JP2009180631A - Navigator, navigation method and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a navigator capable of guiding a route more precisely than conventional ones even in cases of incomplete database of road markings. <P>SOLUTION: The navigator 10 comprises a vehicle location computing section 12 computing the actual location of a vehicle; a road marking DB26 storing the data of road markings; a road marking detection section 14 detecting a road marking from images of a camera 40 mounted on the vehicle; a checking section 24 identifying the location of the road marking detected by checking whether there is the same kind of road marking as the road marking detected by the road marking detection section 14 within the road markings in the prescribed searching range; and a guide content decision section 30 guiding a route based on the location of the road marking when the location of the detected road marking is determined and deciding to guide a route based on the actual location of the vehicle computed by the vehicle position computing section when the location of the detected road marking is not identified. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a navigation apparatus that performs route guidance, and more particularly, to an apparatus mounted on a vehicle or the like.

  Conventionally, a car navigation device that detects a current position of a vehicle and performs route guidance based on the current position is known. Conventional car navigation apparatuses often use mainly the following two methods as a positioning method of the current position of the vehicle.

  The first method is a method using GPS (Global Positioning System). In this method, positioning is performed by receiving signals from a plurality of GPS satellites. In principle, if signals from four GPS satellites can be received, the latitude, longitude, and altitude of the current position can all be determined. However, in this method, a signal from a GPS satellite cannot be stably received in an urban center where there are many tunnels or buildings, or under an elevated structure, and positioning cannot be performed. This drawback is compensated by the following second method.

  The second method is a method of calculating the movement trajectory of the vehicle by integrating the values obtained from the vehicle speed pulse sensor and the gyroscope. In this method, since errors are accumulated simply by integration, correction is performed by comparing with the shape of the route on the map database (map matching). The navigation device calculates the vehicle position even during a period in which the vehicle cannot receive a signal from the GPS satellite by adding the movement locus obtained by the second method to the vehicle position obtained by GPS.

  However, map matching is very effective when the trajectory is characteristic, for example, when the vehicle frequently turns at a right angle. When entering, or when traveling in a grid-like urban area, the position calculation may fail. In addition, since the gyroscope cannot determine that the direction of the vehicle has changed on a turntable in a parking lot or the like, it is not possible to obtain the movement locus of the vehicle.

  Due to the circumstances as described above, the positioning technology using the car navigation device cannot be said to be accurate enough to provide detailed route guidance. For example, in a city area or the like, when there are a plurality of intersections at a short distance, it is not known whether the vehicle is at an intersection to bend or at an intersection just before that. Therefore, in a conventional car navigation apparatus, for example, “the next intersection is on the right” is not guided, but “100 meters ahead, it is on the right” or the like is used.

One technique for improving the positioning accuracy of a vehicle is disclosed in Patent Document 1. In the positioning method described in Patent Document 1, the positioning accuracy is improved by detecting a road marking painted on the road.
JP 2007-108043 A

  However, the number of road markings painted on the road is very large, and it is very difficult to prepare a database that covers the positions where road markings exist. Also, since road markings are frequently repainted, it is more difficult to keep the database up to date. It is also assumed that the road marking cannot be detected properly due to the deterioration of the road marking.

  Therefore, when the method described in Patent Document 1 described above is applied to route guidance, it is assumed that a problem arises in route guidance due to the fact that data in the database is missing. .

  Therefore, an object of the present invention is to provide a navigation device that can provide route guidance with higher accuracy than conventional methods even when a road marking database is incomplete.

  A navigation device of the present invention includes a vehicle position calculation unit that calculates a current position of a vehicle, a road marking database that stores data indicating the position where the road marking exists and the type of the road marking, and a camera mounted on the vehicle. A road marking detection unit for detecting a road marking from the acquired image, and a road marking within a predetermined search range including the current position of the vehicle obtained by the vehicle position calculation unit. A collating unit that reads out from the road marking database and identifies the position of the detected road marking by collating whether or not there is a road marking of the same type as the road marking detected by the road marking detecting unit in the read road marking If the position of the road marking is identified by the collation by the collating unit, route guidance is performed based on the position of the road marking, and the road marking is identified. If no is provided with a route guidance unit for performing route guidance based on the current position of the vehicle calculated by the vehicle position calculating unit.

  Thus, when the position of the road marking detected from the camera image is specified, the route guidance can be improved based on the route guidance based on the road marking. If the detected road marking position is not specified, route guidance is performed based on the current vehicle position calculated by the vehicle position calculation unit. The route guidance can be performed without any delay.

  In the navigation device according to the present invention, the verification unit includes a plurality of road markings having the same arrangement as the plurality of road markings detected by the road marking detection unit in the road markings read from the road marking database. You may collate.

  Thus, by collating a plurality of road markings, even when there are a plurality of road markings of the same type, the possibility of specifying the positions of the detected road markings can be increased.

  In the navigation device of the present invention, the type of the road marking detected by the road marking detection unit and the relative coordinates of the road marking with respect to the vehicle are written in the storage unit, and the data of the relative coordinates is recorded as the vehicle travels. The road marking update part to update may be provided, and the collation part may collate the road marking memorize | stored in the said memory | storage part with the road marking read from the said road marking database.

  By storing the road marking data detected by the road marking detection unit in this way, a plurality of road markings can be appropriately collated.

  In the navigation device of the present invention, the road marking update unit receives the road marking data from the storage unit when the position of the road marking stored in the storage unit is away from the current position of the vehicle by a predetermined threshold or more. It may be deleted.

  With this configuration, the amount of road marking data stored in the storage unit can be appropriately managed.

  The navigation device of the present invention stores in the road marking database data on the direction of the road with the road marking in association with the road marking, and the collating unit performs road marking with the road marking detecting unit. The road marking associated with the direction of the road that matches the traveling direction of the vehicle at the time of detection is read from the road marking database, and the read road marking is compared with the road marking detected by the road marking detection unit. Also good.

  By using road direction data in this way, it is possible to reduce the amount of data read from the road marking database as a target for verification, and to reduce the burden of verification processing.

  In the navigation device of the present invention, the vehicle position calculation unit is connected to a GPS that calculates the absolute position coordinates of the vehicle and a sensor that outputs a signal used to calculate the movement trajectory of the vehicle, and the vehicle calculated by the GPS. The current position of the vehicle is calculated by adding the movement trajectory obtained from the output signal of the sensor to the absolute position coordinates, and the search range used in the verification unit has elapsed since the absolute position coordinates were obtained by the GPS It is good also as a structure changed according to time.

  Since the calculation of the movement locus of the vehicle includes an error, the value of the current position of the vehicle calculated by the vehicle position calculation unit includes an error as time elapses after the absolute position coordinates are obtained by GPS. Accumulate. According to the configuration of the present invention, the search range can be changed according to the magnitude of the current vehicle position detection error. For example, when collation is performed immediately after GPS positioning, since the current position calculation error is small, the search range can be narrowed and collation can be performed quickly. In addition, when collating after a long time has elapsed since GPS positioning, the calculation error of the current position is large, so the search range is widened so that the road marking to be searched does not leak from the search range Can be.

  In the navigation device of the present invention, the road sign detection unit cuts out a window area for road sign detection from an image acquired by the camera, obtains a horizontal frequency distribution of the image of the window area, and in the frequency distribution A crosswalk may be detected as the road marking when the intensity of a specific frequency component is equal to or greater than a predetermined threshold.

  Since the pedestrian crossing has a periodic brightness in the horizontal direction, the frequency distribution in the horizontal direction of the image has a peak in a specific frequency component. Therefore, with the configuration of the present invention, a pedestrian crossing can be detected appropriately.

  The navigation device of the present invention includes a reference image storage unit that stores reference images of various road markings, and the road marking detection unit cuts out a window region for road marking detection from an image acquired by the camera, The road marking may be detected by matching the image of the window area and the reference image stored in the reference image storage unit.

  By performing matching with the reference image in this way, various road markings can be detected. In addition, you may match in order from the road marking of the kind detected with high frequency. Thereby, the processing time required for matching can be shortened. Further, the reference images may be grouped according to the similarity, and when it is determined that the similarity with a certain road marking is high, the road markings of the same group may be preferentially matched.

  In the navigation device of the present invention, the vehicle position calculation unit is connected to a GPS that calculates the absolute position coordinates of the vehicle and a sensor that outputs a signal used to calculate the movement trajectory of the vehicle, and the vehicle calculated by the GPS. The current position of the vehicle is calculated by adding the movement locus obtained from the output signal of the sensor to the absolute position coordinates of the road, and when the position of the road marking is specified by the verification in the verification unit, the road marking The vehicle position may be detected by recalculating the current position of the vehicle based on the position and adding the movement locus to the recalculated position until the absolute position coordinates of the vehicle are obtained by the GPS.

  When the position of the road marking detected from the camera image is specified in this way, the absolute position coordinates of the vehicle are obtained by GPS by calculating the position of the vehicle based on the positional relationship between the road marking and the vehicle. It is possible to calculate the position with high accuracy even during a period when it is not.

  The navigation device according to the present invention includes an aerial image or satellite image whose relationship with absolute coordinates on the ground is known, or an image acquisition unit that acquires a camera image outside the vehicle to which position information of the location where the image was taken is added, and the image acquisition unit A road marking database update unit that detects road markings from the image acquired in step S1 and stores data indicating the types and positions of the detected road markings in the road marking database.

  With this configuration, a road marking database can be automatically created. If the image quality is poor or the road marking itself is worn out, it may not be possible to detect the road marking from the image. If the road marking database is created automatically, all road markings are covered. However, in the present invention, there is no problem because route guidance can be performed appropriately even when there is a defect in the road marking database. That is, it is suitable as a method for constructing a road marking database used in the navigation device of the present invention.

  The navigation device of the present invention includes a search area limiting unit that limits a region for searching for a road marking to a region where a road is displayed among images acquired by the image acquisition unit using road map information. Also good.

  By limiting the search area in this way, the process of searching for road markings can be reduced.

  The navigation device of the present invention stores data indicating the type and position of the road sign detected by the road sign detection unit when it is determined by the collation in the collation unit that the same type of road sign does not exist. Based on the unmatched road marking database and the data stored in the unmatched road marking database, when the number of times the same type of road marking is detected at the same position exceeds a predetermined threshold, the road marking Data may be added to the road marking database.

  Thus, by adding the road marking detected by the camera mounted on the vehicle to the road marking database, it is possible to compensate for the loss of road marking data and to improve the reliability of the road marking database.

  The navigation device according to another aspect of the present invention includes a vehicle position calculation unit connected to a GPS for obtaining an absolute position coordinate of a vehicle and a sensor that outputs a signal used to calculate a movement locus of the vehicle, and a road marking. A road marking database that stores data indicating the position and the type of the road marking; a road marking detection unit that acquires a road image captured by a camera mounted on the vehicle and detects the road marking from the acquired image; The road marking within the predetermined search range including the current position of the vehicle obtained by the vehicle position calculation unit is read from the road marking database, and the road marking detected by the road marking detection unit in the read road marking A collation unit that identifies the detected road sign by collating whether there is a road sign of the same type as the vehicle position calculation unit, the vehicle position calculation unit The current position of the vehicle is calculated by adding the movement locus calculated from the output signal of the sensor to the absolute position coordinates of the road, and when the position of the road marking is specified by the verification in the verification unit, the road The current position of the vehicle is recalculated based on the position of the marking, and the vehicle position is detected by adding a movement locus to the recalculated position until the absolute position coordinates of the vehicle are obtained by the GPS.

  When the position of the road marking detected from the camera image is specified in this way, the absolute position coordinates of the vehicle are obtained by GPS by calculating the position of the vehicle based on the positional relationship between the road marking and the vehicle. It is possible to calculate the position with high accuracy even during a period when it is not.

  According to the navigation method of the present invention, a step of preparing a road marking database storing data indicating a position where a road marking exists and a type of the road marking, and an image of a road photographed by a camera mounted on the vehicle are acquired. A step of detecting a road sign from the acquired image, a step of detecting the current position of the vehicle based on the absolute position coordinates obtained by the GPS and the movement locus of the vehicle, and a predetermined value including the current position of the vehicle The road marking within the search range is read from the road marking database, and the position of the detected road marking is determined by checking whether the read road marking has the same type of road marking as the road marking detected from the image. If the position of the road marking is identified by the identifying step and collation, route guidance is performed based on the position of the road marking. , When the position of the road sign is not specified, and a step of performing route guidance based on the current position of the vehicle.

  With this configuration, as with the navigation device of the present invention, the accuracy of route guidance can be improved, and route guidance can be performed without delay even when data in the road marking database is missing. Various configurations of the navigation device of the present invention can be applied to the navigation method of the present invention.

  The program of the present invention is a program for performing route guidance, the step of acquiring a road image captured by a camera mounted on a vehicle in a computer, and the step of detecting a road marking from the acquired image; A step of acquiring data on the current position of the vehicle from a vehicle position calculation unit that calculates the current position of the vehicle based on the absolute position coordinates obtained by the GPS and the movement locus of the vehicle, and a predetermined value including the current position of the vehicle Road markings within the search range are read from a road marking database storing data indicating the position where the road marking exists and the type of the road marking, and the road detected by the road marking detection unit in the read road marking Identifying the position of the detected road marking by collating whether there is a road marking of the same type as the sign, and the position of the road marking by collation If the position is specified, route guidance is performed based on the position of the road marking. If the position of the road marking is not specified, route guidance is performed based on the calculated current position of the vehicle. Step.

  With this configuration, as with the navigation device of the present invention, the accuracy of route guidance can be improved, and route guidance can be performed without delay even when data in the road marking database is missing. Various configurations of the navigation device of the present invention can be applied to the program of the present invention.

  According to the present invention, it is possible to improve the accuracy of route guidance by specifying road markings, and to perform route guidance without delay even when data in the road marking database is missing. Has an effect.

Hereinafter, a navigation device according to an embodiment of the present invention will be described with reference to the drawings.
FIG. 1 is a diagram showing a configuration of a navigation device 10 according to an embodiment of the present invention. A camera 40 mounted on the vehicle is connected to the navigation device 10. The navigation device 10 acquires an image of a road around the vehicle (hereinafter referred to as “camera image”) taken by the camera 40. For example, the camera 40 is attached to the inside of the windshield of the vehicle and has an imaging range centered on a road ahead of the vehicle.

  The navigation device 10 includes a vehicle position calculation unit 12 that calculates a current position of the vehicle, a road sign detection unit 14 that detects a road sign from a camera image, and relative coordinates of the detected road sign (coordinates with the vehicle position as an origin) A relative coordinate calculation unit 18 for obtaining a position coordinate in the system, a detection data update unit 20 for storing detected road marking data in the temporary storage unit 22, a position (latitude and longitude) where the road marking is painted, and its road A road marking database (hereinafter referred to as “road marking DB”) 26 storing the type of marking is compared with the road marking data stored in the temporary storage unit 22 and the data in the road marking DB 26, and detected from the camera image. A collation unit 24 that identifies the position of the road marking, a guidance content determination unit 30 that determines the content of the route guidance based on the identified position of the road marking, and a route guidance And a route information output unit 32 for performing. Further, the navigation device 10 includes a road marking database detection data update unit 20 (hereinafter referred to as “road marking DB update unit”) 36 for updating the road marking DB 26 and a communication unit 38 that communicates with the server 50. ing.

  The vehicle position calculation unit 12 is connected to the GPS 42, the gyroscope 44, and the vehicle speed pulse sensor 46, and receives signals from these sensors. The GPS 42 receives a signal from a GPS satellite and obtains the absolute position coordinates of the vehicle based on the received signal. When the signal from the GPS satellite cannot be received, the movement trajectory of the vehicle is obtained using the output signals from the gyroscope 44 and the vehicle speed pulse sensor 46, and the movement trajectory is added to the absolute position coordinates obtained by the GPS 42, Detect the current position of the vehicle. In addition to the configuration shown in FIG. 1, the vehicle position calculation unit 12 may calculate the vehicle position using signals from a geomagnetic sensor, a distance sensor, and the like, for example.

  The road sign detection unit 14 has a function of detecting a road sign from the camera image. The road marking detection unit 14 is connected to a road marking template database (hereinafter referred to as “road marking template DB”) 16 that stores road marking templates (reference images). The road sign detection unit 14 cuts out an image by gradually shifting the detection processing target area (window area) from the camera image, and performs matching between the cut-out image and the road sign template stored in the road sign template DB 16. Thus, a road sign is detected from the camera image.

  When the road marking is a pedestrian crossing as shown in FIG. 2A, the brightness periodically changes in the horizontal direction of the image (see FIG. 2B). As shown in FIG. 2A, the road marking detection unit 14 cuts out a window region extending in the horizontal direction, obtains a horizontal frequency distribution of the image by Fourier transform (see FIG. 2C), and the frequency distribution has a peak. Is determined by template matching. That is, when detecting a pedestrian crossing, a frequency distribution template is used. With this configuration, the road sign detection unit 14 can appropriately detect a pedestrian crossing. The road sign detection unit 14 passes the road sign detection result to the relative coordinate calculation unit 18. A detailed description of the processing of the road sign detection unit 14 will be described later.

  The relative coordinate calculation unit 18 obtains the relative coordinates of the road sign with the vehicle position as the origin by obtaining the distance and direction to the road sign detected by the road sign detection unit 14.

  The detection data update unit 20 writes the detected road marking type and its relative coordinates in the temporary storage unit 22. In addition, the detection data update unit 20 sequentially updates the relative coordinates of the road marking stored in the temporary storage unit 22 using the vehicle movement trajectory data obtained by the vehicle position calculation unit 12 of the vehicle. When the absolute value of the relative coordinate becomes a predetermined value or more (for example, 15 m or more), that is, when the distance between the road marking and the vehicle position becomes a predetermined value or more, the detection data update unit 20 Delete road marking data. In the present embodiment, the purpose is to provide route guidance with high accuracy based on the detected position of the road marking, so only the road marking data in the vicinity of the vehicle need be stored.

  FIG. 3A is a diagram illustrating an example of data stored in the temporary storage unit 22. The temporary storage unit 22 stores an ID for identifying the detected road marking, the type of the road marking, and data of a relative x coordinate and a relative y coordinate. FIG. 3B is a diagram showing an example of a road marking corresponding to the data shown in FIG. The origin O indicates the current position of the vehicle 48. The current position O of the vehicle is a value calculated by the vehicle position calculation unit 12. When the vehicle 48 moves in the direction of the arrow, the value of the relative x coordinate becomes a negative value corresponding to the distance traveled.

  The collation unit 24 reads road marking data within the search range including the current position of the vehicle from the road marking DB 26. The search range is determined by the search range determination unit 28. The search range determination unit 28 acquires data on the current position of the vehicle from the vehicle position calculation unit 12 and determines a circular area having a predetermined radius (for example, 100 m) centered on the current position as the search range. Then, the collation unit 24 collates whether the road markings read from the road marking DB 26 include a plurality of road markings having the same arrangement as the plurality of road markings stored in the temporary storage unit 22. For example, when there are road signs arranged in the order of “pedestrian crossing”, “stop line”, and “stop” in the temporary storage unit 22, “pedestrian crossing” and “stop line” arranged in the same order and at almost the same interval. "Check if there is a" stop "road sign. In addition, when there are road signs arranged in the order of “pedestrian crossing”, “stop line”, and “stop” at almost the same interval, the collation unit 24 determines that the road sign is not specified. . The collation unit 24 passes the collation result to the guidance content determination unit 30.

  The guidance content determination unit 30 determines the guidance content based on the collation result. When the position of the detected road marking is specified by the collation unit 24, detailed route guidance (hereinafter referred to as “detail level route” based on the position of the vehicle ascertained from the positional relationship between the road marking and the vehicle. Decide to carry out "guidance"). If the detected position of the road marking is not specified by the collation unit 24, based on the current position of the vehicle calculated by the vehicle position calculation unit 12, the same route guidance (hereinafter referred to as “normal level”) is used. It is decided to carry out "route guidance".

  The route information output unit 32 outputs route information according to the mode determined by the guidance content determination unit 30. Note that the route information output unit 32 performs the same route guidance as before even when the verification result notification is not received from the verification unit 24. For example, in places where route guidance is necessary, such as places to be turned right, route guidance is performed without detecting road markings.

  The communication unit 38 has a function of performing communication with the server 50 having the road marking DB 26. The communication unit 38 downloads the data of the road marking DB 52 stored in the server 50 and stores it in the road marking DB 26.

  The road marking DB update unit 36 updates the road marking data detected by the road marking detection unit 14 but not verified to the road marking DB 26. An unmatched road marking database (hereinafter referred to as “unmatched road marking DB”) 34 is connected to the road marking DB update unit 36. In the unmatched road sign DB 34, road sign data (detected position, type of road sign, etc.) not matched by the matching unit 24 is stored as an unmatched road sign. The road marking DB update unit 36 reads data from the unverified road marking DB 34, and when the number of times the same type of road marking is detected at the same position is equal to or greater than a predetermined threshold, the road marking data is updated to the road marking DB 26. Write to. The communication unit 38 may transmit data stored in the unverified road marking DB 34 to the server 50. As a result, the server 50 can update the road marking DB 52 using the data of unmatched road markings transmitted from the plurality of navigation devices 10.

  FIG. 4 is a diagram illustrating an operation of the navigation device 10 according to the present embodiment. The operation of the navigation device 10 will be described with reference to FIG. The navigation apparatus 10 acquires a camera image from the camera 40 mounted on the vehicle (S10). The road sign detection unit 14 of the navigation device 10 performs a process of detecting a road sign from the acquired camera image. Hereinafter, the operation of the road sign detection unit 14 will be described in detail.

  The road marking detection unit 14 cuts out a window region for performing road marking detection processing from the camera image (S12). The size of the window area to be cut out can be determined in advance based on the mounting position of the camera 40 and the shooting direction. For example, when the camera 40 is mounted downward, the subject appears larger as it goes down in the frame, so 50 × 50 pixels in the lower area of the frame, 30 × 30 pixels in the upper area of the frame, etc. Thus, the size of the window area is determined so that the ratio of the size of the road marking to the window area is constant. Moreover, the road marking detection part 14 cuts out the window area | region extended in a horizontal direction, when detecting a pedestrian crossing. Since the road marking exists only on the road, the road marking detection unit 14 may not cut out a region where no road is reflected as a window region.

  Next, the road marking detection unit 14 performs distortion correction on the image in the cut window region W. As shown in FIG. 5A, in the camera image, the road marking painted on the road appears distorted. In order to appropriately perform template matching, the road marking detection unit 14 performs distortion correction on the image of the cut window region W (see FIG. 5B) to obtain an image as shown in FIG.

  Next, the road marking detection unit 14 detects whether or not a pedestrian crossing is included in the image using the image after distortion correction (S14). The frequency distribution of the image is obtained by Fourier-transforming the image in the window region W in the horizontal direction. The road marking detection unit 14 compares the frequency distribution of the image of the window region W with a frequency distribution template obtained by Fourier transforming the image of the reference pedestrian crossing, and detects the pedestrian crossing.

  Subsequently, the road sign detection unit 14 detects whether a road sign other than the pedestrian crossing is included in the image. Template matching for comparing the road marking template stored in the road marking template DB 16 with the image of the window area is performed (S16). If the similarity is equal to or greater than a predetermined threshold value, it is determined that the image of the window area matches the template, and a road marking corresponding to the template is detected.

  When the image of the window area matches the template (YES in S18), the road sign detection unit 14 passes the road sign detection result to the relative coordinate calculation unit 18. The relative coordinate calculation unit 18 obtains the distance and direction to the road marking based on the angle of the imaging direction of the camera 40 and the cutout position of the window area where the road marking is detected, and obtains the relative coordinates of the road marking viewed from the vehicle. (S20).

  The relative coordinate calculation unit 18 passes the road sign type and relative coordinate data to the detection data update unit 20, and the detection data update unit 20 writes the data indicating the road sign type and position in the temporary storage unit 22 (S22). . The detection data update unit 20 also performs a process of updating the data stored in the temporary storage unit 22, which will be described later with reference to FIG.

  The road marking detection unit 14 determines whether the entire camera image is scanned by the window area (S24). If the scan is not completed (NO in S24), the window area is shifted and an image included in the window area is cut out (S12). After cutting out the image of the window area, the operation is the same as described above.

  When the entire scan of the camera image is completed (YES in S24), the detected road marking and the road marking DB 26 are collated (S28), and the collation result is passed to the guidance content determination unit 30. The guidance content determination unit 30 determines the guidance content based on the collation result. When the position of the road marking detected from the camera image is specified, it is determined to perform detailed level route guidance. When the position of the road marking is not specified, it is determined to perform normal level route guidance based on the current position of the vehicle calculated by the vehicle position calculation unit 12. The route information output unit 32 outputs route information according to the mode determined by the guidance content determination unit 30.

  FIG. 6 is a diagram illustrating an operation of updating the temporary storage unit 22 by the detection data update unit 20. The detection data update unit 20 reads data from the temporary storage unit 22 (S40). The detection data updating unit 20 calculates the movement trajectory of the vehicle from the previous update of the temporary storage unit 22 (S42), and updates the relative coordinates using the calculated movement trajectory (S44). Next, the detection data update unit 20 deletes road marking data whose absolute value of the relative coordinate is equal to or greater than a predetermined threshold (S46), and writes the remaining data in the temporary storage unit 22 (S48). Heretofore, the navigation device 10 according to the present embodiment has been described.

  The navigation device 10 according to the present embodiment compares the road marking detected from the camera image with the road marking DB 26 to identify the position of the road marking, thereby accurately grasping the current position of the vehicle, and at a detailed level. Route guidance can be performed.

  For example, when there are a plurality of intersections at a short distance, the navigation apparatus 10 according to the present embodiment can calculate the position calculation accuracy by the vehicle position calculation unit 12 even if it is not known which intersection the vehicle is near. According to the above, based on the position of the road marking, it is possible to accurately detect which intersection has passed and which intersection is in front. For example, “Next intersection is on the right.” Yes. As another example, a lane for a right turn vehicle may be provided in front of an intersection. In such a case, it is necessary for the right turn car to travel on the right lane, but if the current position of the vehicle is not detected accurately, the presence or absence of the right turn lane cannot be grasped, and an instruction is given quite before the intersection without the right turn lane. It may end up being put out. According to the navigation device 10 of the present embodiment, the detection of the road marking indicates that the vehicle has entered the area having the right turn lane, so the guidance “Please change to the right lane” at an appropriate timing. It can be performed.

  In addition, the navigation device 10 of the present embodiment provides normal level route guidance when no road marking corresponding to the detected road marking is found, and therefore when the data in the road marking DB 26 is missing. However, route guidance can be performed without any delay.

  Next, a configuration for generating the road marking DB 52 used in the present embodiment will be described. Here, the example which produces | generates road marking DB52 which the server 50 has is demonstrated.

  FIG. 7 is a diagram showing a configuration of a road marking database generation device (hereinafter referred to as “road marking DB generation device”) 60. A plurality of road image databases (hereinafter referred to as “road image DB”) 72 that are data sources of the road marking DB 52 are connected to the road marking DB generation device 60. The road image DB 72 stores an image whose correspondence with the absolute position coordinates on the ground is known. Specific examples of the image stored in the road image DB 72 are, for example, a satellite image or aerial image that has been aligned with a map, an out-of-vehicle camera image to which position information photographed by the high-precision GPS 42 is added, and the like. is there.

  The road marking DB generating device 60 includes a map database (hereinafter referred to as “map DB”) 64 that stores road maps, a search area limiting unit 62 that limits a search area for road markings using data in the map DB 64, roads A road sign detection unit 66 for detecting a road sign from the image, a road sign template database (hereinafter referred to as “road display template DB”) 68 storing a road sign template used in the road sign detection unit 66, and a plurality of And an integration unit 70 that integrates road marking data detected from each of the road image DBs 72.

  When the road marking DB generating device 60 acquires road image data from the connected road image DB 72, first, the search area limiting unit 62 limits the area for searching for road marking to only the road area. That is, the search area limiting unit 62 determines an area with a road using the data of the map DB 64 and limits the search area by superimposing the area on the road image. When the road image is a ground image, the road may not be reflected in the upper area in the image depending on the imaging angle. In that case, the search area limiting unit 62 removes the upper area of the image from the search area.

  The road sign detection unit 66 detects a road sign from the road image by the same processing as the road sign detection unit 14 of the navigation device 10 described above. In satellite images and aerial images, the direction of the pedestrian crossing is not necessarily the horizontal direction of the screen. Therefore, when detecting a pedestrian crossing, two-dimensional Fourier transformation is performed on the image of the window area, instead of performing Fourier transformation in the horizontal direction. As a result, when a frequency distribution having a peak is obtained, it is determined that a pedestrian crossing is included.

  The road marking detection unit 66 obtains the absolute position coordinates of the detected road marking using the correspondence between the absolute position coordinates on the ground and the road image. Then, the road sign detection unit 66 passes the detected road sign data to the integration unit 70.

  The integration unit 70 integrates the road marking data detected by the plurality of road marking detection units 66. Various methods are conceivable as integration methods, but in this embodiment, the integration unit 70 takes the logical sum or majority of the detection results by the respective road marking detection units 66. Thereby, even if there is a road marking that could not be detected from the data of a certain road image DB 72, it can be supplemented with the detected road marking from another road image DB 72.

  The road marking DB generation device 60 according to the present embodiment can automatically detect a road marking from a road image and generate a road marking DB 52. Further, by using a plurality of road image DBs 72, road markings that cannot be detected by one road image DB 72 can be complemented by other road image DBs 72, so that the quality of the road marking DB 52 can be improved.

  In addition, by downloading the road marking DB 52 generated by the road marking DB generation device 60 to the navigation device 10 of the present embodiment and using it, the route guidance can be routed without delay even when there is data loss in the road marking DB 52. Can be done. Therefore, the road sign DB generating device 60 is suitable as a device for generating the road sign DB 26 used in the navigation device 10 of the present embodiment.

  The navigation apparatus and the navigation method of the present invention have been described in detail with reference to the embodiment, but the present invention is not limited to the above-described embodiment.

  In the embodiment described above, the navigation device 10 has been described, but a program for realizing the function of the navigation device 10 described above is also included in the scope of the present invention.

  In the above-described embodiment, the search range determination unit 28 determines a circular area having a predetermined radius centered on the current position of the vehicle as the search range. However, the search range may be determined by another method. For example, the radius of the search range may be increased as the elapsed time after the absolute coordinate position is detected by the GPS 42 increases. In the method of obtaining the movement locus of the vehicle using the output signals from the gyroscope 44 and the vehicle speed pulse sensor 46, the error increases with time. Therefore, the search range is enlarged and the road marking to be searched leaks from the search range. I can not. In addition, since the error is small when the elapsed time since the absolute coordinates are acquired by the GPS 42 is small, the search range can be reduced to reduce the burden of collation processing.

  Further, the search range determination unit 28 does not necessarily have to be a circular area, and may determine the shape of the search range according to the traveling direction of the vehicle. For example, the search range may be a fan-shaped shape that expands in the traveling direction of the vehicle or a rectangular shape that extends along the traveling direction.

  In the above-described embodiment, the direction of the road on which the road marking is painted may be stored in the road marking DB 26 in association with the road marking. Thereby, when collating in the collation part 24, the object of the road marking which should be collated can be narrowed down based on the data of the advancing direction of a vehicle, and the direction of a road. For example, a road marking that may match the road marking detected when the vehicle is moving westward is a road marking painted on a road extending east-west, so the search target is on a road extending east-west It can be narrowed down to only road markings. The traveling direction of the vehicle may be obtained by using a magnetic sensor, or may be obtained by specifying a traveling road by map matching with a map.

  In the above-described embodiment, when a road marking is specified by the collating unit 24, the specified road marking data is fed back to the vehicle position calculating unit 12, and the road marking specified by the vehicle position calculating unit 12 is used. The current position of the vehicle may be obtained. Specifically, the current position of the vehicle is calculated based on the position data of the road marking and the relative coordinates of the road marking. Thereby, the current position of the vehicle can be calculated with high accuracy based on the position of the road marking. Further, after obtaining the current position of the vehicle using the road marking, the vehicle position calculation unit 12 may calculate the current position of the vehicle by adding a movement locus to the position of the vehicle.

  In the embodiment described above, the collation unit 24 determines that the position of the road marking cannot be specified when there are a plurality of road marking candidates that match the detected road marking, and performs normal level route guidance. However, when there are a plurality of candidates, it is determined that the matching error matches the minimum road marking, and the guidance content determination unit 30 provides detailed level route guidance based on the position of the road marking. You may decide to do it.

  In the above-described embodiment, an example in which a camera image is acquired from one camera 40 has been described. However, camera images may be acquired from a plurality of cameras. When a plurality of cameras are used, a road sign is detected from each of the plurality of camera images. When the detected road signs overlap, a process of integrating the detection results is performed.

  As described above, according to the present invention, it is possible to improve the accuracy of route guidance, and it is possible to perform route guidance without delay even when data in the road marking database is missing. And is useful as a navigation device mounted on a vehicle or the like.

It is a figure which shows the structure of the navigation apparatus of this Embodiment. (A) It is a figure which shows the camera image and window area | region containing a pedestrian crossing. (B) It is a figure which shows the brightness change of a pedestrian crossing. (C) It is a figure which shows the frequency distribution of a pedestrian crossing. (A) It is a figure which shows the example of the data memorize | stored in the temporary memory part. (B) It is a figure which shows the example of the road marking corresponding to the data shown to Fig.3 (a). It is a figure which shows operation | movement of a navigation apparatus. (A) It is a figure which shows a camera image and a window area | region. (B) It is a figure which shows the image cut out in the window area | region. (C) It is a figure which shows the image after distortion correction. It is a figure which shows operation | movement of an update part. It is a figure which shows the structure of a road marking database production | generation apparatus.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Navigation apparatus 12 Vehicle position calculation part 14 Road marking detection part 16 Road marking template DB
18 Relative coordinate calculation unit 20 Detection data update unit 22 Temporary storage unit 24 Verification unit 26 Road marking DB
28 Search range determination unit 30 Guidance content determination unit 32 Route information output unit 34 Unmatched road marking DB
36 road marking DB update unit 38 communication unit 40 camera 42 GPS
44 Gyroscope 46 Vehicle speed pulse sensor 48 Vehicle 50 Server 52 Road marking DB
60 Road Marking Database Generation Device 62 Search Area Limiting Unit 64 Map DB
66 Road Sign Detection Unit 68 Road Sign Template DB
70 Integration unit 72 Road image DB

Claims (15)

A vehicle position calculation unit for calculating the current position of the vehicle;
A road marking database storing data indicating the position where the road marking exists and the type of the road marking;
A road marking detection unit that acquires an image of a road photographed by a camera mounted on a vehicle and detects a road marking from the acquired image;
The road marking within the predetermined search range including the current position of the vehicle obtained by the vehicle position calculation unit is read from the road marking database, and the road marking detected by the road marking detection unit in the read road marking A collation unit that identifies the position of the detected road marking by collating whether there is a road marking of the same type as,
When the position of the road marking is specified by the verification in the verification section, route guidance is performed based on the position of the road marking, and when the road marking is not specified, the vehicle position calculation section A route guidance unit for performing route guidance based on the current position of the vehicle calculated in
A navigation device comprising:   The said collation part collates whether the road marking read from the said road marking database has several road markings which have the same arrangement | positioning as the several road marking detected by the said road marking detection part. Navigation device. A road marking update unit that writes the type of road marking detected by the road marking detection unit and the relative coordinates of the road marking with respect to the vehicle into the storage unit, and updates the data of the relative coordinates as the vehicle travels. Prepared,
The navigation device according to claim 1 or 2, wherein the collation unit collates a road marking stored in the storage unit with a road marking read from the road marking database.   The road marking update unit deletes the data of the road marking from the storage unit when the position of the road marking stored in the storage unit is more than a predetermined threshold from the current position of the vehicle. Navigation device. In the road marking database, data on the direction of the road with the road marking is stored in association with the road marking,
The verification unit reads from the road marking database a road marking associated with a road direction that matches the traveling direction of the vehicle when the road marking detection unit detects the road marking, and reads the road marking and the road The navigation apparatus according to any one of claims 1 to 4, wherein collation with a road sign detected by a sign detection unit is performed. The vehicle position calculation unit is connected to a GPS for obtaining an absolute position coordinate of the vehicle and a sensor for outputting a signal used for calculation of a movement locus of the vehicle, and the sensor is added to the absolute position coordinate of the vehicle obtained by the GPS. The current position of the vehicle is calculated by adding the movement locus obtained from the output signal of
The navigation device according to claim 1, wherein the search range used by the collation unit is changed according to an elapsed time after obtaining absolute position coordinates by the GPS.   The road marking detection unit cuts out a window area for road marking detection from an image acquired by the camera, obtains a horizontal frequency distribution of the image of the window area, and the intensity of a specific frequency component in the frequency distribution is The navigation apparatus according to any one of claims 1 to 6, wherein a pedestrian crossing is detected as the road marking when the predetermined threshold value is exceeded. A reference image storage unit that stores reference images of various road markings is provided.
The road sign detection unit cuts out a window area for road sign detection from an image acquired by the camera, and detects the road sign by matching the image of the window area with a reference image stored in the reference image storage unit. The navigation device according to any one of claims 1 to 7. The vehicle position calculation unit
Connected to a GPS for obtaining the absolute position coordinates of the vehicle and a sensor for outputting a signal used for calculating the movement locus of the vehicle,
By calculating the current position of the vehicle by adding the movement locus obtained from the output signal of the sensor to the absolute position coordinates of the vehicle obtained by the GPS,
When the position of the road marking is specified by collation in the collation unit, the current position of the vehicle is recalculated based on the position of the road marking, and until the absolute position coordinates of the vehicle are obtained by the GPS, The navigation apparatus according to claim 1, wherein the vehicle position is detected by adding a movement locus to the recalculated position. An image acquisition unit for acquiring an aerial image or satellite image having a known correspondence with the absolute coordinates on the ground, or an out-of-vehicle camera image to which the location information of the shooting location is added;
A road marking database update unit that detects road markings from the image acquired by the image acquisition unit, and stores data indicating the type and position of the detected road markings in the road marking database;
A navigation device according to any one of claims 1 to 9.   The navigation apparatus according to claim 10, further comprising: a search area limiting unit that limits a region for searching for a road marking to an area where a road is displayed among images acquired by the image acquisition unit using road map information. . An unmatched road marking database that stores data indicating the type and position of the road marking detected by the road marking detection unit when it is determined by the matching in the matching unit that the same type of road marking does not exist;
Based on the data stored in the unverified road marking database, when the number of times the same type of road marking is detected at the same position exceeds a predetermined threshold, the road marking data is stored in the road marking database. The navigation device according to any one of claims 1 to 11, which is added. A vehicle position calculation unit connected to a GPS for obtaining the absolute position coordinates of the vehicle and a sensor that outputs a signal used to calculate the movement locus of the vehicle;
A road marking database storing data indicating the position where the road marking exists and the type of the road marking;
A road marking detection unit that acquires an image of a road photographed by a camera mounted on a vehicle and detects a road marking from the acquired image;
The road marking within the predetermined search range including the current position of the vehicle obtained by the vehicle position calculation unit is read from the road marking database, and the road marking detected by the road marking detection unit in the read road marking A matching unit that identifies the detected road marking by collating whether there is a road marking of the same type as,
With
The vehicle position calculation unit
By calculating a current position of the vehicle by adding a movement locus calculated from the output signal of the sensor to the absolute position coordinates of the vehicle obtained by the GPS,
When the position of the road marking is specified by collation in the collation unit, the current position of the vehicle is recalculated based on the position of the road marking, and until the absolute position coordinates of the vehicle are obtained by the GPS, A navigation device that detects a vehicle position by adding a movement locus to a recalculated position. Preparing a road marking database storing data indicating a position where the road marking exists and a type of the road marking;
Obtaining a road image taken with a camera mounted on the vehicle;
Detecting road markings from the acquired images;
Detecting the current position of the vehicle based on the absolute position coordinates obtained by the GPS and the movement locus of the vehicle;
By reading a road marking within a predetermined search range including the current position of the vehicle from the road marking database, and checking whether the road marking of the same type as the road marking detected from the image is in the read road marking Identifying the location of the detected road marking;
When the position of the road marking is specified by collation, route guidance is performed based on the position of the road marking, and when the position of the road marking is not specified, the route is determined based on the current position of the vehicle. A step of providing guidance;
A navigation method comprising: A program that provides route guidance to a computer
Obtaining a road image taken with a camera mounted on the vehicle;
Detecting road markings from the acquired images;
Calculating the current position of the vehicle based on the absolute position coordinates obtained by GPS and the movement trajectory of the vehicle;
A road marking within a predetermined search range including the current position of the vehicle is read from a road marking database storing data indicating a position where the road marking exists and the type of the road marking, and the road is included in the read road marking. Identifying the position of the detected road marking by collating whether there is a road marking of the same type as the road marking detected by the sign detection unit;
When the position of the road marking is specified by the collation, route guidance is performed based on the position of the road marking, and when the position of the road marking is not specified, the calculated current position of the vehicle is set. Based on route guidance,
A program that executes

Images