JP6948042B2 - Position estimation device, position detection method and program - Google Patents

Description

The present invention relates to a position estimation device, a position estimation method and a program.

Although GPS (Global Positioning System) is generally used to estimate the current position of a moving object such as an automobile, GPS accurately positions the current position of the moving object due to interference on the ground. It may not be possible to estimate.

For example, in the valley of a skyscraper in an urban area, the current position estimated by GPS may deviate significantly from the actual current position. Further, for example, when moving in a tunnel, communication cannot be performed between the satellite constituting the GPS and the moving body, and as a result, the current position of the moving body may not be estimated.

A technique for compensating for the inaccuracy of the current position estimated by GPS has been proposed. For example, in Patent Document 1, by averaging a plurality of absolute positions at different times calculated based on GPS signals for the same point, the positioning position calculated based on GPS signals is more accurate. It has been proposed to correct to position information.

Japanese Unexamined Patent Publication No. 2012-208525

However, the technique described in Patent Document 1 is for collecting highly accurate positioning data regarding a stationary object around a road. As a result, for example, even if it is possible to collect highly accurate positions of buildings and roads and create map data, the technique described in Patent Document 1 as described above is used to estimate the current position of the moving body. It is difficult to apply to. In addition, there is almost no technology for estimating the current position of a moving object based on the created map data.

The present invention has been made in view of the above circumstances, and an object of the present invention is to provide a position estimation device or the like capable of accurately estimating the current position of a moving body.

The position estimation device according to the first aspect of the present invention is
Four forward side image feature is captured by a camera mounted on the front, rear, left and right of the moving body, the rearward image, continuously the current image data currently contains an image composed of the left image and right image at regular time intervals The current image acquisition unit to be acquired and
An image for acquiring image map data in which a reference image including a front image, a rear image, a left image, and a right image of a feature obtained by taking a picture in advance and a shooting position information of the reference image are associated with each other. Map acquisition department and
An image matching unit that identifies a reference image corresponding to the current image by collating the current image of the acquired current image data with the reference image of the acquired video map data.
An estimation unit that estimates the current position of the moving body based on the shooting position information associated with the reference image corresponding to the specified current image in the video map data.
It is provided with an output unit that outputs current position data including the estimated current position .
The image collating unit further collates the vector indicating the change of the plurality of current images of the acquired current image data with the vector indicating the change of the reference image of the acquired video map data, thereby both vectors. The reference image having the highest degree of similarity is identified as the reference image corresponding to the change of the plurality of current images.
The estimation unit further estimates the moving direction of the moving body based on the shooting position information associated with the reference image corresponding to the change of the specified plurality of current images in the video map data.
The output section is further characterized in that you output the moving direction data including the estimated direction of movement.

The position estimation method according to the second aspect of the present invention is
Four forward side image feature is captured by a camera mounted on the front, rear, left and right of the moving body, the rearward image, continuously the current image data currently contains an image composed of the left image and right image at regular time intervals To get and
Acquiring video map data in which a reference image including a front image, a rear image, a left image, and a right image of a feature obtained by shooting in advance and a shooting position information of the reference image are associated with each other. When,
Identifying the reference image corresponding to the current image by collating the current image of the acquired current image data with the reference image of the acquired video map data.
To estimate the current position of the moving body based on the shooting position information associated with the reference image corresponding to the specified current image in the video map data.
Only containing and outputting a current position data including the estimated current position,
In specifying the reference image, further, by collating the vector showing the change of the plurality of current images of the acquired current image data with the vector showing the change of the reference image of the acquired video map data. , The reference image having the highest similarity between the two vectors is specified as the reference image corresponding to the change of the plurality of current images.
In estimating the current position of the moving body, the moving body is further based on the shooting position information associated with the reference image corresponding to the change of the specified plurality of current images in the video map data. Estimate the direction of movement,
The output of the current position data is further characterized in that the movement direction data including the estimated movement direction is output.

The program according to the third aspect of the present invention
Computer,
Four forward side image feature is captured by a camera mounted on the front, rear, left and right of the moving body, the rearward image, continuously the current image data currently contains an image composed of the left image and right image at regular time intervals Current image acquisition department to acquire,
An image for acquiring image map data in which a reference image including a front image, a rear image, a left image, and a right image of a feature obtained by taking a picture in advance and a shooting position information of the reference image are associated with each other. Map acquisition department,
An image matching unit that identifies a reference image corresponding to the current image by collating the current image of the acquired current image data with the reference image of the acquired video map data.
An estimation unit that estimates the current position of the moving body based on the shooting position information associated with the reference image corresponding to the specified current image in the video map data.
It functions as an output unit that outputs the current position data including the estimated current position.
The image collating unit further collates the vector indicating the change of the plurality of current images of the acquired current image data with the vector indicating the change of the reference image of the acquired video map data, thereby both vectors. The reference image having the highest degree of similarity is identified as the reference image corresponding to the change of the plurality of current images.
The estimation unit further estimates the moving direction of the moving body based on the shooting position information associated with the reference image corresponding to the change of the specified plurality of current images in the video map data.
The output unit is a program for further functioning to output movement direction data including the estimated movement direction.

According to the present invention, it is possible to accurately estimate the current position of the moving body.

It is a figure which shows the functional structure of the position estimation apparatus which concerns on Embodiment 1 of this invention. It is a figure which shows an example of the structure of the map data included in the video map data. It is a figure which shows an example of the structure of the video data included in the video map data. It is a flowchart which shows an example of the flow of the position estimation processing which concerns on Embodiment 1. FIG. It is a figure which shows the composition example of the present image data which is continuously acquired by the present image acquisition part. It is a flowchart which shows an example of the flow of the present position estimation processing which concerns on Embodiment 1. FIG. It is a flowchart which shows an example of the flow of the moving road estimation process. It is a flowchart which shows an example of the flow of a moving direction estimation process. It is a figure which shows the functional structure of the position estimation apparatus which concerns on one modification. It is a figure which shows the functional structure of the position estimation apparatus which concerns on Embodiment 2 of this invention. It is a flowchart which shows an example of the flow of the part different from the position estimation process which concerns on Embodiment 1 in the position estimation process which concerns on Embodiment 2. It is a flowchart which shows an example of the flow of the present position estimation processing which concerns on Embodiment 2. It is a figure which shows an example of the physical structure of the position estimation apparatus which concerns on this invention.

Embodiments of the present invention will be described with reference to the drawings.

(Embodiment 1)
The position estimation device 100 according to the first embodiment of the present invention is a device for estimating the current position of a moving body and is mounted on an automobile. An automobile is an example of a vehicle included in a moving body, and the moving body includes a vehicle, a person, and the like as an object moving on the ground.

Functionally, as shown in FIG. 1, the position estimation device 100 according to the present embodiment has four cameras 101a to 101d, a current image acquisition unit 102, an approximate position acquisition unit 103, and a video map storage unit 104. A video map acquisition unit 105, an image collation unit 106, an estimation unit 107, and an output unit 108 are provided.

The cameras 101a to 101d are attached to the front, rear, left and right sides of the automobile in order to photograph a feature. Here, examples of features include buildings, roads, and road signs.

Further, each of the cameras 101a to 101d continuously outputs the current image data indicating the captured current image (current image) at regular intervals. The timing at which each of the cameras 101a to 101d currently outputs image data may be appropriately predetermined, and is preferably synchronized.

Specifically, the camera 101a is fixed to the front of the automobile, captures a predetermined range of features in front of the automobile, and continuously outputs current image data at regular intervals. The camera 101b is fixed to the rear part of the automobile, captures a predetermined range of features behind the automobile, and continuously outputs current image data at regular intervals. The camera 101c is fixed to the left side of the automobile, captures a predetermined range of features on the left side of the automobile, and continuously outputs current image data at regular intervals. The camera 101d is fixed to the right side of the automobile, captures a predetermined range of features on the right side of the automobile, and continuously outputs current image data at regular intervals.

In the following, unless each of the cameras 101a to 101d is particularly distinguished, the cameras 101a to 101d will be referred to as a camera 101. In the present embodiment, that describes the example provided four cameras 101.

When the current image data is output from each of the cameras 101, the current image acquisition unit 102 acquires the output current image data.

The approximate position acquisition unit 103 acquires approximate position data indicating the approximate position. The approximate position is the current position of the vehicle estimated by GPS (Global Positioning System). The approximate position data acquired by the approximate position acquisition unit 103 according to the present embodiment indicates the current position of the automobile estimated by transmitting and receiving signals to and from the satellite.

The video map storage unit 104 is a storage unit that stores the video map data 109.

The video map data 109 includes, for example, map data 109a including information represented on a general map and video data 109b including an image of a feature taken at a predetermined point.

As illustrated in FIG. 2, the map data 109a includes data on landmarks (relatively large buildings, facilities, etc. along the road) and roads, which are predetermined features. In the map data 109a, identification information for identifying each is associated with each position information.

The location information of the landmark and the road indicates the position of the feature point included in the area occupied by each, as shown in FIG. 2, for example. The position of the feature point is represented by, for example, latitude and longitude.

The feature points included in the landmark are, for example, points that become corners when the outer edges of the area occupied by the landmark are approximately connected by a straight line. Further, the feature points included in the road are, for example, points that become corners when each side portion of the road is approximately connected by a straight line.

The video data 109b is data collected by, for example, a vehicle or a person having an omnidirectional camera actually moving on a road or the like.

As shown in FIG. 3, the video data 109b associates a reference image including an image of a feature obtained by shooting at a predetermined point (shooting position) with the shooting position information of the reference image. Has been done.

Specifically, in the example of FIG. 3, N is an arbitrary natural number, and images in the front, back, left, and right directions taken at the shooting position are associated with shooting position information indicating the shooting position. The shooting position N is represented by latitude and longitude, as in the case of the above-mentioned feature point positions, for example. Then, the front image N, the rear image N, the left image N, and the right image N, which are the front, rear, left, and right images taken from the points indicated by the shooting positions N, are associated with the shooting position N. Has been done. Here, the front and back are directions that are predetermined along the road, and the left and right are directions that are determined according to the front and back.

The video map acquisition unit 105 acquires the video map data 109 from the video map storage unit 104.

The video map acquisition unit 105 according to the present embodiment acquires video map data narrowed down based on the approximate position indicated by the approximate position data acquired by the approximate position acquisition unit 103. Specifically, the video map acquisition unit 105 acquires video map data including a reference image taken at a shooting position in a predetermined range with respect to the approximate position indicated by the approximate position data.

The image collation unit 106 identifies the reference image corresponding to the current image by collating the current image with the reference image. The current image for collation is the current image included in the current image data acquired by the current image acquisition unit 102. The reference image for collation is a reference image included in the video map data acquired by the video map acquisition unit 105.

Image matching unit 106 according to the present embodiment, further, by matching the reference image and the change of the plurality of current images, specifying the reference image corresponding to the change of the plurality of current images. The changes in the plurality of current images for collation are changes in the plurality of current images included in the current image data acquired by the current image acquisition unit 102, and the changes in the current image data acquired by the current image acquisition unit 102. Detected based on multiple current images included. The reference image for collation is a reference image included in the video map data acquired by the video map acquisition unit 105, as described above.

The estimation unit 107 estimates the current position of the automobile based on the reference image specified by the image matching unit 106. Specifically, the estimation unit 107 estimates the current position of the automobile based on the shooting position information associated with the specified reference image in the video map data 109.

The estimation unit 107 according to the present embodiment further estimates the moving direction of the automobile based on the shooting position information associated with the specified reference image in the video map data 109.

The estimation unit 107 according to the present embodiment further estimates the landmarks located around the automobile and the road on which the automobile is moving based on the video map data 109 and the current image data. The video map data 109 and the current image data used for this estimation are data acquired by the video map acquisition unit 105 and the current image acquisition unit 102, respectively.

The output unit 108 outputs the current position data including the current position estimated by the estimation unit 107.

The output unit 108 according to the present embodiment further outputs movement direction data including the movement direction estimated by the estimation unit 107.

The output unit 108 according to the present embodiment further outputs moving road data indicating the moving road estimated by the estimation unit 107.

So far, the configuration of the position estimation device 100 according to the first embodiment of the present invention has been described. From here, the operation of the position estimation device 100 according to the first embodiment of the present invention will be described.

The position estimation device 100 starts the position estimation process illustrated in FIG. 4, for example, when the power of the automobile is turned on. The position estimation process is a process of estimating the current position of an automobile.

The current image acquisition unit 102 acquires the current image data from each camera 101 (step S101).

Currently, image data is continuously acquired at regular time intervals, as shown in FIG. In the example shown in the figure, the current image data includes shooting time information indicating a shooting time and an image shot by each camera 101 at each shooting time. For example, the shooting time information indicating the time T1 is associated with a front image T1, a rear image T1, a left image T1, and a right image T1, which are images taken at the time T1. The time T1 is, for example, a value representing the elapsed time from a predetermined reference time to the present in units of 1/100 second.

The approximate position acquisition unit 103 acquires the approximate position data (step S102).

The video map acquisition unit 105 acquires the video map data 109 from the video map storage unit 104 with reference to the approximate position data acquired in step S102 (step S103).

Specifically, the video map acquisition unit 105 refers to the approximate position data acquired by the approximate position acquisition unit 103. The video map acquisition unit 105 specifies a predetermined range with respect to the approximate position indicated by the referenced approximate position data. The predetermined range with respect to the approximate position is, for example, a range of several kilometers around the approximate position. The video map acquisition unit 105 acquires video map data including a reference image shot with a point within the specified range as a shooting position. By such processing, the video map acquisition unit 105 narrows down the video map data 109 acquired from the video map storage unit 104 based on the approximate position data.

The image collation unit 106, the estimation unit 107, and the output unit 108 estimate the current position of the automobile based on the current image data and the video map data 109 (step S104).

The detailed flow of the current position estimation process (step S104) is shown in FIG.

As shown in FIG. 6, the image collation unit 106 identifies the reference image corresponding to the current image based on the current image data acquired in step 101 and the video map data 109 acquired in step 103. (Step S111).

Specifically, the image collation unit 106 refers to the current image data acquired in step S101. Further, the image collation unit 106 refers to the video data 109b included in the video map data 109 acquired in step S103. Then, the image collation unit 106 identifies the reference image corresponding to the current image by collating the current image included in the current image data with the reference image included in the video data 109b.

Here, in the collation for specifying the reference image corresponding to the current image, for example, the contour of the feature is detected in each image of the current image and the reference image, and the similarity of the contour is sent out. The contour of the feature may employ a general technique of edge detection in an image, and is detected based on, for example, a change in RGB values in the image. Then, for example, the reference image having the highest similarity is specified as the reference image corresponding to the current image.

The estimation unit 107 estimates the current position of the automobile based on the shooting position information associated with the reference image identified in step S111 in the video map data 109 (step S112).

The output unit 108 outputs the current position data including the current position estimated in step S112 (step S113), and returns to the position estimation process shown in FIG.

As shown in FIG. 4, the estimation unit 107 and the output unit 108 estimate the road on which the automobile is moving based on the current image data and the video map data 109 (step S105).

The detailed flow of the moving road estimation process (step S105) is shown in FIG.

As shown in FIG. 7, the estimation unit 107 is moving with landmarks located in the vicinity based on the current image data acquired in step S101 and the video map data 109 acquired in step S103. The road is estimated (step S121).

Specifically, the estimation unit 107 refers to the current position of the automobile estimated in step S104 based on the current image data and the video map data 109. Here, the current position of the referenced automobile may be the approximate position indicated by the approximate position data acquired in step S102.

The estimation unit 107 estimates landmarks located around the referenced current position based on the map data 109a included in the video map data 109 acquired in step S103. The estimation unit 107 acquires the estimated landmark position information by referring to the map data 109a. Based on the acquired position information and the map data 109a, the estimation unit 107 estimates the road located around the estimated landmark as the road on which the automobile is traveling.

The output unit 108 outputs the moving road data indicating the moving road estimated by the estimation unit 107 (step S122), and returns to the position estimation process shown in FIG. In the moving road data, the road on which the automobile is moving is indicated by, for example, the road identification information included in the map data 109a. In this case, the moving road data includes estimated moving road identification information.

As shown in FIG. 4, the image collation unit 106, the estimation unit 107, and the output unit 108 estimate the moving direction of the automobile based on the current image data and the video map data 109 (step S106).

The detailed flow of the moving direction estimation process (step S106) is shown in FIG.

As shown in FIG. 8, the image collation unit 106 corresponds to a plurality of changes in the current image based on the current image data acquired in step 101 and the video map data 109 acquired in step 103. The reference image is specified (step S131).

Specifically, the image collation unit 106 refers to the current image data continuously acquired in step S101. Further, the image collation unit 106 refers to the video data 109b included in the video map data 109 acquired in step S103. Then, the image collation unit 106 identifies the reference image corresponding to the change of the plurality of current images by collating the change of the plurality of current images included in the current image data with the reference image included in the video data 109b. do.

Here, the changes in the plurality of current images are generally constant when moving in a certain direction on a certain road. Therefore, in the matching to identify the reference image corresponding to the change of the plurality of current images, the change of the plurality of current images can quantitatively expressed by vectorization. Therefore, the reference image corresponding to the change of the plurality of current images includes, for example, a vector indicating the change of the plurality of current images and a vector indicating the change of the reference image along the moving road estimated in step S105. By comparison, the reference image having the highest similarity between the two vectors is identified as the reference image corresponding to the changes in the plurality of current images.

The estimation unit 107 estimates the moving direction of the vehicle based on the shooting position information associated with the reference image identified in step S131 in the video map data 109 (step S132).

The output unit 108 outputs the movement direction data including the movement direction estimated in step S132 (step S133), and returns to the position estimation process shown in FIG. Then, the output unit 108 ends the position estimation process.

So far, Embodiment 1 of the present invention has been described.

According to the present embodiment, the reference image corresponding to the current image taken by the camera attached to the automobile is specified from the reference images taken in advance. Then, the current position of the automobile is estimated based on the shooting position of the specified reference image.

In this way, since the moving position of the automobile can be estimated without using a signal wave such as GPS, the current position of the moving body can be estimated without being affected by interferences on the ground. .. Therefore, it is possible to accurately estimate the current position of the moving body.

According to the present embodiment, from among the reference images taken in advance, the reference image corresponding to the change of a plurality of current images taken by the camera attached to the moving body is specified. Then, the moving direction of the automobile is estimated based on the shooting position of the specified reference image.

In this way, since the moving position of the automobile can be estimated without using a signal wave such as GPS, the moving direction of the moving body can be estimated without being affected by the interference on the ground. .. Therefore, it is possible to accurately estimate the moving direction of the moving body.

In the present embodiment, a predetermined feature is adopted as the landmark, and the landmark around the moving vehicle is estimated based on the information about the landmark, and the road on which the vehicle is moving is estimated. Is estimated. Therefore, it is possible to adopt a feature with a clear position as a landmark, such as a relatively large building or facility along the road, and accurately identify the moving road from the position of the landmark.

In the present embodiment, cameras 101 facing different directions such as front, back, left and right are provided, and the current position and the like are estimated based on the current images taken by each camera 101. This makes it possible to improve the accuracy of estimation as compared with the case of estimating the current position or the like based on the current image in only one direction.

In the present embodiment, the approximate position acquisition unit 103 is provided. As a result, the video map data 109 acquired from the video map storage unit 104 can be narrowed down, and the amount of data acquired from the video map storage unit 104 can be reduced. This makes it possible to shorten the processing time and reduce the processing load in processing such as collation.

(Modification example)
In the first embodiment, an example in which the position estimation device 100 includes the video map storage unit 104 and the video map acquisition unit 105 acquires the video map data 109 from the video map storage unit 104 has been described. However, as shown in FIG. 9, the position estimation device 200 does not have to include the video map storage unit 104. In this case, the video map acquisition unit 205 acquires the video map data 109 from, for example, a server device (not shown) via the network. This network is constructed by a wired line, a wireless line, or a line in which these are combined.

In the first embodiment, the position estimation device 100 needs to include a video map storage unit 104 having a storage capacity corresponding to the amount of data of the video map data 109. Therefore, when the amount of data of the video map data 109 is large, the size of the video map storage unit 104 may be increased and the cost may be increased. In this modification, since the position estimation device 200 does not have to include the video map storage unit 104, it is possible to reduce the size and cost of the position estimation device 200.

(Embodiment 2)
In the first embodiment, an example in which the current position of the automobile as a moving body is estimated based on the current image data and the video map data 109 has been described. In the second embodiment, the current position of the vehicle as a moving body is also estimated by autonomous navigation. Then, the position estimation device according to the second embodiment. The more probable current position is selected and output from the current position estimated based on the current image and the reference image and the current position estimated by autonomous navigation.

As shown in FIG. 10, the position estimation device 300 according to the second embodiment of the present invention includes an autonomous navigation unit 310 in addition to the configuration provided by the position estimation device 100 according to the first embodiment. Further, the position estimation device 300 includes an estimation unit 307 and an output unit 308 instead of the estimation unit 107 and the output unit 108 included in the position estimation device 100 according to the first embodiment. Regarding other points, the configuration of the position estimation device 300 is substantially the same as that of the position estimation device 100 according to the first embodiment.

The autonomous navigation unit 310 estimates the current position of the vehicle by autonomous navigation when the approximate position acquisition unit cannot acquire the approximate position data. In addition, the autonomous navigation unit 310 seeks the first degree of certainty. The first degree of certainty is a value indicating the certainty of the current position estimated by autonomous navigation.

Examples of cases where the approximate position acquisition unit 103 cannot acquire approximate position data include a case where an automobile is traveling in a tunnel and a case where GPS signals cannot be transmitted / received to / from a satellite due to radio interference. ..

The estimation unit 307 obtains a second degree of certainty in addition to the same function as the estimation unit 107 according to the first embodiment. The second degree of certainty is a value indicating the certainty of the current position estimated based on the video map data 109.

The output unit 308 has the same function as the estimation unit 107 according to the first embodiment.

The output unit 308 according to the present embodiment further compares the first certainty degree and the second certainty degree, selects a more probable current position based on the comparison result, and selects the selected current position. Output the current position data including. The current position is selected here from the current position estimated by autonomous navigation and the current position estimated based on the video map data 109.

That is, the output unit 308 determines that the current position estimated by autonomous navigation and the current position estimated based on the video map data 109 by comparing the first confirmation degree and the second confirmation degree. Of these, the more probable one outputs the current position data including the estimated current position.

So far, the configuration of the position estimation device 300 according to the second embodiment of the present invention has been described. From here, the operation of the position estimation device 300 according to the second embodiment of the present invention will be described.

The position estimation device 300 according to the present embodiment executes the position estimation process illustrated in FIG. The figure shows the flow of a portion of the position estimation process according to the present embodiment that is different from the position estimation process according to the first embodiment (see FIG. 4).

When the process of step S101 is executed, the approximate position acquisition unit 103 determines whether or not the approximate position data has been acquired (step S302).

When the approximate position data is acquired within the predetermined determination time, the approximate position acquisition unit 103 determines that the approximate position data has been acquired (step S302; YES). When it is determined that the approximate position data has been acquired (step S302; YES), the video map acquisition unit 105 executes the same process of step S103 as in the first embodiment. Then, the image collation unit 106, the estimation unit 107, and the output unit 108 execute the same process of step S104 as in the first embodiment.

If the approximate position data cannot be acquired within the predetermined determination time, the approximate position acquisition unit 103 determines that the approximate position data has not been acquired (step S302; NO).

When it is determined that the approximate position data has not been acquired (step S302; NO), the autonomous navigation unit 310 estimates the current position of the vehicle by autonomous navigation (step S341).

Specifically, for example, the autonomous navigation unit 310 acquires signals indicating the vehicle speed, acceleration direction, and magnitude of acceleration from each of the vehicle speed pulse generation circuit, the gyro sensor, and the acceleration sensor. The autonomous navigation unit 310 estimates the current position of the vehicle based on the acquired signal and the most recently estimated current position.

The autonomous navigation unit 310 asks the autonomous navigation unit 310 for the first degree of certainty (step S342). The first degree of certainty is calculated according to, for example, the time during which various signals can be acquired.

The video map acquisition unit 105 acquires the video map data 109 from the video map storage unit 104 (step S343). The video map data 109 acquired here is, for example, a predetermined range from the most recently estimated current position. The range here may be set wider than in the case of narrowing down based on the approximate position data in consideration of the fact that the moving body is moving.

The image collation unit 106, the estimation unit 307, and the output unit 308 estimate the current position of the automobile based on the current image data and the video map data 109 (step S344).

The detailed flow of the current position estimation process (step S344) is shown in FIG.

As shown in FIG. 12, the image collation unit 106 executes the same process of step S111 as in the first embodiment. The estimation unit 307 executes the process of step S112 similar to that of the first embodiment.

The estimation unit 307 obtains a second degree of certainty (step S351). In detail, for example, it is calculated according to the similarity corresponding to the reference image specified in step S111.

The output unit 308 compares the first confirmation degree and the second confirmation degree obtained in each of step S342 and step S351 (step S352).

The output unit 308 selects the more probable current position among the current positions estimated in steps S341 and S112 (included in step S344) based on the result of the comparison in step S352. Then, the output unit 308 outputs the current position data including the selected current position (step S353).
In detail, for example, the current position corresponding to the one with the higher degree of certainty is selected. That is, when the first degree of certainty is greater than the second degree of certainty, the current position estimated by autonomous navigation is selected. When the second confirmation degree is larger than the first confirmation degree, the current position estimated based on the current image data and the video map data 109 is selected.

If the first degree of certainty and the second degree of certainty are equal, for example, one of the predetermined current positions is selected.

When the process of step S104 or step S344 is completed, the process of step S105 similar to that of the first embodiment is executed.

So far, the second embodiment of the present invention has been described.

According to the present embodiment, when the current position (approximate position) estimated by GPS cannot be obtained, the current position is estimated by a method different from the collation of the current image and the reference image and the autonomous navigation. Then, based on each degree of certainty, the current position, which is more probable, that is, estimated to have good accuracy, is selected and output.

If the approximate position cannot be obtained, the reference image to be collated with the current image cannot be appropriately narrowed down, and as a result, the accuracy of estimating the current position by collating the current image and the reference image may decrease. In the present embodiment, when the approximate position cannot be obtained, the current position is estimated by a different method as described above, and a more probable current position is selected and output. Therefore, it is possible to suppress a decrease in the estimation accuracy of the current position when the approximate position cannot be obtained.

Here, an example of the physical configuration of the position estimation devices 100, 200, and 300 described so far is shown in FIG. As shown in the figure, the position estimation devices 100, 200, and 300 include cameras 101a to 101d and a processing unit 1001, which are mounted on the automobile 1002. As shown in the figure, the cameras 101a to 101d and the processing unit 1001 are connected by wirings 1003a to 1003d. It should be noted that each of the cameras 101a to 101d and the processing unit 1001 may be configured so as to be able to exchange image data at present, and may be connected wirelessly as well as by wire.

The processing unit 1001 includes a CPU (Control Processing Unit) 1004, a RAM (Random Access Memory) 1005, a ROM (Read Only Memory) 1006, a flash memory 1007, a communication I / F (interface) 1008, and a GPS I / F (GPS interface) 1009. , Connector 1110 and the like. Each unit 1004 to 1110 in the processing unit 1001 is connected by an internal bus 1111 and can transmit and receive data.

The communication I / F (interface) 1008 is an interface for communicating via a mobile phone communication network, a wireless LAN (Local Area Network), or the like. The GPS I / F (GPS interface) 1009 is an interface for communicating with GPS satellites and acquiring the approximate position of an automobile. The connector 1110 is a member to which the wirings 1003a to 1003d are connected.

Each function of the position estimation devices 100, 200, and 300 is, for example, by executing a software program (simply referred to as a "program") in which the CPU 1004 is pre-installed in the processing unit 1001 by using the RAM 1005 as a workspace. , It is realized in the whole processing unit 1001.

Specifically, for example, the functions of the current image acquisition unit 102, the approximate position acquisition unit 103, the video map acquisition unit 105, 205, the image collation unit 106, the estimation unit 107, 307, the output unit 108, 308, and the autonomous navigation unit 310 are described. It is realized by the CPU 1004 that executes the program as described above. Further, for example, the function of the video map storage unit 104 is realized by the flash memory 1007.

In the position estimation device 300 according to the second embodiment, in order to obtain a signal used by the autonomous navigation unit 310 for estimating the current position, a vehicle speed pulse generation circuit, a gyro sensor, an acceleration sensor, etc. of a vehicle (not shown) are further added to the automobile. It will be installed.

Although the embodiments and modifications of the present invention have been described above, the present invention is not limited thereto. For example, the present invention also includes a form in which a part or all of the embodiments and modifications described above are appropriately combined, and a form in which the form is appropriately modified.

Some or all of the above embodiments may also be described, but not limited to:
(Appendix 1)
The current image acquisition unit that acquires the current image data including the current image of the feature taken by the camera attached to the moving body, and the current image acquisition unit.
A video map acquisition unit that acquires video map data associated with a reference image including an image of a feature obtained by shooting in advance and shooting position information of the reference image.
An image matching unit that identifies a reference image corresponding to the current image by collating the current image of the acquired current image data with the reference image of the acquired video map data.
An estimation unit that estimates the current position of the moving body based on the shooting position information associated with the specified reference image in the video map data.
A position estimation device including an output unit that outputs current position data including the estimated current position.
(Appendix 2)
The image matching unit further identifies a reference image corresponding to the change in the current image by collating the change in the current image of the acquired current image data with the reference image in the acquired video map data.
The estimation unit further estimates the moving direction of the moving body based on the shooting position information associated with the specified reference image in the video map data.
The position estimation device according to Appendix 1, wherein the output unit further outputs movement direction data including the estimated movement direction.
(Appendix 3)
The video map data further includes location information regarding landmarks and roads that are predetermined features.
The estimation unit further estimates the landmarks located around the moving body and the road on which the moving body is moving, based on the acquired video map data and the acquired current image data. death,
The position estimation device according to Appendix 1 or 2, wherein the output unit further outputs moving road data indicating the estimated moving road.
(Appendix 4)
The moving body is a vehicle.
The current image acquisition unit is any one of Appendix 1 to 3, characterized in that the current image acquisition unit acquires current image data including a current image in which a feature is photographed by cameras attached to each of the front, rear, left and right sides of the moving body. The position estimation device described in 1.
(Appendix 5)
Further provided with an approximate position acquisition unit for acquiring approximate position data indicating an approximate position which is the current position of the moving body estimated by GPS.
The video map acquisition unit is any one of Appendix 1 to 4, characterized in that the video map acquisition unit acquires video map data including a reference image in a predetermined range with respect to the approximate position indicated by the acquired approximate position data. The position estimation device described in 1.
(Appendix 6)
When the approximate position acquisition unit cannot acquire the approximate position data, the current position of the moving body is estimated by autonomous navigation, and the first certainty indicating the certainty of the current position estimated by the autonomous navigation is obtained. With more parts
The estimation unit further obtains a second degree of certainty indicating the certainty of the current position estimated based on the video map data.
The output unit has a current position estimated by the autonomous navigation by comparing the first confirmation degree and the second confirmation degree, and a current position estimated based on the video map data. The position estimation device according to Appendix 5, wherein the more probable current position data including the estimated current position is output.
(Appendix 7)
The position estimation device according to any one of Supplementary note 1 to 6, further comprising at least one or more cameras.
(Appendix 8)
The position estimation device according to any one of Supplementary note 1 to 7, wherein the video map acquisition unit acquires the video map data from a server device via a network.
(Appendix 9)
A video map storage unit for storing the video map data is further provided.
The position estimation device according to any one of Supplementary note 1 to 8, wherein the video map acquisition unit acquires the video map data from the video map storage unit.
(Appendix 10)
Acquiring the current image data including the current image of the feature taken by the camera attached to the moving body, and
Acquiring video map data in which a reference image including an image of a feature obtained by shooting in advance and the shooting position information of the reference image are associated with each other.
Identifying the reference image corresponding to the current image by collating the current image of the acquired current image data with the reference image of the acquired video map data.
Estimating the current position of the moving body based on the shooting position information associated with the specified reference image in the video map data.
A position estimation method including outputting current position data including the estimated current position.
(Appendix 11)
Computer,
The current image acquisition unit, which acquires the current image data including the current image of the feature taken by the camera attached to the moving body.
A video map acquisition unit that acquires video map data associated with a reference image including an image of a feature obtained by shooting in advance and shooting position information of the reference image.
An image matching unit that identifies a reference image corresponding to the current image by collating the current image of the acquired current image data with the reference image of the acquired video map data.
An estimation unit that estimates the current position of the moving body based on the shooting position information associated with the specified reference image in the video map data.
An output unit that outputs current position data including the estimated current position,
A program to function as.

The present invention can be usefully used, for example, in a car navigation system, a device for managing the operation status of a vehicle, and the like.

100,200,300 Position estimation device 101,101a-101d Camera 102 Current image acquisition unit 103 Approximate position acquisition unit 104 Video map storage unit 105,205 Video map acquisition unit 106 Image matching unit 107 Estimating unit 108 Output unit 109 Video map data 109a Map data 109b Video data 310 Autonomous Navigation Department

Claims (7)

Four forward side image feature is captured by a camera mounted on the front, rear, left and right of the moving body, the rearward image, continuously the current image data currently contains an image composed of the left image and right image at regular time intervals The current image acquisition unit to be acquired and
An image for acquiring image map data in which a reference image including a front image, a rear image, a left image, and a right image of a feature obtained by taking a picture in advance and a shooting position information of the reference image are associated with each other. Map acquisition department and
An image matching unit that identifies a reference image corresponding to the current image by collating the current image of the acquired current image data with the reference image of the acquired video map data.
An estimation unit that estimates the current position of the moving body based on the shooting position information associated with the reference image corresponding to the specified current image in the video map data.
It is provided with an output unit that outputs current position data including the estimated current position.
The image collating unit further collates the vector indicating the change of the plurality of current images of the acquired current image data with the vector indicating the change of the reference image of the acquired video map data, thereby both vectors. The reference image having the highest degree of similarity is identified as the reference image corresponding to the change of the plurality of current images.
The estimation unit further estimates the moving direction of the moving body based on the shooting position information associated with the reference image corresponding to the change of the specified plurality of current images in the video map data.
The output unit is a position estimation device that further outputs movement direction data including the estimated movement direction. The video map data further includes location information regarding landmarks and roads that are predetermined features.
The estimation unit further estimates the landmarks located around the moving body and the road on which the moving body is moving, based on the acquired video map data and the acquired current image data. death,
The position estimation device according to claim 1, wherein the output unit further outputs moving road data indicating the estimated moving road. Further provided with an approximate position acquisition unit for acquiring approximate position data indicating an approximate position which is the current position of the moving body estimated by GPS.
The video map acquisition unit according to claim 1 or 2 , wherein the video map acquisition unit acquires video map data including a reference image in a predetermined range with respect to the approximate position indicated by the acquired approximate position data. Position estimation device. When the approximate position acquisition unit cannot acquire the approximate position data, the current position of the moving body is estimated by autonomous navigation, and the first degree of certainty indicating the certainty of the current position estimated by the autonomous navigation is obtained. With more navigation
The estimation unit further obtains a second degree of certainty indicating the certainty of the current position estimated based on the video map data.
The output unit has a current position estimated by the autonomous navigation by comparing the first confirmation degree and the second confirmation degree, a current position estimated based on the video map data, and a current position. The position estimation device according to claim 3 , wherein the more probable current position data including the estimated current position is output. The position estimation device according to any one of claims 1 to 4 , wherein the video map acquisition unit acquires the video map data from a server device via a network. Four forward side image feature is captured by a camera mounted on the front, rear, left and right of the moving body, the rearward image, continuously the current image data currently contains an image composed of the left image and right image at regular time intervals To get and
Acquiring video map data in which a reference image including a front image, a rear image, a left image, and a right image of a feature obtained by shooting in advance and a shooting position information of the reference image are associated with each other. When,
Identifying the reference image corresponding to the current image by collating the current image of the acquired current image data with the reference image of the acquired video map data.
To estimate the current position of the moving body based on the shooting position information associated with the reference image corresponding to the specified current image in the video map data.
Including outputting the current position data including the estimated current position.
In specifying the reference image, further, by collating the vector showing the change of the plurality of current images of the acquired current image data with the vector showing the change of the reference image of the acquired video map data. , The reference image having the highest similarity between the two vectors is specified as the reference image corresponding to the change of the plurality of current images.
In estimating the current position of the moving body, the moving body is further based on the shooting position information associated with the reference image corresponding to the change of the specified plurality of current images in the video map data. Estimate the direction of movement,
The position estimation method is characterized in that the output of the current position data further outputs the movement direction data including the estimated movement direction. Computer,
Four forward side image feature is captured by a camera mounted on the front, rear, left and right of the moving body, the rearward image, continuously the current image data currently contains an image composed of the left image and right image at regular time intervals Current image acquisition department to acquire,
An image for acquiring image map data in which a reference image including a front image, a rear image, a left image, and a right image of a feature obtained by taking a picture in advance and a shooting position information of the reference image are associated with each other. Map acquisition department,
An image matching unit that identifies a reference image corresponding to the current image by collating the current image of the acquired current image data with the reference image of the acquired video map data.
An estimation unit that estimates the current position of the moving body based on the shooting position information associated with the reference image corresponding to the specified current image in the video map data.
It functions as an output unit that outputs the current position data including the estimated current position.
The image collating unit further collates the vector indicating the change of the plurality of current images of the acquired current image data with the vector indicating the change of the reference image of the acquired video map data, thereby both vectors. The reference image having the highest degree of similarity is identified as the reference image corresponding to the change of the plurality of current images.
The estimation unit further estimates the moving direction of the moving body based on the shooting position information associated with the reference image corresponding to the change of the specified plurality of current images in the video map data.
The output unit is a program for further functioning to output movement direction data including the estimated movement direction.

Images