JP5742559B2 - POSITION DETERMINING DEVICE, NAVIGATION DEVICE, POSITION DETERMINING METHOD, AND PROGRAM - Google Patents

Description

  The present invention relates to a position determination device, a navigation device, a position determination method, and a program, and more particularly, to a position determination device, a navigation device, and a position determination that determine a position of a vehicle on a map using a landscape image taken from an in-vehicle camera. It relates to methods and programs.

  Conventionally, as this type of position determination device, pattern information indicating road markings such as white lines on the road and position information indicating the coordinates of the feature points of the road markings (world coordinates) are registered in advance, When it is determined by pattern matching using pattern information that there is a road marking in the image captured by the in-vehicle camera, feature points are extracted from the captured image, and the vehicle coordinate system is based on the provisional current position of the vehicle. It has been proposed to calculate the coordinates of the feature points of the road marking (vehicle coordinates), and to calculate the current position of the vehicle using the calculated vehicle coordinates of the feature points of the road marking and the corresponding position information (for example, Patent Document 1). With this device, it is possible to calculate the current position with higher accuracy than the provisional current position of the vehicle obtained by the position measurement using the GPS signal, the gyroscope, and the signal from the lane sensor.

JP 2007-108043 A

  However, the characteristic part of the landscape image from the vehicle, which is performed for the purpose of matching the pattern information indicating the road marking performed by the above-described device with the photographed image and calculating the current position with high accuracy as in the above-described device. Since it is necessary to perform the matching process between the pattern information indicating the captured image and the captured image while the vehicle is traveling, the vehicle position may not be correctly determined. While the vehicle is traveling, depending on the traveling environment, the speed of movement of the road markings and characteristic parts in the captured image increases, and the time required to obtain a captured image that matches the pattern information decreases. That is, when the range of the shooting position on the road where a captured image that matches the pattern information can be obtained becomes narrow, the chance that it is determined that the pattern information and the captured image match is reduced. was there.

  The main object of the position determination apparatus, navigation apparatus, position determination method, and program of the present invention is to more appropriately determine the position of the vehicle on the map.

  The position determination apparatus, navigation apparatus, position determination method, and program of the present invention employ the following means in order to achieve the main object described above.

The first position determination device of the present invention is
A plurality of reference data obtained by extracting characteristic parts of a landscape image taken from an in-vehicle camera at intervals in the traveling direction on the road are stored in association with the corresponding positions on the map, and taken from the in-vehicle camera. Matching processing is performed to determine whether or not the scenery image to be matched and the plurality of reference data match, and based on the position on the map corresponding to the reference data determined to match by the matching processing A position determination device for determining a position of a vehicle on a map,
The matching process is performed when the degree of coincidence between the reference data and the determination data obtained by performing predetermined image processing for extracting a characteristic part from a landscape image photographed from a vehicle is equal to or greater than a determination threshold value. It is performed by determining that they match and determining that they do not match when the degree of matching is less than the threshold for determination,
The determination threshold is adjusted so as to decrease as the vehicle speed increases and to decrease as the steering wheel steering amount increases.
It is characterized by that.

  In the first position determination device of the present invention, the matching process is performed by matching the reference data with the determination data obtained by performing predetermined image processing for extracting a characteristic part from the landscape image captured from the vehicle. It is determined by matching when the degree of is equal to or greater than the determination threshold, and is determined not to match when the degree of matching is less than the determination threshold. It is adjusted to tend to be smaller and smaller as the steering wheel steering amount is larger. When the vehicle speed is relatively high or the steering amount is relatively large, the moving speed (steering wheel steering) of a characteristic part in the landscape image (for example, a roadside feature such as a roadside building or a road sign) When the amount is relatively large, the moving speed in a direction corresponding to the direction opposite to the turning direction of the vehicle in particular is relatively large, and the time during which a landscape image having a characteristic portion matching the reference data can be taken is short. . Therefore, in such a case, by reducing the determination threshold value, it can be easily determined that the scenery image and the reference data match by the matching process. As a result, the position of the vehicle on the map can be determined more appropriately. Here, as the predetermined image processing, one including at least edge detection processing can be used.

  In the first position determination apparatus of the present invention, the vehicle includes a width direction vehicle position detection unit that detects a vehicle position in the width direction of the road on which the vehicle is traveling, and the threshold value for determination is determined based on whether the detected vehicle position is the travel distance. It is also possible to adjust the tendency so that the closer to the road side of the road, the smaller the trend is. This is because the closer the vehicle position is to the side of the road in the width direction of the running road, the higher the moving speed of the characteristic part on the side of the road side (left side or right side) closer in the landscape image. based on. In this case, the width direction vehicle position detecting means is means for detecting a traveling lane when traveling on a road having a plurality of lanes, and the threshold for determination is determined based on the detected traveling lane. It is also possible to adjust the tendency so that the closer to the roadside of the road, the smaller the tendency is.

  In the first position determination apparatus of the present invention, the vehicle further includes a width direction vehicle position detection unit that detects a vehicle position in the width direction of the traveling road, and the determination threshold value is the width of the traveling road. The detected vehicle position in the direction may be adjusted so as to become smaller as the distance from the vehicle position at the time of capturing the landscape image from which the reference data is obtained is adjusted. This is because the farther away from the vehicle position at the time of shooting the landscape image in which the vehicle position obtained the reference data in the width direction of the running road, the characteristic side on the far side (left side or right side) in the landscape image It is based on the tendency that the moving speed of a part becomes large. In this case, the width direction vehicle position detecting means is means for detecting a traveling lane when traveling on a road having a plurality of lanes, and the determination threshold is calculated based on whether the detected traveling lane is the reference data. It is also possible to adjust the tendency so that the distance from the traveling lane at the time of shooting the landscape image obtained becomes smaller.

  Furthermore, in the first position determination apparatus of the present invention, the determination threshold tends to be smaller as the width of the traveling road is smaller, and tends to be smaller as the curvature radius of the traveling road is smaller. It is also possible to adjust the distance to be at least one of the tendency to become smaller as the distance from the nearest intersection on the road is shorter. This is because the smaller the width of the running road, the smaller the radius of curvature of the running road, the shorter the distance from the intersection on the road that entered after turning the intersection, the shorter the landscape image It is based on the tendency that the moving speed of the characteristic part inside becomes large.

  Alternatively, in the first position determination device of the present invention, the threshold for determination tends to decrease as the number of lanes on a traveling road decreases, and tends to decrease as the lane width of a traveling road decreases. It can also be adjusted so that it becomes at least one of the tendency which becomes small, so that the sidewalk width of the road which is carried out is narrow. This is because the smaller the number of road lanes, the narrower the lane width of the road, the narrower the sidewalk width of the road, the smaller the width of the road, and the smaller the width of the running road, This is based on the tendency that the moving speed of the characteristic part in the image increases.

  In the first position determination apparatus of the present invention in which the determination threshold is adjusted such that the vehicle position tends to decrease as the vehicle position moves away from the vehicle position at the time of capturing a landscape image from which reference data is obtained. The data is divided into left side determination data and right side determination data on the left side when viewed from the vehicle, and the reference data is divided into left side reference data and left side reference data on the left side when viewed from the vehicle, The degree of matching between the left determination data and the left reference data is equal to or greater than the left determination threshold, and the degree of coincidence between the right determination data and the right reference data is equal to or greater than the right determination threshold. And the degree of matching between the left determination data and the left reference data is less than the left determination threshold, or the right determination data and the left reference data When the degree of coincidence with the side reference data is less than the right-side determination threshold, it is determined that they do not coincide with each other, and the left-side determination threshold is detected in the width direction of the running road. The vehicle position is adjusted so as to decrease as it moves further leftward from the vehicle position at the time of capturing the landscape image from which the reference data was obtained, and the right-side determination threshold is set in the width direction of the running road. The detected vehicle position may be adjusted so as to become smaller as it moves to the right side from the vehicle position at the time of capturing the landscape image from which the reference data was obtained. In this way, the determination threshold value can be adjusted more appropriately.

The second position determination apparatus of the present invention is
A plurality of reference data obtained by extracting characteristic parts of a landscape image taken from an in-vehicle camera at intervals in the traveling direction on the road are stored in association with the corresponding positions on the map, and taken from the in-vehicle camera. Matching processing is performed to determine whether or not the scenery image to be matched and the plurality of reference data match, and based on the position on the map corresponding to the reference data determined to match by the matching processing A position determination device for determining a position of a vehicle on a map,
A width direction vehicle position detecting means for detecting a vehicle position in the width direction of the running road,
The matching process is performed when the degree of coincidence between the reference data and the determination data obtained by performing predetermined image processing for extracting a characteristic part from a landscape image photographed from a vehicle is equal to or greater than a determination threshold value. It is performed by determining that they match and determining that they do not match when the degree of matching is less than the threshold for determination,
The determination threshold is adjusted so that the detected vehicle position tends to decrease as the detected vehicle position is closer to the side of the road on which the vehicle is running.
It is characterized by that.

  In the second position determination apparatus of the present invention, the matching process is performed by matching the determination data obtained by performing predetermined image processing for extracting a characteristic portion on the landscape image captured from the vehicle and the reference data. It is determined that it is coincident when the degree is equal to or greater than the determination threshold, and is determined not to coincide when the degree of coincidence is less than the determination threshold, and the determination threshold is running The vehicle position in the width direction of the road is adjusted to tend to be smaller as it is closer to the side of the road on which the vehicle is traveling. If the vehicle position in the width direction of the running road is relatively close to the roadside of the road, a characteristic part (for example, the roadside The moving speed of roadside features such as buildings and road signs is relatively high, and the time during which a landscape image having a characteristic portion matching the reference data can be taken is short. Therefore, in such a case, by reducing the determination threshold value, it can be easily determined that the scenery image and the reference data match by the matching process. As a result, the position of the vehicle on the map can be determined more appropriately. Here, as the predetermined image processing, one including at least edge detection processing can be used.

The third position determination apparatus of the present invention is
A plurality of reference data obtained by extracting characteristic parts of a landscape image taken from an in-vehicle camera at intervals in the traveling direction on the road are stored in association with the corresponding positions on the map, and taken from the in-vehicle camera. Matching processing is performed to determine whether or not the scenery image to be matched and the plurality of reference data match, and based on the position on the map corresponding to the reference data determined to match by the matching processing A position determination device for determining a position of a vehicle on a map,
A width direction vehicle position detecting means for detecting a vehicle position in the width direction of the running road,
The matching process is performed when the degree of coincidence between the reference data and the determination data obtained by performing predetermined image processing for extracting a characteristic part from a landscape image photographed from a vehicle is equal to or greater than a determination threshold value. It is performed by determining that they match and determining that they do not match when the degree of matching is less than the threshold for determination,
The threshold value for determination is adjusted so that the detected vehicle position in the width direction of the running road becomes smaller as the distance from the vehicle position at the time of capturing a landscape image obtained from the reference data decreases.
It is characterized by that.

  In the third position determination apparatus according to the present invention, the matching process is performed between the determination data obtained by performing predetermined image processing for extracting a characteristic portion from a landscape image captured from a vehicle and the reference data. When the degree of matching is greater than or equal to the determination threshold value, it is determined that they match, and when the degree of matching is less than the determination threshold value, it is determined that they do not match. In the width direction of the road, the vehicle position is adjusted so as to become smaller as the distance from the vehicle position at the time of taking a landscape image from which reference data is obtained. If the vehicle position is relatively far from the vehicle position at the time of shooting the landscape image from which reference data was obtained in the width direction of the road that is running, the direction side (left side or right side) in the landscape image ) Characteristic parts (for example, roadside buildings, roadside features such as road signs, etc.) are relatively fast, and the time during which a landscape image having a characteristic part matching the reference data can be taken is short. . Therefore, in such a case, by reducing the determination threshold value, it can be easily determined that the scenery image and the reference data match by the matching process. As a result, the position of the vehicle on the map can be determined more appropriately. Here, as the predetermined image processing, one including at least edge detection processing can be used.

  The navigation device of the present invention extracts the characteristic portion of the landscape image taken from the in-vehicle camera at intervals in the traveling direction on the road, that is, the position determination device of the present invention of any of the above-described aspects. A plurality of reference data obtained in association with the corresponding positions on the map, and matching for determining whether or not the landscape image photographed from the in-vehicle camera matches the plurality of reference data A position determination device that performs a process and determines a position on a map of a vehicle based on a position on a map corresponding to reference data determined to be matched by the matching process. The degree of coincidence between the determination data obtained by performing predetermined image processing for extracting a characteristic portion from a landscape image captured from the reference data and the reference data is less than the determination threshold. It is determined by determining that they match when the degree of matching is less than the determination threshold, and the determination threshold tends to decrease as the vehicle speed increases. And the first position determination device or the width direction vehicle position detection means for detecting the vehicle position in the width direction of the running road, the adjustment being made such that the steering wheel steering amount increases. The processing is coincident when the degree of coincidence between the determination data obtained by performing predetermined image processing for extracting a characteristic portion on the landscape image taken from the vehicle and the reference data is equal to or greater than the determination threshold. And when the degree of coincidence is less than the determination threshold value, it is determined that they do not match. The determination threshold value is determined based on the detected vehicle position. A second position determination device, or a width direction vehicle position detection means for detecting a vehicle position in the width direction of the road on which the vehicle is running The matching process is performed when the degree of coincidence between the reference data and the determination data obtained by performing predetermined image processing for extracting a characteristic part from a landscape image taken from a vehicle is equal to or greater than a determination threshold value. And the determination threshold is determined in the width direction of the running road. A third position determination device that is adjusted such that the detected vehicle position tends to decrease as the vehicle position moves away from the vehicle position at the time of capturing the landscape image from which the reference data was obtained. The gist is that route guidance to the destination is performed using the determined position of the vehicle on the map.

  Since the navigation device of the present invention includes the position determination device of the present invention according to any one of the aspects described above, it is possible to more appropriately determine the effect of the position determination device described above, for example, the position of the vehicle on the map. The same effects as those that can be achieved can be achieved.

The first position determination method of the present invention includes:
A plurality of reference data obtained by extracting characteristic portions of a landscape image taken from an in-vehicle camera on the road at intervals in the running direction are stored in association with the corresponding positions on the map, and taken from the vehicle. A matching process for determining whether or not the landscape image matches the plurality of reference data, and a position on the map corresponding to the reference data determined to be constant by the matching process Determining the position of the vehicle on the map, and a position determination method comprising:
The matching process is performed when the degree of coincidence between the reference data and the determination data obtained by performing predetermined image processing for extracting a characteristic part from a landscape image photographed from a vehicle is equal to or greater than a determination threshold value. It is performed by determining that they match and determining that they do not match when the degree of matching is less than the threshold for determination,
The determination threshold is adjusted so as to decrease as the vehicle speed increases and to decrease as the steering wheel steering amount increases.
It is characterized by that.

  In the first position determination method of the present invention, the matching process matches the reference data with the determination data obtained by performing predetermined image processing for extracting a characteristic part from the landscape image captured from the vehicle. It is determined by matching when the degree of is equal to or greater than the determination threshold, and is determined not to match when the degree of matching is less than the determination threshold. It is adjusted to tend to be smaller and smaller as the steering wheel steering amount is larger. When the vehicle speed is relatively high or the steering amount is relatively large, the moving speed (steering wheel steering) of a characteristic part in the landscape image (for example, a roadside feature such as a roadside building or a road sign) When the amount is relatively large, the moving speed in a direction corresponding to the direction opposite to the turning direction of the vehicle in particular is relatively large, and the time during which a landscape image having a characteristic portion matching the reference data can be taken is short. . Therefore, in such a case, by reducing the determination threshold value, it is possible to increase the chance of determining that the landscape image and the reference data match by the matching process. As a result, the position of the vehicle on the map can be determined more appropriately. Here, as the predetermined image processing, one including at least edge detection processing can be used.

The second position determination method of the present invention includes:
A plurality of reference data obtained by extracting characteristic portions of a landscape image taken from an in-vehicle camera on the road at intervals in the running direction are stored in association with the corresponding positions on the map, and taken from the vehicle. A matching process for determining whether or not the landscape image matches the plurality of reference data, and a position on the map corresponding to the reference data determined to be constant by the matching process Determining the position of the vehicle on the map, and a position determination method comprising:
The matching process is performed when the degree of coincidence between the reference data and the determination data obtained by performing predetermined image processing for extracting a characteristic part from a landscape image photographed from a vehicle is equal to or greater than a determination threshold value. It is performed by determining that they match and determining that they do not match when the degree of matching is less than the threshold for determination,
The determination threshold value is adjusted so that the vehicle position in the width direction of the traveling road is closer to the side of the road on which the traveling road is closer to be smaller.
It is characterized by that.

  In the second position determination method of the present invention, the matching process is performed by matching the determination data and the reference data obtained by performing predetermined image processing for extracting a characteristic part on the landscape image captured from the vehicle. It is determined that it is coincident when the degree is equal to or greater than the determination threshold, and is determined not to coincide when the degree of coincidence is less than the determination threshold, and the determination threshold is running The vehicle position in the width direction of the road is adjusted to tend to be smaller as it is closer to the side of the road on which the vehicle is traveling. If the vehicle position in the width direction of the running road is relatively close to the roadside of the road, a characteristic part (for example, the roadside The moving speed of roadside features such as buildings and road signs is relatively high, and the time during which a landscape image having a characteristic portion matching the reference data can be taken is short. Therefore, in such a case, by reducing the determination threshold value, it can be easily determined that the scenery image and the reference data match by the matching process. As a result, the position of the vehicle on the map can be determined more appropriately. Here, as the predetermined image processing, one including at least edge detection processing can be used.

The third position determination method of the present invention includes:
A plurality of reference data obtained by extracting characteristic portions of a landscape image taken from an in-vehicle camera on the road at intervals in the running direction are stored in association with the corresponding positions on the map, and taken from the vehicle. A matching process for determining whether or not the landscape image matches the plurality of reference data, and a position on the map corresponding to the reference data determined to be constant by the matching process Determining the position of the vehicle on the map, and a position determination method comprising:
The matching process is performed when the degree of coincidence between the reference data and the determination data obtained by performing predetermined image processing for extracting a characteristic part from a landscape image photographed from a vehicle is equal to or greater than a determination threshold value. It is performed by determining that they match and determining that they do not match when the degree of matching is less than the threshold for determination,
The determination threshold value is adjusted so that the vehicle position in the width direction of the running road becomes smaller as the vehicle position moves away from the vehicle position at the time of capturing a landscape image obtained from the reference data.
It is characterized by that.

  In the third position determination method of the present invention, the matching process is performed between the determination data obtained by performing predetermined image processing for extracting a characteristic portion from a landscape image captured from a vehicle and the reference data. When the degree of matching is greater than or equal to the determination threshold value, it is determined that they match, and when the degree of matching is less than the determination threshold value, it is determined that they do not match. In the width direction of the road, the vehicle position is adjusted so as to become smaller as the distance from the vehicle position at the time of taking a landscape image from which reference data is obtained. If the vehicle position is relatively far from the vehicle position at the time of shooting the landscape image from which reference data was obtained in the width direction of the road that is running, the direction side (left side or right side) in the landscape image ) Characteristic parts (for example, roadside buildings, roadside features such as road signs, etc.) are relatively fast, and the time during which a landscape image having a characteristic part matching the reference data can be taken is short. . Therefore, in such a case, by reducing the determination threshold value, it can be easily determined that the scenery image and the reference data match by the matching process. As a result, the position of the vehicle on the map can be determined more appropriately. Here, as the predetermined image processing, one including at least edge detection processing can be used.

  The program of the present invention is for causing one or more computers to implement the steps of the first to third position determination methods described above. This program may be recorded on a computer-readable recording medium (for example, hard disk, ROM, FD, CD, DVD, etc.), or from a computer via a transmission medium (communication network such as the Internet or LAN). It may be distributed to another computer, or may be exchanged in any other form. If this program is executed by one computer or shared by a plurality of computers, each step of the above-described position determination method is realized, so that the same effect as that of the position determination method can be obtained.

1 is a configuration diagram showing an outline of a configuration of a navigation system 10 including a navigation device 20 as an embodiment of the present invention. It is explanatory drawing explaining an example of a mode that the reference data used for the matching process with a landscape image are produced. It is explanatory drawing explaining an example of several relationship with a road, its width | variety Dw, sidewalk width Pw, and lane width Lw. 3 is a flowchart showing an example of a vehicle position determination processing routine executed by an electronic control unit 30. It is explanatory drawing which shows an example of the distance Dc from the nearest intersection in a road. It is explanatory drawing explaining an example of the several relationship of the parameter and coefficient for setting the threshold value for determination Mref. It is explanatory drawing which shows the example of the coefficient setting map of the travel lane coefficient k3. It is explanatory drawing explaining the example of a mode that the driving lane difference coefficient k4 is set. It is explanatory drawing explaining an example of the some relationship between the parameter and coefficient for setting the threshold value for determination Mref. It is explanatory drawing explaining an example of a mode that a vehicle turns left at an intersection and drive | works. It is a flowchart which shows an example of the own vehicle position determination processing routine performed by the electronic control unit 30 of a modification. It is explanatory drawing which shows an example of the left travel lane difference coefficient setting map. It is explanatory drawing which shows an example of the map for right lane difference coefficient setting. It is explanatory drawing which shows an example of a mode that reference data is divided | segmented into left reference data and right reference data.

  Next, the form for implementing this invention is demonstrated using an Example.

  FIG. 1 is a configuration diagram showing an outline of a configuration of a navigation system 10 including a navigation device 20 as an embodiment of the present invention. The navigation system 10 according to the embodiment includes an in-vehicle device 12 centering on various sensors mounted on a vehicle (automobile) (not shown) and a navigation device 20 that operates by receiving power supply from an in-vehicle battery (not shown). As shown in the drawing, the navigation device 20 of the embodiment has a rectangular screen for displaying characters and images, for example, a display 22 configured as a liquid crystal display, an organic electroluminescence display, and the like, and is attached to the screen of the display 22. For example, a touch panel 24 using a resistance film method or a capacitance method, an electronic control unit 30 for controlling the entire apparatus, and a hard disk drive (hereinafter referred to as HDD) 40 as a large-capacity memory for storing various application software and map data. With.

  The electronic control unit 30 is configured as a microprocessor centered on the CPU 32. In addition to the CPU 32, the ROM 34 that stores various processing programs, the RAM 36 that temporarily stores data, and the nonvolatile that holds the stored data. Flash memory 38 and an input / output port (not shown). In addition to the signal from the touch panel 24 for detecting the touch position of the operator and the data read from the HDD 40, various signals from the in-vehicle device 12 are input to the electronic control unit 30 via the in-vehicle network via the input port. Yes. The signal from the in-vehicle device 12 includes a signal from a GPS receiver 50 that receives a signal from a GPS (Global Positioning System) satellite via a GPS antenna, and a pulse signal every time a vehicle drive shaft or wheel rotates a certain amount. A signal from the vehicle speed sensor 52 that outputs a signal, a signal from the acceleration sensor 54 that detects the acceleration of the vehicle, a signal from the direction sensor 56 that includes, for example, a gyro sensor that detects the traveling direction of the vehicle and its change, a vehicle handle A signal from a steering amount sensor 58 that detects a rotation amount (rotation angle) from a reference position (a position corresponding to the traveling direction) as a steering wheel steering amount, a moving image or a still image of a landscape image installed in front of the vehicle There is a signal (captured image) from the camera 59 that can be photographed. From the electronic control unit 30, a display signal to the display 22, an audio signal to the speaker 26 with a built-in amplifier, data to be written to the HDD 40, and the like are output via an output port.

  In the navigation device 20 of the embodiment, the electronic control unit 30 reads out necessary application software and map data from the HDD 40 and executes various processes. For example, as shown in the functional block of FIG. 1 representing functions realized by hardware, software, or a combination thereof, the electronic control unit 30 is based on a signal from the GPS receiver 50, a signal from the direction sensor 56, or the like. Location processing by the location processing unit 60 that determines the vehicle position that is the current position of the vehicle, map display processing by the map display processing unit 62 that displays a map on the display 22 using map data, and the purpose from the determined vehicle position Searching for a travel route to the ground, displaying a map by the map display processing unit 62, navigation processing by the navigation processing unit 64 that provides route guidance by voice output from the speaker 26, and various images by the image processing unit 66 used for these processings Execute processing. The map data is stored in a map data database (hereinafter referred to as map data DB) 42 of the HDD 40.

  In the location processing by the electronic control unit 30, the vehicle position is calculated by so-called GPS navigation based on the GPS signal from the GPS receiver 50, or the vehicle speed pulse signal or acceleration from the vehicle speed sensor 52 when the GPS signal cannot be received. Map data when the vehicle position is calculated by so-called autonomous navigation based on the acceleration of the vehicle from the sensor 54 and the vehicle direction from the direction sensor 56, or when the calculated vehicle position deviates from the road on the map. The vehicle position is estimated by performing so-called map matching that corrects the vehicle position to a position on the road using. Further, in the location processing, matching processing is performed as image recognition processing for determining whether or not a landscape image photographed by the camera 59 matches a plurality of reference data stored in the HDD 40 in advance. The position on the map associated with the reference data determined to be correct is determined as the correct vehicle position instead of the vehicle position estimated by GPS navigation, autonomous navigation, or the like. Here, the plurality of reference data stored in the HDD 40 are feature points of a landscape image taken at intervals in the traveling direction on the road from an in-vehicle camera of the vehicle that actually traveled (for example, the outline of a roadside building, a window Data obtained by extracting roadside features such as frames, signboards and road signs, road markings, etc.), together with shooting information such as the position of the vehicle when shooting a landscape image, and a reference data database of the HDD 40 ( (Hereinafter referred to as reference data DB) 44.

  Here, a method for creating reference data used for matching processing with a landscape image photographed by the camera 59 will be described. FIG. 2 is an explanatory diagram for explaining an example of creating reference data used for matching processing with a landscape image. In the embodiment, as shown in the upper part of the figure, the vehicle for creating reference data travels while taking a landscape image in front of the vehicle from the in-vehicle camera with a predetermined reference data interval D on the road. The position (latitude and longitude) on the map of the vehicle at the time of shooting and the direction of the vehicle (direction along the road direction) are acquired as accurate information based on signals from various sensors. In the embodiment, the reference data interval D can be obtained on many roads in an accurate range within a range in which the calculation amount of the electronic control unit 30 included in the navigation device 20 of the embodiment is not excessive by location processing. As the distance on the road, a distance (for example, several tens of meters) determined in advance by experiments and analysis based on the performance of the electronic control unit 30 and the HDD 40 is used. Subsequently, as shown in the lower left part of the figure, the feature point of the landscape image is extracted and binarized by performing an edge detection process etc. on the photographed landscape image by the on-board electronic control unit A point image is created, and the feature point image is subjected to enhancement processing that emphasizes feature points suitable for landscape image recognition and discards feature points around the emphasized feature points, noise removal processing, etc. Create reference data. In addition, when creating reference data from a photographed landscape image, the feature point image is emphasized so that road markings such as white lines and arrows on the road are emphasized. Detects the driving lane. Detailed descriptions of well-known image processing methods such as edge detection processing, enhancement processing, and noise removal processing are omitted. The plurality of reference data created in this manner are stored in the reference data DB 44 of the HDD 40 of the navigation device 20 mounted on the vehicle of the embodiment, as shown in the right side of the middle stage in the drawing, for each reference data, And the driving lane (hereinafter referred to as the shooting position, shooting direction, and driving lane at the time of shooting), respectively.

  The map data DB 42 of the HDD 40 includes road information, facility information, and the like. The road information includes, for example, the position (latitude and longitude) on the map indicated by the nodes at both ends of each road, the distance of each road, and the region. (Urban area, suburb), type (general road, highway), slope, number of traffic lights, as shown on the lower right side of FIG. 2, the width Dw, the number of lanes Ln, the lane width Lw, the sidewalk width Pw, road shape (straight line or curve), radius of curvature R, road type (one-way road, two-way road, road with a median, etc.), the aforementioned reference data interval D, and the like are included. FIG. 3 is an explanatory diagram illustrating an example of a plurality of relationships between a road and its width Dw, sidewalk width Pw, and lane width Lw. In the figure, (a) on the left shows a road with a road type of two-way traffic and a lane number Ln of 6 on one side, and the center (b) shows a one-way road (for example, an expressway or an elevated road). And the right side (c) shows a road with a road type of one-way and a lane number Ln of 4. In FIG. 3, the sidewalk width Pw of the road is equal on both the left and right sides, and the lane widths Lw are all equal. However, if all roads are based on the same standard, for example, the sidewalk width Pw is It may be expressed as an average value of widths, or the lane width Lw may be expressed as an average value of lane widths of all lanes.

  Next, the operation of the navigation device 20 of the embodiment configured as described above, particularly the operation when determining the position of the vehicle during traveling will be described. FIG. 4 is a flowchart illustrating an example of a vehicle position determination processing routine executed by the location processing unit 60 of the electronic control unit 30 in cooperation with the image processing unit 66. In this routine, the current vehicle traveling direction (direction along the road direction) based on the vehicle position estimated by GPS navigation, autonomous navigation, or the like and the signal from the direction sensor 56 is referred to the reference data DB 44, respectively. It is repeatedly executed every time the shooting position and the shooting direction associated with one of the data match.

  When the own vehicle position determination processing routine is executed, the CPU 32 of the electronic control unit 30 first takes an image of the estimated own vehicle position, which is the currently estimated own vehicle position, by the camera 59 when this routine is executed. Data for determination obtained from a landscape image, estimated vehicle position and reference data of shooting position and shooting direction that coincide with the current vehicle traveling direction (hereinafter referred to as vehicle traveling direction), estimated vehicle position and vehicle traveling direction Corresponding to the shooting travel lane Ls corresponding to the coincident shooting position and shooting direction, the number of road lanes Ln, the lane width Lw, the sidewalk width Pw, the curvature radius R, the estimated own vehicle position and the vehicle traveling direction at the estimated own vehicle position. The distance Dc from the nearest intersection on the road corresponding to the road shape and type of the road, the estimated vehicle position and the vehicle traveling direction, and the signal from the vehicle speed sensor 52 A process of inputting data necessary for determination, such as the vehicle speed V calculated based on the above, the steering wheel amount St obtained as a signal from the steering amount sensor 58, and the determination reference threshold value M0 used in this determination process, is executed (step S100). ).

  Here, in the embodiment, the estimated vehicle position is input by the location processing unit 60 estimated by GPS navigation, autonomous navigation, or the like. For the determination data, the same image processing method used when the reference data is created from the landscape image described with reference to FIG. 2 is applied to the landscape image captured by the camera 59 when this routine is executed. It was supposed to input what was created. The reference data and the shooting travel lane Ls are input from the reference data DB 44 corresponding to the shooting position and shooting direction corresponding to the estimated own vehicle position and vehicle traveling direction, respectively. The number of road lanes Ln, the lane width Lw, the sidewalk width Pw, and the radius of curvature R are input from the map data DB 42 as information stored on the road at the estimated vehicle position. As the road shape and road type, information stored as road information corresponding to the estimated own vehicle position and vehicle traveling direction is input from the map data DB 42. In the embodiment, the distance Dc from the nearest intersection is determined for each of a plurality of points obtained by dividing the nodes at both ends of each road into unit distances, and is based on the relationship with other roads and the position and distance of each road. What is calculated and stored in the map data DB 42 is input. An example of the distance Dc from the nearest intersection on the road in FIG. 5 is shown. The determination reference threshold value M0 is a process of this routine for determining whether or not the coincidence rate M indicating the degree of coincidence between the determination data and the reference data is greater than or equal to the determination threshold value Mref (that is, the matching process of the embodiment). The reference value of the determination threshold value Mref is determined in advance as a comparatively small value within a range where it can be determined that the determination data matches the reference data.

  When the data is input in this way, the same method as the image processing method when the driving lane is detected when the reference data described with reference to FIG. 2 is generated, and when the determination data is generated from the landscape image photographed by the camera 59 To detect the travel lane Lr on the road on which the vehicle is traveling (step S110), and calculate the travel lane difference ΔL obtained by subtracting the captured travel lane Ls from the detected travel lane Lr. (Step S120). Here, in a country or region where the automobile is on the left side like Japan, the driving lane Lr is the value of the leftmost lane as the value 1 and the value from the left side when the road on which the vehicle is driving has multiple lanes. The first lane is a value indicated by an integer, and the value is 1 when the road on which the vehicle is traveling has only one lane. The shooting lane Ls as shooting information in the reference data DB 44 is also an integer equal to or greater than 1.

  Subsequently, the vehicle speed coefficient k1, the steering amount coefficient k2, and the travel to be multiplied by the determination reference threshold M0 based on the input vehicle speed V, the steering wheel steering amount St, the detected travel lane Lr, and the calculated travel lane difference ΔL, respectively. A lane coefficient k3 and a travel lane difference coefficient k4 are set (steps S130 to S150). Here, in the embodiment, for each coefficient, the relationship between the parameter and the coefficient is determined in advance and stored in the ROM 34 as a coefficient setting map for each coefficient. When a parameter is given, the corresponding coefficient is derived from the stored map. And set it.

  FIG. 6 shows an example of a plurality of relationships between parameters and coefficients for setting the determination threshold value Mref, FIG. 7 shows an example of a coefficient setting map for the travel lane coefficient k3, and FIG. 8 shows the travel lane coefficient k4. The explanatory view explaining the example of a mode of setting up is shown. In FIG. 6, the right side is an example of a coefficient setting map. As shown in FIG. 6, as the vehicle speed V increases, the vehicle speed coefficient k1 decreases across the value 1, and as the steering wheel amount St increases, the steering amount coefficient k2 decreases from the value 1, and the travel lane Lr is closer to the roadside. As the travel lane coefficient k3 decreases and the absolute value of the travel lane difference ΔL increases (the travel lane Lr moves away from the travel lane Ls during shooting), the travel lane difference coefficient k4 decreases from the value 1. Each coefficient is set based on this. As for the steering amount coefficient k2, the value 1 is uniformly set in a range where the steering amount St of the steering wheel is equal to or less than a predetermined amount as a dead zone in the vicinity of the value 0. Further, as shown in FIG. 7, for the driving lane coefficient k3, a different coefficient setting map is selected for each input road type and the number of lanes Ln, and the driving lane Lr is applied to the selected coefficient setting map. Can be set. For example, in the road (a) on the left side of the figure (the road type is a face-to-face road and the number of lanes Ln is a value of 6), the smaller the travel lane Lr (that is, the closer to the left side of the road), the more the value 1 is crossed. The driving lane coefficient k3 is set so as to be small at the center, and the road (b) in the center (the road type is a one-way road and the number of lanes Ln is 3) or the road (c) on the right side of the figure ( The road lane coefficient k3 is set so that the road lane Lr becomes smaller from the value 1 as the road lane Lr is away from the center lane and closer to the left or right side of the road, when the road type is a one-way road and the number of lanes Ln is 4. Is done. Furthermore, by setting the travel lane difference coefficient k4 based on the absolute value of the travel lane difference ΔL, as shown in FIG. 8, the travel lane Lr is separated from the travel lane Ls on the left or right side in the road width direction as shown in FIG. In other words, it can be said that the value is set to be smaller from 1. The reason for setting each coefficient in this way will be described next.

  First, when the vehicle speed V and the steering wheel amount St are relatively large, the moving speed of the feature point in the landscape image taken by the camera 59 (the steering wheel steering amount St is a direction corresponding to a direction opposite to the turning direction of the vehicle in particular. The time during which a landscape image having a feature point matching the reference data can be taken is extremely short. In other words, in such a case, the range of the shooting position on the road where a landscape image having a feature point matching the reference data in the reference data DB 44 can be shot is quite narrow. For this reason, the opportunity to determine that the landscape image and the reference data match by the matching process is reduced, and this is true even though the landscape image may be originally recognized as matching the reference data. A so-called recognition failure that cannot be recognized occurs. On the other hand, as the vehicle speed V and the steering wheel amount St are increased, the vehicle speed coefficient k1 and the steering amount coefficient k2 are decreased, and a determination threshold value Mref, which will be described later, obtained by multiplying the determination reference threshold value M0 by each coefficient is decreased. By this, it can be easily determined that the determination data and the reference data are matched by the matching process (the chances of being determined to be matched can be increased), and are recognized as being originally matched. It is possible to suppress omissions in recognition of landscape images that may be present (increase the recognition rate).

  Further, when the driving lane Lr is close to the road side, the vehicle position is close to the road side on the left side or the right side from the center in the width direction of the driving road, and the road in the landscape image taken by the camera 59 is close. Movement of feature points on the side (feature points on the left side of the image if close to the left side of the road from the center in the width direction of the road, and feature points on the right side of the image if close to the road side on the right side of the center in the width direction of the road) This is a case where the speed is relatively high and a recognition failure of a landscape image is likely to occur. Therefore, the closer the driving lane Lr is to the side of the road, the smaller the driving lane coefficient k3 and the determination threshold value Mref can be made easier to determine that the determination data and the reference data match. It is possible to suppress omissions in recognition of landscape images. Further, when the absolute value of the travel lane difference ΔL is large (when the travel lane Lr is away from the travel lane Ls at the time of photographing), when photographing a landscape image in which the vehicle position has obtained reference data in the width direction of the traveling road The feature point in the landscape image captured by the camera 59 (the image is displayed when the vehicle position is distant to the left from the vehicle position at the time of capturing the landscape image obtained from the reference data). When moving to the right side from the vehicle position at the time of shooting the landscape image from which the feature points on the left side and the reference data are obtained, the movement speed of the feature points on the right side of the image is relatively high, which causes a failure to recognize the landscape image. Therefore, the smaller the absolute value of the traveling lane difference ΔL is, the smaller the traveling lane difference coefficient k4 is set to decrease the determination threshold value Mref, thereby making it easier to determine that the determination data matches the reference data. It is possible to suppress the omission of recognition of landscape images.

  Subsequently, the lane number coefficient k5, the lane width coefficient k6, and the sidewalk width coefficient k7 to be multiplied by the determination reference threshold M0 are set based on the input parameters of the lane number Ln, the lane width Lw, and the sidewalk width Pw, respectively (step S160) When the input road shape is a curve shape, a curvature radius coefficient k8 to be multiplied by the determination reference threshold value M0 is set based on the curvature radius R, and when the input road shape is a straight shape, the value is set to the curvature radius radius coefficient k8. 1 is set (step S170), and an intersection distance coefficient k9 to be multiplied by the determination reference threshold M0 is set based on the input distance Dc from the nearest intersection (step S180). Here, in the embodiment, for each coefficient, the relationship between the parameter and the coefficient is determined in advance and stored in the ROM 34 as a coefficient setting map for each coefficient. When a parameter is given, the corresponding coefficient is derived from the stored map. And set it.

  FIG. 9 shows an example of a plurality of relationships between parameters and coefficients for setting the determination threshold value Mref. In the drawing, the right side is an example of a coefficient setting map. As shown in the figure, the smaller the lane number Ln, the smaller the lane number coefficient k5 across the value 1, the smaller the lane width Lw, the smaller the lane width coefficient k6 across the value 1, and the smaller the sidewalk width Pw, the sidewalk When the width coefficient k7 decreases across the value 1 and the road shape is a curve shape, the smaller the radius of curvature R, the smaller the radius of curvature coefficient k8 crosses the value 1, and the distance Dc from the nearest intersection is a predetermined value described later. Each coefficient is set based on each parameter so that the intersection distance coefficient k9 becomes smaller in the range less than 1 as the distance is shorter than the distance D1. As for the sidewalk width coefficient k7, when there is no sidewalk on the road running with the sidewalk width Pw of 0, a lower limit value less than the predetermined value 1 of the sidewalk width coefficient k7 is set. . As for the intersection distance coefficient k9, the value 1 is set when the distance Dc from the nearest intersection is equal to or greater than the predetermined distance D1. The reason for setting each coefficient in this way will be described next.

  First, when the width of the running road is relatively small, the moving speed of the feature points in the landscape image captured by the camera 59 is relatively large, and the recognition of the landscape image is likely to occur. On the other hand, when the road width is relatively small, if the determination threshold value Mref is reduced, it can be easily determined that the determination data matches the reference data. As a case where the width of the road is relatively small, it is possible to consider a case where the number of lanes on the road is small, a case where the lane width is narrow, a case where the sidewalk width is narrow, etc. In the embodiment, the smaller the lane number Ln, the smaller the lane The narrower the width Lw and the narrower the sidewalk width Pw, the smaller the coefficients and the smaller the threshold value for judgment Mref, so that recognition failure of the landscape image can be suppressed.

  Further, when the running road has a curved shape and the radius of curvature R is relatively small, the moving speed of the feature point in the landscape image taken by the camera 59 (particularly in the direction corresponding to the direction opposite to the turning direction of the vehicle). ) Is relatively large, and the range of shooting positions on the road where a landscape image having a feature point matching the reference data can be shot is quite narrow, so that recognition of the landscape image is likely to occur. On the other hand, as the curvature radius R of the road is smaller, the curvature radius coefficient k8 is decreased to decrease the determination threshold value Mref, so that it can be easily determined that the determination data and the reference data match. It is possible to suppress omissions in recognition of landscape images. Furthermore, as can be seen from the example of the vehicle shown in FIG. 10 turning left at the intersection, the vehicle enters the road diagonally with respect to the road direction when the vehicle turns right or left at the intersection and enters the road. However, after the vehicle has traveled to some extent, the vehicle travels in a direction that substantially matches the road direction. For this reason, in the embodiment, since the intersection distance coefficient k9 is decreased and the determination threshold Mref is decreased as the distance Dc from the nearest intersection on the road is shorter than the predetermined distance D1, the determination data and the reference data are reduced. Can be easily determined to match, and recognition failure of landscape images can be suppressed. Therefore, the predetermined distance D1 can be a distance determined in advance through experiments and analysis as a distance until the vehicle travels in a direction substantially coincident with the road direction on the road on which the vehicle that has turned right or left enters the intersection.

  When the threshold values k1 to k9 are set in this way, a value obtained by multiplying the determination reference threshold value M0 by all the threshold values k1 to k9 is set as the determination threshold value Mref (step S190), and the matching rate M between the determination data and the reference data is set. Calculation is performed (step S200), and the calculated match rate M is compared with the determination threshold value Mref (step S210). Here, the matching rate M represents the degree to which the determination data and the reference data match. For example, the corresponding pixels of both data (the total number of pixels and the aspect ratio are the same) as a binarized image. Can be calculated as a ratio of the number of pixels having the same value (value 0 or value 1) to the total number of pixels.

  When the calculated coincidence rate M is less than the determination threshold Mref, it is determined that the determination data and the reference data do not match, and the reference data of the shooting position and shooting direction that match the estimated vehicle position and vehicle traveling direction, and It is determined whether or not a total of three determination processes, that is, a total of three reference data of the two reference data before and after the road and one determination data, has been performed (step S220). If the determination process has not been performed three times in total, the process returns to step S100, and the shooting position and the shooting that match the estimated vehicle position and the vehicle traveling direction previously input in step S100 are replaced without changing the determination data. Instead of the reference data of the direction, the reference data (one reference data ahead) in the vehicle traveling direction is input on the road at the photographing position, and the processes in steps S100 to S210 are repeated. Again, when the coincidence rate M is less than the determination threshold value Mref and the determination process is not performed three times in total (steps S210 and S220), the process returns to step S100, and the determination data is not changed and is first changed to step S100. Instead of the reference data of the shooting position and shooting direction that match the estimated host vehicle position and vehicle traveling direction input in step S100, step S100 to step S100 are performed. The process of S210 is repeated.

  When the determination result that the coincidence rate M is less than the determination threshold value Mref is repeatedly obtained three times (steps S210 and S220), the estimated vehicle position is determined as the vehicle position (step S230), and the vehicle position determination process is performed. End the routine. On the other hand, when it is determined that the match rate M is equal to or greater than the determination threshold value Mref while repeating such processing three times, the shooting position corresponding to the reference data that matches the determination data at that time is determined as the own vehicle position. (Step S240), and the vehicle position determination processing routine is terminated. By such processing, it is possible to specify the vehicle position that is more accurate than the estimated vehicle position estimated by GPS navigation, autonomous navigation, or the like in the location processing. When the vehicle position is specified in this way, the next vehicle position is estimated by GPS navigation, autonomous navigation, or the like with the specified vehicle position as a reference.

  The processing of each step of this routine is basically performed by the location processing unit 60 of the electronic control unit 30, and the creation of determination data in the input processing of step S100 is performed by the image processing unit 66 of the electronic control unit 30. It is.

  According to the navigation device 20 of the embodiment described above, a plurality of reference data obtained by extracting feature points of a landscape image taken from a camera 59 mounted on the road with an interval in the direction of travel are respectively taken correspondingly. It is stored in the reference data DB 44 in association with the position, and a matching process is performed to determine whether or not a landscape image taken from the camera 59 mounted on the vehicle matches a plurality of reference data. In the method of determining the vehicle position that is the current position of the vehicle based on the shooting position corresponding to the reference data determined to be, the matching process extracts feature points from the landscape image shot from the camera 59 When the matching rate M between the determination data obtained by performing predetermined image processing and the reference data is equal to or higher than the determination threshold Mref The determination threshold Mref is determined to be smaller as the vehicle speed V is larger, and the steering wheel steering amount St is smaller than the determination threshold Mref. Since the adjustment is performed such that the larger the value is, the smaller the determination data is, the easier it is to determine that the determination data and the reference data match. As a result, the vehicle position can be determined more appropriately.

  Further, according to the navigation device 20 of the embodiment, the vehicle position in the width direction of the traveling road is traveling so that the traveling lane Lr becomes smaller as it is closer to the roadside of the traveling road. Since the determination threshold value Mref is adjusted so as to become smaller as it is closer to the road side of the road, it can be more easily determined that the determination data matches the reference data, and so-called recognition omission of a landscape image is prevented. It can be suppressed more. Further, the reference value of the vehicle position is obtained so that the absolute value of the travel lane difference ΔL increases (the travel lane Lr becomes farther from the travel lane Ls at the time of shooting), that is, the vehicle position in the width direction of the road on which the vehicle is traveling Since the determination threshold value Mref is adjusted so as to become smaller as the distance from the vehicle position at the time of landscape image shooting increases, it can be further easily determined that the determination data and the reference data match, Recognition failure can be further suppressed.

  In addition, according to the navigation device 20 of the embodiment, the smaller the width of the road on which it is traveling (specifically, the smaller the lane number Ln, the narrower the lane width Lw, the narrower the sidewalk width Pw), the smaller. In addition, as the traveling road is curved and the curvature radius R is smaller, the smaller the distance Dc from the nearest intersection on the traveling road is, the smaller it is. Since the determination threshold value Mref is adjusted, it can be easily determined that the determination data matches the reference data, and the recognition failure of the landscape image can be further suppressed. As a result, the vehicle position can be determined more appropriately.

  In the navigation device 20 according to the embodiment, the determination threshold value Mref used for determining the coincidence rate M between the determination data and the reference data is such that the traveling lane Lr becomes closer to the road as the vehicle speed V increases and the steering wheel steering amount St increases. The closer the distance Dc from the nearest intersection is, the closer the absolute value of the driving lane difference ΔL is, the smaller the number of lanes Ln, the narrower the lane width Lw, the narrower the sidewalk width Pw, and the smaller the radius of curvature R. It is assumed that the adjustment is made so that the shorter the value is, that is, the adjustment is made based on nine parameters, but instead of the adjustment based on the number of lanes Ln, the lane width Lw, and the sidewalk width Pw. Alternatively, the determination threshold value Mref may be adjusted so as to be shorter as the road width Dw is smaller. Further, the determination threshold Mref may be adjusted based on only a part (one or more) of the nine parameters including the road width Dw added to the nine parameters. Adjustments may be made based on other parameters in addition to the ten parameters.

  Therefore, for example, the determination threshold value Mref used for determining the matching rate M between the determination data and the reference data is not adjusted based on parameters other than the vehicle speed V and the steering wheel steering amount St, and decreases as the vehicle speed V increases. The steering wheel steering amount St may be adjusted to be smaller as the steering wheel steering amount St is larger. Further, the determination threshold value Mref may be adjusted so as to decrease as the traveling lane Lr is closer to the roadside without being adjusted based on parameters other than the traveling lane Lr. Further, the determination threshold Mref may be adjusted so as to decrease as the absolute value of the travel lane difference ΔL increases without being adjusted based on parameters other than the travel lane difference ΔL.

  In the navigation device 20 of the embodiment, the determination threshold value Mref used for determining the matching rate M between the determination data and the reference data is adjusted so as to decrease as the travel lane Lr is closer to the roadside. Since the determination threshold value Mref may be adjusted to be smaller as the vehicle position is closer to the roadside in the width direction of the traveling road, the vehicle position in the width direction of the road can be determined with respect to the determination data without detecting the traveling lane Lr. It is good also as what detects by image processing and adjusts threshold value Mref for judgment based on this detection position.

  In the navigation device 20 of the embodiment, the determination threshold Mref used for determining the coincidence rate M between the determination data and the reference data is adjusted so as to decrease as the travel lane Lr moves away from the shooting travel lane Ls. However, since the determination threshold value Mref may be adjusted to be smaller as the vehicle position moves away from the vehicle position at the time of shooting the landscape image in which the reference data is obtained in the width direction of the road on which the vehicle is traveling, the traveling lane Lr is detected. Alternatively, the vehicle position in the width direction of the road may be detected by image processing on the determination data, and the determination threshold value Mref may be adjusted based on the detected position.

  In the navigation apparatus 20 of the embodiment, the vehicle position is determined by comparing the coincidence rate M between the determination data and the reference data with the determination threshold value Mref. However, the determination data and the reference data are respectively transmitted from the vehicle. In view of this, the vehicle position may be determined by dividing the left side data and the right side data and then comparing the matching rate and the determination threshold value on the left side and the right side, respectively. In this case, the vehicle position determination process routine of FIG. 11 may be executed instead of the vehicle position determination process routine of FIG. In the routine of FIG. 11, the processing of step S300 is executed instead of the processing of step S150 of the routine of FIG. 4, and the processing of steps S310 to S340 is executed instead of the processing of steps S190 to S210 of the routine of FIG. Is the same as the routine of FIG. Therefore, the same process is given the same step number, and the detailed description thereof is omitted.

  In the routine of FIG. 11, when the vehicle speed coefficient k1, the steering amount coefficient k2, and the travel lane coefficient k3 are set (steps S130 and S140), the left side in the landscape image is determined based on the calculated travel lane difference ΔL. A left travel lane difference coefficient k4L and a right travel lane difference coefficient k4R for determining the right side in the landscape image are set (step S300), and other coefficients k5 to k9 are set (steps S160 to S180). The left travel lane difference coefficient k4L is set using the left travel lane difference coefficient setting map stored in the ROM 34 with the relationship between the travel lane difference ΔL and the left travel lane difference coefficient k4L determined in advance. k4R is set using a right travel lane difference coefficient setting map stored in the ROM 34 with the relationship between the travel lane difference ΔL and the right travel lane difference coefficient k4R determined in advance. FIG. 12 shows an example of a left-hand travel lane difference coefficient setting map, and FIG. 13 shows an example of a right-hand drive lane difference coefficient setting map. The contents of each map will be described later.

  Subsequently, a value obtained by multiplying the vehicle speed coefficient k1, the steering amount coefficient k2, the travel lane coefficient k3, the left travel lane difference coefficient k4L, and other coefficients k5 to k9 is set as the left determination threshold MLref, and the vehicle speed coefficient k1 An image obtained by multiplying the steering amount coefficient k2, the travel lane coefficient k3, the right travel lane difference coefficient k4R, and other coefficients k5 to k9 as the right determination threshold MLref (step S310), and viewing the determination data from the vehicle The data is divided into left-side left determination data and right-side right determination data as data, and reference data is divided into left-side left reference data and right-side right reference data as image data viewed from the vehicle (step S320). FIG. 14 shows an example of how the reference data is divided into left reference data and right reference data. In the example of FIG. 14, the reference data is divided in half on the left side and the right side. The determination data is also divided in the same manner as the reference data. The reference data may be divided in advance and stored in the reference data DB 44 of the HDD 40.

  Next, the matching rate ML between the left determination data and the right reference data is calculated in the same manner as the matching rate M, and the matching rate MR between the right determination data and the right reference data is the same as the matching rate M. (Step S320), the calculated match rate ML is compared with the left determination threshold value MLref, and the match rate MR is compared with the right determination threshold value MRref (step S340). Then, while the processes in steps S100 to S340 are repeated three times using different reference data, the match rate ML is determined to be three times continuously, less than the left determination threshold value MLref, or the match rate MR is less than the right determination threshold value MRref. Sometimes, the estimated vehicle position is determined as the vehicle position (step S230). On the other hand, when it is determined that the match rate ML is equal to or greater than the left-side determination threshold value MLref and the match rate MR is equal to or greater than the right-side determination threshold value MRref while the processing in steps S100 to S340 is repeated three times using different reference data. Then, the shooting position corresponding to the reference data that matches the determination data is determined as the vehicle position (step S240), and the vehicle position determination processing routine is terminated.

  Here, the left traveling lane difference coefficient setting map and the right traveling lane difference coefficient setting map illustrated in FIGS. 12 and 13 will be described. As shown in FIG. 12, the left-side travel lane difference coefficient k4L is equal to the right-hand side of the shooting-time travel lane Ls when the travel-lane difference ΔL is 0 or more, that is, whether the travel lane Lr is the same as the shooting-time driving lane Ls. A value of 1 is set when the vehicle is in a lane. Further, the left lane difference coefficient k4L is set to be smaller as the lane difference ΔL is smaller than 0, that is, to be smaller as the lane Lr is further to the left from the shooting lane Ls. Is done. This is because the moving speed of the feature point on the left side in the landscape image captured by the camera 59 tends to be relatively higher as the traveling lane Lr is farther leftward in the width direction of the road compared to the traveling lane Ls at the time of shooting. This tendency is based on the fact that scenery image recognition failure occurs. Further, as shown in FIG. 13, the right travel lane difference coefficient k4R is determined when the travel lane difference ΔL is 0 or less, that is, whether the travel lane Lr is the same as the travel lane Ls at the time of shooting or from the travel lane Ls at the time of shooting. A value of 1 is set for the left lane. Further, the right travel lane difference coefficient k4R is set so as to decrease as the travel lane difference ΔL is greater than 0, that is, as the travel lane Lr is further away from the shooting travel lane Ls to the right. The This is because the moving speed of the feature point on the right side in the landscape image captured by the camera 59 tends to be relatively higher as the traveling lane Lr is further to the right in the width direction of the road than the traveling lane Ls at the time of shooting. This tendency is based on the fact that scenery image recognition failure occurs. The vehicle position is determined by using the left determination threshold value Mref and the right determination threshold value Mref, which are set using these coefficients, for comparison with the left match rate ML and the right match rate MR, respectively. Recognition failure is suppressed, and the vehicle position can be determined more appropriately.

  In the embodiment, the position determination device according to the present invention is applied to the navigation device 20 including the display 22 and the HDD 40. However, for example, the location of the electronic control unit 30 is connected to a server communicably connected to the navigation device including the display. The server may function as the position determination device of the present invention by performing the function of the processing unit 60 and a part of the function of the image processing unit 66 necessary for the location processing. Further, instead of the navigation device 20, devices such as a digital camera, a portable terminal, and a drive recorder may function as the position determination device of the present invention. In this case, such a device may be installed in the vehicle interior, and the vehicle position may be determined using a landscape image taken by a camera provided in the device.

  In the embodiments, the position determination device of the present invention is applied to the navigation device 20. However, the position determination method may be used, and each step of the position determination method may be realized by one or more computers. The program may be in the form of Further, such a program may be stored in a recording medium.

  The correspondence between the main elements of the embodiment and the main elements of the invention described in the column of means for solving the problems will be described. In the embodiment, a plurality of reference data is stored in the reference data DB 42 of the HDD 40, and a matching process is performed to determine whether or not a landscape image captured from the camera 59 mounted on the vehicle matches the plurality of reference data. A location processing unit 60 of the electronic control unit 30 that executes the vehicle position determination processing routine of FIG. 4 or FIG. 11 that determines the vehicle position based on the imaging position corresponding to the reference data determined to match. A part of the image processing unit 66 corresponds to a “position determination device”. Further, the location processing unit 60 of the electronic control unit 30 that executes the process of step S110 of the host vehicle position determination process routine of FIG. 4 or FIG. 11 that detects the travel lane Lr using the determination data is “width direction vehicle position detection”. It corresponds to “means”.

  The correspondence between the main elements of the embodiment and the main elements of the invention described in the column of means for solving the problem is the same as that of the embodiment described in the column of means for solving the problem. Therefore, the elements of the invention described in the column of means for solving the problems are not limited. That is, the interpretation of the invention described in the column of means for solving the problems should be made based on the description of the column, and the examples are those of the invention described in the column of means for solving the problems. It is only a specific example.

  As mentioned above, although the form for implementing this invention was demonstrated using the Example, this invention is not limited at all to such an Example, In the range which does not deviate from the summary of this invention, it is with various forms. Of course, it can be implemented.

  The present invention can be used in the manufacturing industry of navigation devices and position determination devices.

  10 navigation system, 12 in-vehicle device, 20 navigation device, 22 display, 24 touch panel, 26 speaker, 30 electronic control unit, 32 CPU, 34 ROM, 36 RAM, 38 flash memory, 40 hard disk drive (HDD), 42 for map data Database (Map Data DB), 44 Reference Data Database (Reference Data DB), 50 GPS Receiver, 52 Vehicle Speed Sensor, 54 Acceleration Sensor, 56 Direction Sensor, 58 Steering Amount Sensor, 59 Camera, 60 Location Processing Unit, 62 Map Display processing unit, 64 navigation processing unit, 66 image processing unit.

Claims (11)

A plurality of reference data obtained by extracting characteristic parts of a landscape image taken from an in-vehicle camera at intervals in the traveling direction on the road are stored in association with the corresponding positions on the map, and taken from the in-vehicle camera. Matching processing is performed to determine whether or not the scenery image to be matched and the plurality of reference data match, and based on the position on the map corresponding to the reference data determined to match by the matching processing A position determination device for determining a position of a vehicle on a map,
The matching process is performed when the degree of coincidence between the reference data and the determination data obtained by performing predetermined image processing for extracting a characteristic part from a landscape image photographed from a vehicle is equal to or greater than a determination threshold value. It is performed by determining that they match and determining that they do not match when the degree of matching is less than the threshold for determination,
The determination threshold is adjusted so as to decrease as the vehicle speed increases and to decrease as the steering wheel steering amount increases.
Position determination device. The position determination device according to claim 1,
A width direction vehicle position detecting means for detecting a vehicle position in the width direction of the running road,
The determination threshold is adjusted so that the detected vehicle position tends to decrease as the detected vehicle position is closer to the side of the road on which the vehicle is running.
Position determination device. The position determination device according to claim 2,
The width direction vehicle position detecting means is means for detecting a traveling lane when traveling on a road having a plurality of lanes,
The threshold for determination is adjusted to tend to be smaller as the detected travel lane is closer to the side of the road on which the vehicle is traveling.
Position determination device. The position determination device according to claim 1,
A width direction vehicle position detecting means for detecting a vehicle position in the width direction of the running road,
The threshold value for determination is adjusted so that the detected vehicle position in the width direction of the running road becomes smaller as the distance from the vehicle position at the time of capturing a landscape image obtained from the reference data decreases.
Position determination device. The position determination device according to claim 4,
The width direction vehicle position detecting means is means for detecting a traveling lane when traveling on a road having a plurality of lanes,
The determination threshold is adjusted such that the detected travel lane tends to become smaller as the distance from the travel lane at the time of capturing a landscape image from which the reference data is obtained.
Position determination device. The position determination device according to any one of claims 1 to 5,
The threshold for judgment tends to be smaller as the width of the running road is smaller, the tendency to be smaller as the radius of curvature of the running road is smaller, and the distance from the nearest intersection on the running road is shorter Adjusted to be at least one of the tendency to become smaller,
Position determination device. The position determination device according to any one of claims 1 to 6,
The threshold value for judgment tends to decrease as the number of lanes on the road that travels decreases, tends to decrease as the lane width of the road that travels decreases, and tends to decrease as the sidewalk width of the road that travels decreases. Adjusted to be at least one of
Position determination device. The position determination device according to claim 4 or 5,
The matching process includes
The determination data is divided into left determination data on the left side and right determination data on the right side when viewed from the vehicle, and the reference data is divided into left reference data on the left side and right reference data on the right side when viewed from the vehicle. And
The degree of coincidence between the left determination data and the left reference data is equal to or greater than the left determination threshold, and the degree of coincidence between the right determination data and the right reference data is equal to or greater than the right determination threshold. The degree of matching between the left-side determination data and the left-side reference data is less than the left-side determination threshold value, or the degree of matching between the right-side determination data and the right-side reference data Is determined to be inconsistent when the value is less than the right determination threshold,
The left-side determination threshold is adjusted so that the detected vehicle position in the width direction of the traveling road decreases toward the left side from the vehicle position at the time of capturing a landscape image obtained from the reference data. Become
The determination threshold for the right side is adjusted so that the detected vehicle position in the width direction of the traveling road tends to decrease as it moves to the right side from the vehicle position at the time of shooting the landscape image obtained from the reference data. Become,
Position determination device.   A navigation device comprising the position determination device according to any one of claims 1 to 8, and performing route guidance to a destination using a position on a map of a vehicle determined by the position determination device. A plurality of reference data obtained by extracting characteristic portions of a landscape image taken from an in-vehicle camera on the road at intervals in the running direction are stored in association with the corresponding positions on the map, and taken from the vehicle. A matching process for determining whether or not the landscape image matches the plurality of reference data, and a position on the map corresponding to the reference data determined to be constant by the matching process Determining the position of the vehicle on the map, and a position determination method executed by a position determination device that executes :
The matching process is performed when the degree of coincidence between the reference data and the determination data obtained by performing predetermined image processing for extracting a characteristic part from a landscape image photographed from a vehicle is equal to or greater than a determination threshold value. It is performed by determining that they match and determining that they do not match when the degree of matching is less than the threshold for determination,
The determination threshold is adjusted so as to decrease as the vehicle speed increases and to decrease as the steering wheel steering amount increases.
Position determination method. A program for causing the one or more computers to function as the position determination device by executing each step of the position determination method according to claim 10 by one or more computers .

Images