JP5353199B2 - Map information collection system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a map information collecting system for accurately collecting map information as much as possible. <P>SOLUTION: The map information collecting system A1 includes: an image recognition means 14 recognizing a prescribed object from a vehicle outside image of a plurality of viewpoints; a vector information calculation means 15 calculating the vector information of the movement of an object in the vehicle outside image corresponding to at least one viewpoint out of the plurality of viewpoints; an object location estimation means 16 estimating the location of the object on the basis of the vector information of the object and the traveling information/position information of a vehicle; an object appearance estimation means 17 estimating the appearance of the object in a prescribed range of the other vehicle outside image on the basis of the location of the object estimated from one vehicle outside image; a same object determination means 18 for determining that the object recognized in the one vehicle outside image and the other vehicle outside images is one and the same object when an object is recognized within a prescribed range in the other estimated vehicle outside image. <P>COPYRIGHT: (C)2010,JPO&amp;INPIT

Description

  The present invention relates to a map information collection system that collects map information by recognizing an image from a camera mounted on a vehicle.

  As this type of map information collection system, there is one disclosed in Patent Document 1. This map information collection system includes a camera mounted on a vehicle, image recognition means for recognizing a landmark as a predetermined object from an image photographed by the camera, and photographing for evaluating difficulty of the image recognition from various photographing conditions. Condition evaluation means, transmission history management means for confirming from the transmission history whether map information related to the landmark obtained as an image recognition result has been transmitted to the map center in a certain period in the past, landmark recognition score, photographing condition, and transmission There is a transmission condition determining means for determining whether or not to transmit the map information to the map center in consideration of the history. According to such a map information collection system, when the shooting conditions are good, there is no transmission history, and map information is acquired with a high recognition score, the map information can be transmitted as update information to the map center. . Thereby, the accuracy of update information can be improved and the number of transmissions can be reduced.

JP 2008-51612 A

  However, since the conventional map information collecting system basically recognizes only an image from one camera, it cannot collect so much map information related to landmarks due to limitations such as camera direction and viewpoint.

  To increase the amount of information collected, a plurality of cameras may be mounted on the vehicle, and landmarks may be image-recognized for images from the cameras. However, for example, the same landmark may be captured in a plurality of images captured from different viewpoints. In such a case, since the plurality of landmarks recognized in the images in the plurality of images cannot be determined as the same object, the map information related to the landmarks cannot be collected correctly, and the same map information is duplicated. There was a risk of being sent or registered.

  The present invention has been conceived under the circumstances described above, and an object thereof is to provide a map information collecting system capable of collecting as much and as accurately map information as possible.

  In order to solve the above problems, the present invention takes the following technical means.

  According to the first aspect of the present invention, a plurality of photographing means that are mounted on a vehicle and continuously capture the outside of the vehicle from a plurality of viewpoints, and traveling information on the traveling speed and traveling direction of the vehicle. Travel information acquisition means for sequentially acquiring position information acquisition means for sequentially acquiring position information relating to the ground position of the vehicle, images of a plurality of viewpoints obtained by the plurality of photographing means, and travel information and position of the vehicle A recording means for recording information in time series, an image recognition means for recognizing a predetermined object from the outside image of the plurality of viewpoints, and an outside image corresponding to at least one of the plurality of viewpoints. Vector information calculating means for calculating vector information about how the object moves in the image, vector information of the object, and travel information and position of the vehicle Object location estimating means for estimating the location of the object based on the information; for the outside image of the plurality of viewpoints, the location of the object estimated from one outside image, and the traveling information and position information of the vehicle; Based on the above, the object appearance predicting means for predicting that the object will appear in a predetermined range of another outside image, and the object is recognized in the predetermined range of the predicted other outside image. In this case, there is provided a map information collection system comprising: an identical object determination unit that determines that the object recognized in the one vehicle exterior image and the other vehicle exterior image is the same object.

  According to a second aspect of the present invention, a plurality of photographing means mounted on a vehicle, capturing the outside of the vehicle from a plurality of viewpoints and continuously photographing, and including at least a pair of photographing ranges, A plurality of viewpoints obtained by the traveling information acquisition means for sequentially acquiring the traveling information regarding the traveling speed and the traveling direction of the vehicle, the position information acquisition means for sequentially acquiring the position information regarding the ground position of the vehicle, and the plurality of photographing means. A recording means for recording the vehicle outside image and the travel information and position information of the vehicle in time series, an image recognition means for recognizing a predetermined object from the outside image of the plurality of viewpoints, and photographing among the outside images of the plurality of viewpoints Disparity information calculating means for calculating disparity information related to the difference in appearance of the object in the images outside the vehicle with reference to a set of images outside the vehicle with overlapping ranges; and the object Based on the parallax information and the travel information and position information of the vehicle, the object location estimation means for estimating the location of the object and the outside images of the multiple viewpoints were estimated from a set of outside images with overlapping shooting ranges. Based on the location of the object and the travel information and position information of the vehicle, the object appearance prediction means for predicting that the object will appear in a predetermined range of another outside image, and the predicted When the object is recognized within a predetermined range of another vehicle outside image, the same object determining means is provided that determines the object recognized in the one vehicle outside image and the other vehicle outside image as the same object. A map information collection system is provided.

  According to the third aspect of the present invention, a plurality of imaging means mounted on a vehicle and capturing continuously the outside of the vehicle from a plurality of viewpoints, and travel information on the travel speed and travel direction of the vehicle. Travel information acquisition means for sequentially acquiring position information acquisition means for sequentially acquiring position information relating to the ground position of the vehicle, images of a plurality of viewpoints obtained by the plurality of photographing means, and travel information and position of the vehicle A recording means for recording information in time series, an image recognition means for recognizing a predetermined object from the multiple-viewpoint exterior image, and an exterior image corresponding to at least one of the multiple-viewpoint exterior images. , Object positioning means for measuring the distance and direction to the object recognized in the image outside the vehicle, and positioning information obtained by measuring the distance and direction to the object And the object location estimating means for estimating the location of the object based on the travel information and the position information of the vehicle, and the location of the object estimated from a single exterior image and the location of the plurality of viewpoints Object appearance prediction means for predicting that the object will appear in a predetermined range of another outside image based on the travel information and position information of the vehicle, and a predetermined range of the predicted other outside image When the object is recognized, there is provided a map information collecting system comprising: an identical object judging means for judging that the object recognized in the one outside image and the other outside image is the same object. The

  According to the technology disclosed by the present invention, for example, a predetermined object is image-recognized from a plurality of viewpoints of an outside vehicle image, the location of the object is estimated from one outside vehicle image, and the appearance of the object in the other image is detected. Predict. Thereby, even if it is the target object image | photographed in the image outside a vehicle of a different viewpoint, since it can determine whether these target objects are the same objects, as much map information regarding this kind of target object as possible and It can be collected accurately.

  Other features and advantages of the present invention will become more apparent from the detailed description given below with reference to the accompanying drawings.

  Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the drawings.

  1 to 5 show an embodiment of a map information collecting system according to the present invention. The map information collection system A1 of the present embodiment is used in combination with a so-called navigation system B, and is installed in a vehicle (not shown) together with the navigation system B.

  The map information collection system A1 includes an A-view camera 1a, a B-view camera 1b, an image processing unit 10, a travel information acquisition unit 11, a position information acquisition unit 12, a recording unit 13, an image recognition unit 14, and a vector information calculation unit 15. , An object location estimation unit 16, an object appearance prediction unit 17, and an identical object determination unit 18. The navigation system B includes a navigation control unit 20, a navigation data storage unit 21, and a display unit 22. Image processing unit 10, travel information acquisition unit 11, position information acquisition unit 12, image recognition unit 14, vector information calculation unit 15, object location estimation unit 16, object appearance prediction unit 17, identical object determination unit 18, and navigation The control unit 20 is realized by a microcomputer including a CPU and a ROM. The recording unit 13 and the navigation data storage unit 21 are realized by a hard disk device, for example. The display unit 22 is realized by a liquid crystal display device, for example.

  As shown in FIG. 2, the A viewpoint camera 1a is set with a viewpoint facing the front of the vehicle, and continuously captures the front of the vehicle at predetermined time intervals. An image taken by the A-viewpoint camera 1a is referred to as an A-viewpoint outside-vehicle image. The B viewpoint camera 1b has a viewpoint that faces the left side of the vehicle, for example, and continuously captures the left side of the vehicle at predetermined time intervals. An image photographed by the B-viewpoint camera 1b is referred to as a B-viewpoint outside-vehicle image. The B-viewpoint camera 1b can switch shooting modes such as a standard image quality mode and a high image quality mode. The A-viewpoint and B-viewpoint external images have image contents as illustrated in FIG. 3 as an example.

  The image processing unit 10 performs various types of image processing on the outside images from the A viewpoint and B viewpoint cameras 1 a and 1 b, and transfers these outside images to the recording unit 13. The image processing unit 10 has a function as shooting mode switching means for switching and controlling the shooting mode of the B-viewpoint camera 1b.

  The travel information acquisition unit 11 sequentially acquires travel information regarding the travel speed and the travel direction of the vehicle at predetermined time intervals, and transfers the travel information to the recording unit 13. The traveling speed is measured by a speedometer installed in the vehicle, and the traveling direction is detected by a gyro sensor installed in the vehicle.

  The position information acquisition unit 12 sequentially acquires position information regarding the ground position of the vehicle at predetermined time intervals, and transfers this position information to the recording unit 13. The position information is created by receiving a GPS signal from a GPS satellite.

  In the recording unit 13, the A-viewpoint and B-viewpoint outside-vehicle images, vehicle travel information, and position information are recorded in association with each other in time series. As illustrated in FIG. 4 as an example, the outside image of the A viewpoint and the B viewpoint is recorded in order of time passage from top to bottom. For example, when traveling at a speed of 30 km in the east direction, the traveling information is recorded as “E30”. The location information is recorded as “E134.56.4.0 N34.41.3.3.3” by, for example, latitude and longitude. Location information to be described later is also recorded by, for example, latitude and longitude.

  The image recognition unit 14 recognizes, for example, a static object such as a building, a store, or a road from the images outside the A viewpoint and the B viewpoint by image recognition processing. In this image recognition process, for example, the contour or pattern shape of a characteristic object is extracted from the image outside the vehicle, and the object is identified by performing pattern matching with the object model stored in the database in advance. To do. In such an image recognition process, for example, a sign or a sign can be recognized, and not only image information of the object but also text information can be obtained incidentally.

  The vector information calculation unit 15 refers to a plurality of continuous A-view images outside the vehicle, and calculates vector information regarding how the object moves in these out-of-vehicle images. Specifically, as shown in FIG. 5, the vector information calculation unit 15 performs relative processing in which the target object appears to be apparently shifted between a plurality of images having a predetermined time interval after the target object is image-recognized. The amount of movement and its direction are calculated as vector information. The vector information is as shown by a white arrow in the figure. For example, if the arrow indicating the vector information is large, it can be seen that the object exists at a relatively close distance from the vehicle. If the arrow is small, the object exists at a relatively far distance from the vehicle. From such vector information, the approximate distance and direction to the object with the vehicle position as the reference position can be determined.

  The object location estimation unit 16 indicates that the object actually exists based on the vector information of the object and the traveling information and position information of the vehicle corresponding to the outside image of the A viewpoint from which the vector information is calculated. Estimate the location on the waxy ground. Such location estimation is performed by performing predetermined calculations on the vector information, the travel information, and the position information, and determining the distance and direction from the vehicle whose travel position is known to the object. The location information obtained as the estimation result is recorded in the recording unit 13 in association with the corresponding outside image of the A viewpoint. Further, when the location information is obtained, the flag “#F” is set in the outside image of the viewpoint A based on the location information.

  The object appearance prediction unit 17 uses the same object as the image at a certain point of the B-view outside image based on the location information of the object obtained from the A-view image and the travel information and position information of the vehicle. Predict that an object will appear. Prediction of the appearance of such an object is performed by performing predetermined calculations on location information, travel information, and position information, and when the currently traveling vehicle travels in a certain direction for a predetermined time, the target is captured in the shooting range of the B viewpoint camera 1b. This is done by predicting that an object will come in. As a prediction result, a predicted elapsed time from the time when the image of the A viewpoint outside the vehicle with the flag “#F” is set is obtained. When the predicted elapsed time has passed, a high image quality mode switching signal is transmitted from the object appearance prediction unit 17 to the image processing unit 10. The image processing unit 10 that has received the high image quality mode switching signal switches the shooting mode of the B-viewpoint camera 1b from the standard image quality mode to the high image quality mode. The high image quality mode is continued for, for example, a predetermined time, and after the predetermined time, the shooting mode of the B-viewpoint camera 1b is switched to the standard image quality mode again.

  If the object is recognized in the predicted B-view vehicle outside image, the same object determination unit 18 is the same as the object recognized in the A-view vehicle image based on the prediction. Determine whether or not. When it is determined that the objects are the same as the determination result, in addition to the location information regarding the object, information indicating the corresponding image and name is transferred to the navigation control unit 20 as map information. When it is determined that the objects are not the same as the determination result, it is determined that the object photographed with the A-viewpoint outside image and the object photographed with the B-viewpoint exterior image are different from each other, and map information regarding each object Is transferred to the navigation control unit 20.

  The navigation control unit 20 reads, for example, navigation data related to a map image for display from the navigation data storage unit 21, and based on this navigation data and vehicle travel information and position information, a surrounding map and the vehicle mark that the vehicle is currently traveling are read. Is displayed on the display unit 22. When map information is transferred from the map information collection system A1, the navigation control unit 20 updates navigation data based on the map information. For example, when an object such as a newly constructed building or store is image-recognized and map information is transferred, a landmark relating to the object is reflected in the navigation data, and the display unit 22 has a land that has not existed before. The mark is displayed in the surrounding map.

  Next, the operation procedure of the map information collection system A1 will be described with reference to FIG.

  First, the image processing unit 10 acquires images outside the vehicle from the cameras 1a and 1b of the A viewpoint and the B viewpoint (S1).

  Simultaneously with the acquisition of the outside image, the travel information acquisition unit 11 and the position information acquisition unit 12 acquire the travel information and position information of the vehicle (S2, S3).

  The recording unit 13 records, for example, a time-stamp indicating acquisition date and time, an A-side image and a B-viewpoint outside-vehicle image, travel information, and position information (S4).

  Next, the image recognizing unit 14 performs image recognition processing on the recorded A-viewpoint outside-vehicle image (S5). As a result, a predetermined object that matches the object model is recognized from the outside image of the A viewpoint. Note that the image recognition process may be performed on the A-side image and the B-side image. Further, it may be performed simultaneously with the recording of the outside image.

  Next, the vector information calculation unit 15 calculates the vector information of the object from a plurality of continuous images of the A viewpoint outside the vehicle (S6).

  Next, the target object location estimation unit 16 estimates the location of the target object based on the calculated vector information and the travel information and position information of the vehicle corresponding to the vehicle outside image of the viewpoint A based on the calculated vector information (S7). . At this time, the location information of the object is recorded in the recording unit 13 in association with the corresponding outside image of the A viewpoint, and the flag “#F” is set on the outside image of the A viewpoint serving as a reference. Note that the process of estimating the location of the object may be performed on the object recognized in the B-viewpoint outside vehicle image by calculating the vector information of the object from the B-viewpoint outside-vehicle image. In this case, the location information of the object is recorded in association with the B-viewpoint outside-vehicle image.

  Next, the object appearance prediction unit 17 is the same as the vehicle-outside image at the B viewpoint after a predetermined time has elapsed from the vehicle-outside image with the flag “#F” based on the location information of the object and the travel information and position information of the vehicle. It is predicted that an object will appear (S8). As the prediction result, the predicted elapsed time from the time stamp corresponding to the flag “#F” is calculated. Such location estimation and appearance prediction of the object are executed in real time simultaneously with the recording of the outside image.

  When the predicted elapsed time is reached (S9: YES), the image processing unit 10 switches the shooting mode of the B-viewpoint camera 1b to the high image quality mode (S10). As a result, a high-quality vehicle exterior image can be obtained from the B-viewpoint camera 1b. Until the predicted time point comes (S9: NO), the process returns to S1 and the processes up to S8 are repeatedly executed. During the processing from S1 to S8, a B-viewpoint outside-vehicle image having standard image quality is acquired. When the process is repeated after returning to S1, the object recognition process may be performed on the B-viewpoint outside-vehicle image, and the location information of the object may be recorded.

  Thereafter, the image recognition unit 14 performs image recognition processing on the recorded B-viewpoint outside-vehicle image (S11). At this time, the in-vehicle image of the B viewpoint at the time of prediction is recorded with high image quality by switching the shooting mode. Therefore, the recognition rate of the object by the image recognition process is increased particularly for the B viewpoint external image at the time of prediction.

  Next, the same object determination unit 18 determines whether or not the object recognized in the B-view outside image at the prediction time point is the same as the object recognized in the A-view outside image (S12). ). As the determination process, it is easily determined based on whether or not the types and names of the identified objects match.

  When it is determined that the objects are the same as the determination result (S12: YES), the same object determination unit 18 displays the location information recorded for the object and information indicating the image and name of the corresponding object model as map information. To the navigation control unit 20 (S13).

  In S12, when it is determined that the object recognized in the A-side image and the object recognized in the B-side image are different (S12: NO), the same object determination unit 18 Map information relating to the object is transferred to the navigation control unit 20 (S14). Such a series of processing from S1 to S14 is repeatedly performed.

  Therefore, according to the map information collection system A1 of the present embodiment, the identity of the object recognized by the A-side image and the object recognized by the B-side image is determined, and only correct map information is obtained. Since it can be selected and transferred to the navigation system B, it is possible to collect as much and as accurately map information about this type of object as possible.

  7 to 14 show other embodiments of the map information collecting system according to the present invention. Note that the same or similar components and various processes as those according to the above-described embodiment are denoted by the same reference numerals and description thereof is omitted.

  As shown in FIG. 8, the map information collection system A2 in FIG. 7 is provided with an A ′ viewpoint camera 1a ′ so that the shooting range overlaps with the A viewpoint camera 1a. In addition, the map information collection system A2 is provided with a parallax information calculation unit 15 'that calculates parallax information in order to estimate the location of the object from the images outside the vehicle at the A viewpoint and the A' viewpoint.

  The A ′ viewpoint outside vehicle image captured by the A ′ viewpoint camera 1a ′ has an image content as indicated by a broken line in FIG. 9, and is slightly shifted from the A viewpoint vehicle outside image. The outside image of the A ′ viewpoint is also recorded in the recording unit 13 in association with the travel information and position information of the vehicle.

  The disparity information calculation unit 15 ′ refers to the A viewpoint and the outside image of the A ′ viewpoint, and calculates disparity information related to the appearance of the object recognized from these outside images. Specifically, the parallax information calculation unit 15 ′ calculates the relative shift amount and the direction in which the target object appears to be shifted between these images after the target object is image-recognized from the A viewpoint and the outside image of the A ′ viewpoint. Calculated as parallax information. This disparity information is similar to the vector information in the above-described embodiment, and is obtained as vector data. For example, when the vector value of the parallax information is large, it can be seen that the object is present at a relatively close distance from the vehicle, and when the vector value is small, the object is present at a relatively distant distance from the vehicle. From the parallax information as the vector data, the approximate distance and direction to the object with the vehicle position as the reference position can be determined. Based on the parallax information of the target object and the traveling information and the position information of the vehicle corresponding to the outside image of the A viewpoint and the A ′ viewpoint that are the calculation target of the parallax information, the target object location estimation unit 16 Estimate the ground location where the object will actually exist.

  The operation procedure of the map information collection system A2 will be described with reference to FIG.

  First, the image processing unit 10 acquires images outside the vehicle from the A viewpoint, the A ′ viewpoint, and the B viewpoint cameras 1a, 1a ′, and 1b (S21). Thereafter, similarly to the processes from S2 to S4 in FIG. 6, the processes from S22 to S24 are executed.

  Next, the image recognition unit 14 performs an image recognition process on the recorded A-viewpoint and A′-viewpoint external images (S25). Thus, a predetermined object that matches the object model is recognized from the images outside the A viewpoint and the A ′ viewpoint.

  Next, the parallax information calculation unit 15 calculates the parallax information of the target object from the outside images of the A viewpoint and the A ′ viewpoint in which the target object is recognized (S26).

  Next, the object location estimating unit 16 determines the location of the object based on the calculated disparity information and the travel information and position information of the vehicle corresponding to the outside image of the A viewpoint and the A ′ viewpoint that are the calculation targets. Is estimated (S27). Thereafter, similarly to the processing from S8 to S14 in FIG. 6, the processing from S28 to S34 is executed.

  Therefore, the map information collection system A2 of the present embodiment can collect as much and as accurately the map information regarding the object as possible according to the embodiment described above. In addition, when estimating the location of an object, it is only necessary to refer to a set of images outside the vehicle taken at an A viewpoint and an A ′ viewpoint at a certain moment. Appearance prediction of an object can be performed and a series of processing can be performed smoothly.

  As shown in FIG. 12, the map information collecting system A3 in FIG. 11 is provided with a radar 1c that transmits a radar wave toward the photographing range of the camera 1a at the A viewpoint and receives the reflected wave. Further, the map information collection system A3 is provided with an object positioning unit 19 that measures the distance and direction to the object in accordance with a signal from the radar 1c. As shown in FIG. 12 and FIG. 13, the radar 1 c oscillates a radar wave within a predetermined range, thereby measuring not only the distance to the object reflected by the radar wave but also the direction. The

  The object positioning unit 19 refers to the outside image of the A viewpoint, and when the direction of the object recognized in the outside image matches the direction based on the signal from the radar 1c, the object positioning unit 19 detects the object based on the signal from the radar 1c. Create the distance and direction to the object as positioning information. This positioning information is used as it is for the location estimation of the object, and the object location estimation unit 16 is based on the positioning information obtained by the radar 1c and the traveling information and position information of the vehicle. Estimate the location.

  The operation procedure of the map information collection system A3 will be described with reference to FIG.

  First, as map information collection processing, the processing of S41-S45 is performed similarly to the processing of S1-S5 of FIG.

  Thereafter, the object positioning unit 19 measures the distance and direction to the object based on the signal from the radar 1c with respect to the object recognized by the A-viewpoint outside vehicle image (S46). The distance and direction to the target object thus measured are transmitted to the target object location estimating unit 16 as positioning information.

  Next, the target object location estimating unit 16 estimates the location of the target object based on the positioning information and the travel information and position information corresponding to the measurement time (S47). Thereafter, similarly to the processes from S8 to S14 in FIG. 6, the processes from S48 to S54 are executed.

  Therefore, the map information collection system A3 of the present embodiment can collect as much and as accurately map information about the object as possible according to the embodiment described above. Further, since the location of the object is immediately estimated based on the A-view outside image taken at a certain moment and the signal from the radar 1c, the acquisition of the location information and the prediction of the appearance of the object are almost at the same time. And a series of processes can be executed smoothly.

  In addition, this invention is not limited to said embodiment.

  The configuration shown in the above embodiment is merely an example, and the design can be changed as appropriate according to the specification.

  For example, three or more cameras that photograph outside the vehicle may be provided. As the number of cameras installed increases, more map information about the object can be collected.

  As the map information collection system, for example, a function mounted on a vehicle and an information processing function for performing image recognition or the like may be separately provided, and this information processing function may be provided to a computer in a data center. Examples of the information processing function include an image recognition unit, a vector information calculation unit, a parallax information calculation unit, an object location estimation unit, an object appearance prediction unit, and an identical object determination unit. In this case, the shooting mode cannot be switched, but once the vehicle outside image is recorded, image recognition processing can be performed with high accuracy by a general-purpose computer or the like having high computing ability, and the map information of the object can be obtained. More and more can be acquired.

It is a block diagram which shows one Embodiment of the map information collection system which concerns on this invention. It is a schematic diagram which shows the arrangement | positioning form of the camera of A viewpoint and B viewpoint. It is explanatory drawing for demonstrating the image outside a vehicle of A viewpoint and B viewpoint. It is explanatory drawing for demonstrating the recording structure of an image and information. It is explanatory drawing for demonstrating the calculation method of vector information. It is a flowchart which shows the operation | movement procedure by the map information collection system of FIG. It is a block diagram which shows other embodiment of the map information collection system which concerns on this invention. It is a schematic diagram which shows the arrangement | positioning form of the camera of A viewpoint, A 'viewpoint, and B viewpoint. It is explanatory drawing for demonstrating the vehicle outside image of A viewpoint and A 'viewpoint. It is a flowchart which shows the operation | movement procedure by the map information collection system of FIG. It is a block diagram which shows other embodiment of the map information collection system which concerns on this invention. It is a schematic diagram which shows the arrangement | positioning form of the camera of A viewpoint and B viewpoint, and a radar. It is explanatory drawing for demonstrating the positioning method about a target object. It is a flowchart which shows the operation | movement procedure by the map information collection system of FIG.

Explanation of symbols

1a A viewpoint camera (photographing means)
1a 'A' viewpoint camera (photographing means)
1b B viewpoint camera (photographing means)
1c Radar 10 Image processing unit (shooting mode switching means)
11 Travel information acquisition unit (travel information acquisition means)
12 position information acquisition unit (position information acquisition means)
13 Recording section (recording means)
14 Image recognition unit (image recognition means)
15 Vector information calculation unit (vector information calculation means)
15 'parallax information calculation unit (parallax information calculation means)
16 Target location estimation part (Target location estimation means)
17 Object appearance prediction part (object appearance prediction means)
18 Same thing judgment part (same thing judgment means)
19 Object positioning unit (object positioning means)

Claims (4)

A plurality of photographing means mounted on the vehicle and capturing the outside of the vehicle from a plurality of viewpoints continuously;
Travel information acquisition means for sequentially acquiring travel information relating to the travel speed and travel direction of the vehicle;
Position information acquisition means for sequentially acquiring position information relating to the ground position of the vehicle;
Recording means for recording images of a plurality of viewpoints obtained by the plurality of photographing means and vehicle travel information and position information in time series;
Image recognizing means for recognizing a predetermined object from the plurality of viewpoints of the vehicle exterior image;
A vector information calculating unit that refers to an outside image corresponding to at least one of the plurality of viewpoints, and calculates vector information about how the object moves in the outside image;
Object location estimating means for estimating the location of the object based on vector information of the object and travel information and position information of the vehicle;
The above-mentioned object appears in a predetermined range of another outside image based on the location of the object estimated from one outside-of-vehicle image and the travel information and position information of the vehicle. An object appearance prediction means for predicting waxing;
When the object is recognized in a predetermined range of the predicted other outside image, the same object determining means for determining the object recognized in the one outside image and the other outside image as the same object;
A map information collection system. A plurality of photographing means mounted on the vehicle, capturing the outside of the vehicle from a plurality of viewpoints and continuously photographing, and including at least a pair of photographing ranges,
Travel information acquisition means for sequentially acquiring travel information relating to the travel speed and travel direction of the vehicle;
Position information acquisition means for sequentially acquiring position information relating to the ground position of the vehicle;
Recording means for recording images of a plurality of viewpoints obtained by the plurality of photographing means and vehicle travel information and position information in time series;
Image recognizing means for recognizing a predetermined object from the plurality of viewpoints of the vehicle exterior image;
Parallax information calculation means for calculating parallax information related to the difference in appearance of the object in the outside images with reference to the outside images of a set of imaging ranges that overlap among the outside images of the plurality of viewpoints;
Object location estimating means for estimating the location of the object based on the parallax information of the object and travel information and position information of the vehicle;
Based on the location of the object estimated from the images outside the vehicle with overlapping shooting ranges and the travel information and position information of the vehicle, the object within the predetermined range of the other images Object appearance prediction means for predicting that
When the object is recognized in a predetermined range of the predicted other outside image, the same object determining means for determining the object recognized in the one outside image and the other outside image as the same object;
A map information collection system. A plurality of photographing means mounted on the vehicle and capturing the outside of the vehicle from a plurality of viewpoints continuously;
Travel information acquisition means for sequentially acquiring travel information relating to the travel speed and travel direction of the vehicle;
Position information acquisition means for sequentially acquiring position information relating to the ground position of the vehicle;
Recording means for recording images of a plurality of viewpoints obtained by the plurality of photographing means and vehicle travel information and position information in time series;
Image recognizing means for recognizing a predetermined object from the plurality of viewpoints of the vehicle exterior image;
An object positioning unit that refers to an image outside the vehicle corresponding to at least one viewpoint among the images outside the vehicle of the plurality of viewpoints, and measures a distance and a direction to the object recognized in the image outside the vehicle;
An object location estimating means for estimating the location of the object based on positioning information obtained by measuring the distance and direction to the object, and the traveling information and position information of the vehicle;
The above-mentioned object appears in a predetermined range of another outside image based on the location of the object estimated from one outside-of-vehicle image and the travel information and position information of the vehicle. An object appearance prediction means for predicting waxing;
When the object is recognized in a predetermined range of the predicted other outside image, the same object determining means for determining the object recognized in the one outside image and the other outside image as the same object;
A map information collection system. At the same time that the plurality of viewpoints outside the vehicle image are captured from the plurality of photographing means to the recording means, the object location estimating means and the object appearance predicting means are executing predetermined operations in real time,
When the time estimated by the object location estimating means reaches the time predicted by the object appearance predicting means, the other outside vehicle image is taken with higher image quality than the image quality before the predicted time. The map information collecting system according to claim 1, further comprising photographing mode switching means for switching a photographing mode of the photographing means.

Images