioning data included in the received probe data, and the same location as the subject position of the shot image specified in S304 is shot from the shooting position of the shot image. A plurality of photographed images are read out from the storage medium.

  Next, a cross-correlation coefficient between the captured image included in the received probe data and the entire captured image read in S306 is calculated (S308). Here, the cross-correlation coefficient for the entire image is calculated in order to determine whether or not two images are taken from the same viewpoint from the same place. In the present embodiment, the cross-correlation coefficient for each captured image read in S306 is calculated using the captured image included in the received probe data as a template. Note that the larger the cross-correlation coefficient, the higher the similarity between the two images. Since the calculation of the cross-correlation coefficient is a well-known technique, a detailed description thereof is omitted here.

  Next, it is determined whether or not the cross correlation coefficient of the entire image is equal to or greater than a predetermined threshold (S310).

  If the cross-correlation coefficient of the entire image is not equal to or greater than the threshold value, the viewpoint of the data is moved (S312). Here, the viewpoint conversion of the captured image included in the received probe data is performed so that the cross-correlation coefficient of the entire image is equal to or greater than the threshold value, and the process returns to S308.

  For example, when the viewpoint is moved in the X-axis direction and the cross-correlation coefficient in the entire image increases, the viewpoint is further moved in the X-axis direction. When the viewpoint is moved in the X-axis direction, the viewpoint is moved in the opposite direction when the cross-correlation coefficient in the entire image becomes small. In this way, the viewpoint is moved in the X-axis, Y-axis, and Z-axis directions.

  In addition, the process of S308, S310, and S312 is performed sequentially about each picked-up image read in S306, and the image in which the cross correlation coefficient in the whole image became more than a threshold value is extracted. Although not shown, an image having a cross-correlation coefficient less than a threshold value in the entire image is not extracted even though the viewpoint conversion is performed in S312.

  Next, the variance value of each pixel data is calculated (S312). Specifically, a variance value of pixel data of a specific pixel is calculated for each captured image in which the cross-correlation coefficient is equal to or greater than a predetermined threshold value in S310. Here, the variance value of the pixel data is calculated for all the pixels of the captured image. In the present embodiment, a variance value is calculated for each of the three primary colors (red (R), green (G), and blue (B)) as pixel data. In this embodiment, the variance value of the pixel data is calculated for all the pixels of the captured image. However, it is not always necessary to calculate the pixel data variance value for all the pixels of the photographed image. For example, the processing load is such that the pixel data variance value is calculated for each specific pixel specified at regular intervals. May be reduced.

  FIG. 8 shows an example of the frequency distribution of color information at the coordinates (X, Y) of each photographed image taken from the same viewpoint as the photographed image included in the received probe data. In the figure, (a) represents the frequency distribution of red (R), (b) represents the frequency distribution of green (G), and (c) represents the frequency distribution of blue (B).

  Here, it is assumed that a large number of captured images are transmitted from an unspecified number of probe cars 2 and a large number of captured images obtained by capturing the same place from the same viewpoint as the captured image included in the received probe data are extracted. In addition, a large number of extracted captured images include a moving object that can move and a moving object that does not include a moving object.

  Moving objects that can be moved, such as parked vehicles and pedestrians parked on the road shoulder, will not move in a relatively short time, so they are only included in some captured images, whereas buildings, guardrails, etc. Since the feature does not move, it is always included in all captured images. In other words, the variance value of pixel data in a specific pixel of many captured images taken from the same location from the same point is small when the captured image does not include a moving object, but the captured image includes a moving object. If it is, it will be bigger.

  Here, it is determined whether or not a moving object is included in the captured image based on whether or not the variance value of the pixel data in the specific pixel is equal to or greater than a predetermined threshold (S314). In the present embodiment, whether an object that moves is included in the captured image based on whether or not all the dispersion values of the three primary colors (red (R), green (G), and blue (B)) are equal to or greater than a threshold value. Determine whether or not.

  Here, when all the dispersion values of the three primary colors (red (R), green (G), and blue (B)) of a specific pixel are less than the threshold value, the determination in S314 is YES, and the dispersion value is Probe data including a captured image that is less than the threshold is registered in a map data generation DB (not shown) (S316).

  If at least one of the three primary colors (red (R), green (G), and blue (B)) of a specific pixel is greater than or equal to a threshold value, the determination in S314 is NO, The process proceeds to S318 without registering the photographed image whose variance value is equal to or greater than the threshold value in the map data generation DB.

  In S318, it is determined whether or not the processing for all the pixels has been completed. Here, when the processing for all the pixels has not been completed, the determination in S318 is NO, and the processing returns to S312 and the processing of S312 to S316 is performed on the other pixels.

  When the processing for all the pixels is completed, the determination in S318 is YES, and the process proceeds to S202 in FIG.

  In S202, the feature map data is generated. Specifically, the absolute position of each part of the subject included in the captured image is calculated based on the probe data registered in the map data generation DB to generate three-dimensional feature map data, which is registered in the map database 11. .

  Next, it is determined whether there is newly received data (S204). Here, when probe data is newly received, the determination in S204 is YES, and the process returns to S200. If no new probe data is received, the determination in S204 is NO and the process is terminated. In this way, feature map data representing a three-dimensional feature map that does not include a movable object that can move as shown in FIG. 7 is generated.

  The control unit 25 of the in-vehicle device 20 mounted on the probe vehicle 2 acquires new feature map data by the map data generation device 10 and updates the map data stored in the map data storage unit 12. Update processing is performed.

  FIG. 9 shows a flowchart of the map data update process. When an operation for instructing the update of the map data is performed by the user of the in-vehicle device 20, the control unit 25 performs the process illustrated in FIG.

  First, a transmission request for updated map data is made to the map data generation device 10 (S400), and it is determined whether or not the updated map data has been received (S402).

When the updated map data is received, the map data is updated according to the updated map data (S404). When the map data update process is completed, this process ends.

  In addition, the travel control unit 25 of the in-vehicle device 20 mounted on the probe vehicle 2 is input from the distance image sensor 21 and the three-dimensional feature map data not including the moving object generated by the map data generation device 10. A moving object is detected on the basis of captured images around the vehicle, and a driving support process is performed for driving support so as to avoid the moving object.

  FIG. 10 shows a flowchart of the driving support process. When an operation for instructing the start of driving support is performed by the user of the in-vehicle device 20, the driving control unit 25 performs the process shown in FIG. 10 in parallel with the probe data collecting process shown in FIG.

  First, a travel plan is created (S500). Specifically, an optimum guide route from the departure point to the destination is searched according to the user operation, and the searched guide route is displayed on the display unit. In the present embodiment, the lane in which the mounted vehicle travels is specified on the guide route.

  Next, the mobile object detection process of S600 is performed. FIG. 11 shows a flowchart of this moving object detection process. In this moving object detection process, first, feature map data around the current position is acquired (S602). Specifically, a captured image with the viewpoint where the mounted vehicle is located is acquired from the map data storage unit 23.

  Next, forward image data is acquired from the distance image sensor 21 (S604). Here, the captured image and distance measurement data at the current position captured by the distance image sensor 21 are acquired.

  Next, with respect to the captured image with the viewpoint where the mounted vehicle acquired in S602 is located, and the captured image at the current position captured by the distance image sensor 21 acquired in S604, the correlation between the entire images. The number is calculated (S606). The calculation of the cross-correlation coefficient for the entire image can be performed as in S308.

  Next, it is determined whether or not the cross correlation coefficient of the entire image is equal to or greater than a predetermined threshold (S608).

  Here, when the cross-correlation coefficient in the entire image does not exceed the threshold value, the viewpoint of the data is moved (S610). Here, the captured image is moved from the viewpoint where the mounted vehicle acquired in S602 is positioned so that the cross-correlation coefficient of the entire image is equal to or greater than the threshold, and the process returns to S606.

  If the viewpoint is moved in this way and the cross-correlation coefficient in the entire image is equal to or greater than the threshold value, this viewpoint can be set as the current position of the mounted vehicle. In the present embodiment, the position detection accuracy is low in the configuration in which the current position is calculated based on information input from the position detection device 22 and the current position of the mounted vehicle is corrected by performing the map matching process. When the cross-correlation coefficient in the entire image is equal to or greater than the threshold value, the current position of the mounted vehicle is corrected so that the position of this viewpoint is the current position.

  As described above, when the cross-correlation coefficient of the entire image is equal to or greater than the threshold value, the captured image with the viewpoint where the mounted vehicle acquired in S602 is located, and the distance image sensor 21 acquired in S604. Each pixel data variance value is calculated for the photographed image at the current position taken by (S614). The calculation of each pixel data dispersion value can be performed in the same manner as in S312. That is, whether or not an object to be moved is included in the captured image based on whether or not all the dispersion values of the three primary colors (red (R), green (G), and blue (B)) are greater than or equal to a threshold value. Determination is made (S616).

  Here, when at least one of the three primary colors (red (R), green (G), and blue (B)) of the color of the specific pixel is greater than or equal to the threshold value, the determination in S616 is YES, and the distance It is determined that a moving object that can move in the captured image at the current position captured by the image sensor 21 is not detected (S618), and the process proceeds to S502 in FIG.

  If all the dispersion values of the three primary colors (red (R), green (G), and blue (B)) of the specific pixel are equal to or greater than the threshold, the determination in S6106 is NO, and the distance image The process proceeds to S502 in FIG. 10 without determining that a moving object capable of moving to the captured image at the current position captured by the sensor 21 has been detected.

  In S502 of FIG. 10, the planned position is calculated while avoiding the moving object. Here, first, a subject included in a captured image input from the distance image sensor 21 that does not exist in the three-dimensional feature map is recognized as a moving object, and a relative distance from the moving object is determined from the distance image sensor 21. Calculation is performed based on the input distance measurement data, and the planned travel position of the mounted vehicle is calculated so as to avoid moving objects. For example, as shown in FIG. 12, when the parked vehicle A is stopped on the shoulder of the lane of the destination and this parked vehicle A is recognized as a moving object, as shown by the arrow B in the figure. The planned travel position of the mounted vehicle is calculated so as to avoid this parked vehicle A.

  Next, the difference (error amount) between the planned travel position of the mounted vehicle calculated in S502 and the actual position of the mounted vehicle is calculated (S504).

  Next, the vehicle drive control unit 30 is instructed to reduce the difference between the planned travel position of the mounted vehicle and the actual position of the mounted vehicle (S506). Specifically, a signal that instructs the brake ECU, the steering ECU, and the engine ECU to reduce the difference between the planned travel position of the mounted vehicle and the actual position of the mounted vehicle is output.

  Next, it is determined whether or not the error amount is greater than or equal to a predetermined determination value (S508). If the error amount is larger than the determination value, the determination in S508 is NO and the process returns to S502. Therefore, the instruction to the vehicle drive control unit 30 is continued.

  If the error amount is equal to or smaller than the determination value, it is next determined whether the mounted vehicle has arrived at the destination based on whether the mounted vehicle has entered a certain area around the destination (S510).

  Here, when the mounted vehicle has not entered the fixed area around the destination, the determination in S510 is NO and the process returns to S600. Further, when the mounted vehicle enters a certain area around the destination, the determination in S510 is YES, and this process is terminated.

  According to the configuration described above, a plurality of probe data including a photographed image obtained by photographing the same place from the same spot is extracted from the probe data stored and stored in the storage medium, and the photographed image obtained by photographing the same place is specified. Based on a dispersion value of pixel data of pixels, a captured image that does not include a movable object that can move is specified, and probe data that includes a captured image that does not include the moving object is registered as map data generation data. Based on the probe data registered in the map data generation data, the absolute position of each area of the subject included in the captured image is calculated to generate three-dimensional feature map data that does not include a moving object. That is, it is possible to generate 3D map data that does not include a moving object.

  In addition, since the viewpoint data of the captured image included in the probe data stored and stored in the storage medium is moved to extract the probe data including a plurality of captured images obtained by capturing the same location from the same location, the captured data is somewhat at the captured position. Even if there is a discrepancy, it can be extracted as probe data including a photographed image obtained by photographing the same place from the same spot.

  In addition, the feature map data at the same place as the photographed image taken by the photographing means is acquired from the storage medium storing the three-dimensional feature map data generated by the map data generating device, and the photograph taken by the photographing means is taken. Based on the image and the variance value of the pixel data of the specific pixel of the feature map based on the feature map data, the movable vehicle is detected so as to avoid the detected moving object. Support is provided. That is, it is possible to perform the travel assistance of the mounted vehicle so as to avoid the moving object using the three-dimensional map data not including the moving object.

  In addition, the viewpoint of the photographed image taken by the photographing means so that the error amount between the feature map around the mounted vehicle specified by the three-dimensional feature map data and the photographed image photographed by the photographing means is less than the threshold value. , And the current position of the mounted vehicle specified by the current position specifying means is corrected so that the position of this viewpoint is the current position, so that the current position of the mounted vehicle can be specified with high accuracy.

  In addition, this invention is not limited to the said embodiment, Based on the meaning of this invention, it can implement with a various form.

  For example, in the above-described embodiment, a configuration has been described in which communication is performed with an unspecified number of vehicles equipped with a data collection device, and map data is generated using information collected by the data collection device. The information collected by the data collection device may not be acquired. For example, the information collected by the data collection device may be acquired via a storage medium such as a memory card.

  Moreover, in the said embodiment, when the map data generation apparatus 10 receives probe data, the probe data containing the some picked-up image image | photographed in the same place as the picked-up image contained in the received probe data from a storage medium. Based on the variance value of the pixel data for each area of the plurality of photographed images extracted and photographed at the same place, a photographed image that does not include a movable object that can be moved is identified, and the movable object is included Although a configuration has been shown in which probe data including an uncaptured captured image is registered as map data generation data, the above-described processing may be performed at a timing other than the timing at which probe data is received.

  In the above embodiment, the variance value for each color information is calculated as the pixel data. However, the variance value of the luminance information may be calculated as the pixel data, and the variance values of both the color information and the luminance information are calculated. You may make it do.

  In the above embodiment, the cross-correlation coefficient for each captured image read in S306 is calculated in S308 using the captured image included in the received probe data as a template. It is not limited to such a method.

  In the above-described embodiment, the example in which the map data generation device 10 generates the three-dimensional feature map data that does not include the movable object that can move is described. However, the three-dimensional ground map that does not include the moving object has been described. Two-dimensional feature map data may be generated from the map data.

  The correspondence relationship between the configuration in the above embodiment and the configuration of the claims will be described. S200 and S302 correspond to storage processing means, S304 to S318 correspond to map data generation data registration means, and S202. The distance map sensor 21 corresponds to the feature map data generation means, the distance image sensor 21 corresponds to the photographing means, S602 corresponds to the feature map data acquisition means, S614 to S618 correspond to the moving object detection means, and S502 to S510 travel. The processing for specifying the current position of the mounted vehicle based on the information for specifying the current position input from the position detection device 22 corresponds to the current position specifying means. S606 to S610 are viewpoint moving means. If the cross-correlation coefficient of the entire image is equal to or greater than the threshold due to the viewpoint movement in S610, the position of this viewpoint is set as the current position. Sea urchin, processing for correcting the current position of the equipped vehicle corresponding to the current position correction means.

DESCRIPTION OF SYMBOLS 1 Information center 2 Probe vehicle 10 Map data generation apparatus 11 Map database 12 Communication part 13 Control part 20 Car-mounted apparatus 21 Distance image sensor 22 Position detection apparatus 23 Map data storage part 24 Communication part 25 Control part 30 Vehicle drive control part

Claims (5)

A map data generation device that acquires information collected by the data collection device from a plurality of vehicles equipped with a data collection device, and generates map data using the acquired information,
The data collection device specifies image data representing a photographed image obtained by photographing the periphery of the mounted vehicle, distance measurement data for identifying the distance between each part of the subject of the photographed image, and the position and orientation of the mounted vehicle. To collect probe data in association with positioning data for
When receiving the probe data collected by the data collection device, storage processing means for accumulating and storing the received probe data in a storage medium;
From the probe data stored and stored in the storage medium, a plurality of probe data including a photographed image obtained by photographing the same place from the same spot is extracted, and specific pixels of the photographed image obtained by photographing the same place from the same spot are extracted. Map data generation that identifies a captured image that does not include a movable object that can move based on the variance of pixel data, and registers probe data that includes the captured image that does not include the moving object as map data generation data Data registration means,
A ground for generating three-dimensional feature map data not including the moving object by calculating the absolute position of each area of the subject included in the captured image based on the probe data registered in the map data generation data A map data generation device comprising: a drawing data generation unit;   The map data generation data registration means includes probe data including a plurality of photographed images obtained by photographing the same place from the same spot by moving the viewpoint of the photographed image contained in the probe data stored and stored in the storage medium. The map data generation device according to claim 1, wherein the map data generation device is extracted.   The map data generation apparatus according to claim 1, wherein the pixel data is at least one of color data and luminance data. Photographing means for photographing the periphery of the vehicle;
A feature for acquiring feature map data at the same location as a photographed image photographed by the photographing means from a storage medium storing three-dimensional feature map data generated by the map data generating device according to claim 1 Figure data acquisition means;
It is possible to move based on a variance value of pixel data of specific pixels of a feature map based on a photographed image photographed by the photographing means and the feature map data obtained by the feature map data obtaining means. Moving object detection means for detecting a moving object;
A travel support device, comprising: travel support means for supporting travel of the mounted vehicle so as to avoid the moving object detected by the moving object detection means. Current position specifying means for specifying the current position of the mounted vehicle;
Based on the current position of the mounted vehicle specified by the current position specifying means, the three-dimensional feature map data around the mounted vehicle is acquired, and the mounted vehicle periphery specified by the three-dimensional feature map data Viewpoint moving means for moving the viewpoint of the photographed image photographed by the photographing means so that an error amount between the feature map of the feature image and the photographed image photographed by the photographing means is less than a threshold value;
The travel support apparatus according to claim 4, further comprising: a current position correcting unit that corrects a current position of the mounted vehicle so that a position of the viewpoint moved by the viewpoint moving unit is a current position. .

Images