JP3333223B2 - Roadway recognition device for mobile vehicles - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus for recognizing a traveling path of a moving vehicle, and more particularly, to inputting an external environment of the moving vehicle as an image and recognizing a traveling path shape, a vehicle position, a vehicle posture, and the like based on the image. To a device for doing so.

[0002]

2. Description of the Related Art In an autonomous traveling vehicle, it is very important to accurately recognize the shape of a traveling road, the position of a vehicle, the posture of a vehicle, and the like on a traveling road that is currently traveling. When performing such travel road recognition, a road image is usually input by a camera or the like, and feature amounts such as road lane lines are extracted based on the road image, and the contents of a knowledge base prepared in advance are extracted. With reference to this feature amount, the process of recognizing the traveling road is performed. The knowledge base includes information on the conditions of image input, such as camera height, installation angle, zoom magnification, etc., information on the road structure such as the number of lanes and lane width of the running road, and the current vehicle speed. Information such as weather and weather is prepared. By performing geometric calculation processing between these pieces of information and the extracted feature values, it is possible to recognize the shape of the traveling road, the vehicle position, the vehicle attitude, and the like. For example, Japanese Patent Application Laid-Open No. 64-26913 discloses that
A traveling path recognition device for an autonomous traveling vehicle based on such a principle is disclosed.

[0003]

As described above, the process of recognizing a traveling road is performed by a geometric calculation process based on features extracted from a road image and information in a knowledge base. A complicated geometric operation generally requires a considerably long operation time. For this reason, the conventional travel path recognition device has a problem that the recognition time is long.

Accordingly, an object of the present invention is to provide an apparatus for recognizing a traveling path of a mobile vehicle, which can perform recognition processing at a higher speed.

[0005]

Means for Solving the Problems (1) In a first aspect of the present invention, in a traveling road recognition device for a moving vehicle, an image showing an environment outside the moving vehicle is input and features of the traveling road are shown from the image. A feature amount extracting means for extracting a feature amount, a travel path recognizing means for recognizing a shape of the travel route based on the feature amount, and a first information holding knowledge necessary for recognizing the shape of the travel route . A knowledge base, and a second knowledge base in which recognition results regarding the shape of the traveling path corresponding to the specific feature amount are registered. 2 is searched, and when a feature amount similar to the extracted feature amount is registered, the recognition result of the traveling path corresponding to the similar feature amount is output as it is, and a similarity to the extracted feature amount is output. When the feature value to be registered is not registered , A recognition result is obtained by referring to the knowledge held in the first knowledge base, the recognition result is output, and a process of newly registering the recognition result in the second knowledge base is performed. Things. (2) The second invention of the present application is a moving device according to the first invention.
In the apparatus for recognizing a traveling road of a vehicle, a feature amount extracting unit extracts a line candidate point constituting a road lane line based on the input road image and a road image input unit for inputting a road image of the traveling road. And a line candidate point extracting unit, and a set of extracted line candidate points is used as a feature amount. (3) The third invention of the present application is the first or second invention described above.
In the travel path recognition device in a mobile vehicle according to the traveling path recognized hand
At the bottom, recognition of the vehicle position as well as the shape of the travel path is performed.
And the first knowledge base contains the vehicle position
Also have the necessary knowledge to recognize
In addition, the second knowledge base contains recognition results regarding the vehicle position.
The result is also registered. (4) The fourth invention of the present application relates to the first to third inventions.
Travel path recognition means for a traveling vehicle
To recognize the attitude of the vehicle as well as the shape of the roadway
Further functions are provided, and the first knowledge base contains information on vehicle attitude.
To have the knowledge necessary to perform
Then, the second knowledge base contains the recognition results regarding the vehicle attitude.
Is also registered.

[0006]

[Operation] On a traveling road such as an expressway where the road structure does not change much, it is highly possible that a pattern similar to the combination of the feature value extracted from the road image and the recognition result of the traveling road appears. For example, it is assumed that a recognition process of a traveling road is performed based on a certain feature amount A, and a recognition result B is obtained. Subsequently, assuming that a feature A ′ similar to the feature A is obtained, a recognition result B ′ obtained by performing a recognition process based on the feature A ′ is often similar to the recognition result B. . Therefore, when a recognition result B is obtained by performing a recognition process on the feature amount A using the first knowledge base, a combination of the feature amount A and the recognition result B is registered in the second knowledge base at that time. To Then, when the feature amount A ′ is obtained later, a similar feature amount A is searched from the second knowledge base, and the recognition result B registered in combination with the feature amount A is extracted as the feature amount A. There is no problem if it is used as a recognition result for '.

The present invention pays attention to such a point, and when it can be approximated by past recognition results registered in the second knowledge base, the traveling path recognition processing (complex geometrical processing) using the first knowledge base is performed. In this case, an approximate recognition result in the second knowledge base is used as it is, without performing a time-consuming operation due to the necessity of a mathematical operation. Also, in the knowledge base
Using the knowledge and the features of the running path extracted during running,
Various recognitions about the route can be performed.
You.

[0008]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiments according to the present invention will be described below based on embodiments illustrating the present invention. FIG. 1 is a block diagram showing a basic configuration of a traveling road recognition device for a mobile vehicle according to one embodiment of the present invention. This device is composed of a feature amount extracting means 10, a traveling road recognizing means 20, a global knowledge base 30, and a local knowledge base 40. The feature amount extracting means 10 has a function of inputting an image indicating the environment of the outside of the moving vehicle and extracting a feature amount indicating the feature of the traveling road from the image. In this embodiment,
It comprises a road image input unit 11 and a line candidate point extracting unit 12. The road image input unit 11 is, specifically, a camera attached to a moving vehicle, and a road image ahead of the traveling road is captured by the camera. The line candidate point extracting unit 12 performs a process of extracting a line candidate point forming a road lane line based on the road image. For example, if the road lane line is constituted by a white line, a pixel having a density value corresponding to the white line is extracted as a line candidate point on the road image.
Theoretically, a line connecting these line candidate points is a road lane. In this embodiment, the set W of the line candidate points is handled as a feature amount extracted from the road image.

The travel path recognizing means 20 calculates the candidate point set W
Has the function of obtaining the travel path shape X, the vehicle position Y, and the vehicle attitude Z based on the The task of recognizing the traveling path, which is the purpose of this apparatus, is nothing less than the task of obtaining information related to the traveling path such as the traveling path shape X, the vehicle position Y, and the vehicle attitude Z. In the apparatus according to the present invention, two methods are prepared as a method for performing such recognition by the traveling road recognition means 20. The first method is a method performed with reference to the global knowledge base 30, and the second method is a method performed with reference to the local knowledge base 40.

[0010] The global knowledge base 30 holds knowledge necessary for recognizing a traveling route.
In this embodiment, specifically, for the camera used in the road image input unit 11, information such as the mounting height, the installation angle, and the zoom magnification, the number of lanes on the road on which the vehicle is traveling,
Information such as lane width, and information such as current vehicle speed and weather are prepared. These pieces of information are listed as an example. In short, any information may be prepared in the global knowledge base 30 as long as the information can be used for the recognition of the traveling path. In the first recognition method, the travel path is recognized based on the feature amount extracted by the feature amount extraction unit 10, that is, information on the candidate point set W and information prepared in the global knowledge base 30. . As described above, since such a recognition process is a geometric calculation process, the calculation time is long. In addition, since various methods are known about a specific recognition processing method, description is omitted here. When such a recognition process is performed, a travel path shape X, a vehicle position Y, and a vehicle posture Z are output as recognition results. The feature of the present invention is that the recognition result obtained by referring to the global knowledge base 30 is
The point is to register.

In the local knowledge base 40, a recognition result of a traveling path corresponding to a specific feature amount (that is, a candidate point set W) is registered. As described above, each time a recognition process that refers to the global knowledge base 30 is performed, the recognition result is registered in the local knowledge base 40.
This registration is performed by associating a specific feature with a recognition result for the feature. In the example shown in the local knowledge base 40 of FIG. 1, the recognition results Xi, Yi, Zi are registered for the candidate point set Wi (i =
1, 2, 3, ...). This means that when the candidate point set Wi was obtained in the past, as a result of performing a geometric operation with reference to the global knowledge base 30, recognition results Xi, Yi, and Zi were obtained. are doing.

The second recognition method performed by the travel path recognition means 20 is a method of searching for a similar feature amount from the local knowledge base 40 and diverting the recognition result registered for the searched feature amount as it is. . For example, assuming that the candidate point set Wk is obtained from the feature amount extracting means 10, the local knowledge base 40 is searched for a candidate point set similar to the candidate point set Wk. If W3 is similar to Wk, X3, Y3, Z3, which are the recognition results registered for W3,
As the recognition result for the candidate point set Wk. Such a recognition method is merely an operation of searching a database, and a recognition result can be obtained in a very short time.

A feature of the travel path recognition apparatus according to the present invention resides in that a first recognition method using the global knowledge base 30 and a second recognition method using the local knowledge base 40 are used together. When the candidate point set Wk is obtained from the feature amount extraction means 10, the travel path recognition means 20 first tries the second recognition method with priority. That is, a search similar to the candidate point set Wk is searched from the local knowledge base 40.
If a similar thing is found, the previously registered recognition result is output as it is as the recognition result of the candidate point set Wk. If a similar one is not registered, the first recognition method calculates the recognition results Xk, Yk, and Zk, outputs them, and outputs Wk → Xk, Yk,
The corresponding relationship Zk is newly registered in the local knowledge base 40.

By employing such a method, when a candidate point set similar to the past is obtained, the second recognition method can be applied, and the time-consuming first recognition method can be minimized. . It should be noted that as a process of searching for a similar candidate point set from the local knowledge base 40 with respect to the candidate point set Wk, it is preferable to perform a similarity determination allowing some ambiguity. For example, production rules, fuzzy rules, neural nets,
By performing a similarity determination using the above, a highly flexible determination can be made. In addition, information on an error between the candidate point set Wk and a candidate point set determined to be similar to the candidate point set (for example, W3) is accumulated, and the local knowledge base 40 is appropriately modified based on the information on the error. It is also possible to perform various processing. Since the information registered in the local knowledge base 40 gradually increases, if there is a limit on the memory capacity or the like, the oldest registered contents or the least frequently searched contents are sequentially deleted. You can do it.

Embodiment 1 Utilizing a Map Database Next, an embodiment in which a road database can be recognized using a map database will be described with reference to the block diagram of FIG. In this embodiment, the feature amount extracting means 10 and the travel path recognizing means 20 are the same as those in the above-described embodiment. The knowledge base 50 is, as in the previous embodiment,
Global Knowledge Base 30 and Local Knowledge Base 4
It may be composed of two types of databases, that is, 0, or may be composed of only the global knowledge base 30 as in a conventional apparatus. The features of this device are
Further, the point is that a map database 60 is added. In order to use the map database 60, a segment counting section 61 and an environment generating section 62 are provided.

The map database 60 divides a traveling road into a plurality of road segments and prepares map data for each road segment. Although any method may be used for dividing the road segment, in this embodiment, as shown in FIG. 3, a broken white line appears.
The cycle is divided as one road segment. That is, in the example shown in FIG. 3, the area from the upper end of the dashed white line as the center line to the upper end of the next dashed white line is defined as one road segment, and the road segments 1, 2, 2,.
3, 4, ... are defined. In the map database 60, a segment list as shown in FIG. 4 is prepared corresponding to such road segments. This segment list prepares map data (for example, road width W, curvature R, gradient T, etc.) for each road segment.

The operation of this device is as follows. First, similarly to the above-described embodiment, the feature amount (candidate point set W) is extracted by the feature amount extracting unit 10 and is provided to the traveling road recognizing unit 20. On the other hand, the road image input by the road image input unit 11 is also provided to the segment counting unit 61.
The segment counting section 61 has a function of recognizing the segment position of the own vehicle by counting the defined road segments. For example, as in the example shown in FIG. 3, when a road segment is defined for each cycle of a broken white line,
Based on the input road image, it recognizes a dashed white line,
At present, it recognizes on which road segment the vehicle is located. More specifically, if the number of dashed white lines that have passed is counted based on the input road image, this count value corresponds to the road segment number that is currently passing. For example, in FIG. 3, when the vehicle starts traveling with the road segment 1 as a starting point, if the count value becomes 4, the segment position where the own vehicle is present can be specified as the road segment 4. actually,
It is preferable to set a point where visual recognition is easy, such as a junction of a highway, as a starting point of a road segment.

The segment counting section 61 gives information on the recognized segment position to the map database 60. The map database 60 searches for map data corresponding to the given segment position, and provides this to the environment generation unit 62. For example, when road segment 4 is given as a segment position, map data (road width W4, curvature R4, gradient T4, et
c.) and, if necessary, map data corresponding to the road segment ahead thereof is given to the environment generating unit 62. The environment generating unit 62 generates a predicted road environment for the currently traveling road segment 4 and the road segment ahead of the road segment 4 based on such map data. The travel path recognition means 20 performs travel path recognition based on the candidate point set W given as a feature amount. In this recognition work, the travel path recognition means 20 refers to the knowledge information in the knowledge base 50, and generates an environment. The road environment generated in 62 can be used as reference data. The road environment generated by the environment generating unit 62 is information including the content of how wide the road is, how it is curved, and what kind of gradient it has. As a result, the computational burden of the travel path recognition process in the travel path recognition means 20 is greatly reduced. For example, it is possible to more finely extract a feature amount in a direction in which the traveling path is expected to be curved, or to estimate and compensate for a feature amount hidden by an obstacle such as a preceding vehicle. .

A candidate point set W given as a feature quantity;
If the reference data provided from the environment generating unit 62 greatly differs from the reference data, it is considered that an error has occurred in the segment count value in the segment counting unit 61. In such a case, a signal requesting map data for some road segments in the vicinity is given from the traveling road recognition means 20 to the map database 60, and reference data that matches the candidate point set W is generated. It is preferable that the obtained road segment is searched for, and thereby the position correction for the segment counting unit 61 is performed.

Second Embodiment Utilizing a Map Database Next, another embodiment capable of recognizing a traveling route using a map database will be described with reference to the block diagram of FIG. The feature of this embodiment is that the vehicle position Y and the vehicle attitude Z can be accurately recognized. In this embodiment, the feature amount extracting means 10, the traveling path recognizing means 20, and the knowledge base 50 are the same as those in the above-described embodiment. Other features of this device include a map database 70, an absolute position detector 71, a scene generator 72,
Is added. The absolute position detection unit 71 has a function of detecting the current position of the vehicle as an absolute position using a GPS using a satellite, a geomagnetic sensor, a wheel speed sensor, or the like. However, in this device, the detection accuracy of the absolute position may not be so high. In other words,
It is sufficient that the absolute position detection unit 71 can detect a rough position of the own vehicle.

The map database 70 is necessary for generating scenes (corresponding to the external environment observed at the points and the images input by the road image input unit 11) at several points to be traveled. Information is prepared as a database. For example, a three-dimensional shape of a white line at a certain point, a three-dimensional coordinate value of a building existing at the point, and the like are prepared as map data. The scene generation unit 72 has a function of generating a scene at that point based on the map data. Since the absolute position of the own vehicle detected by the absolute position detecting unit 71 is given to the scene generating unit 72, the map database 70 is searched and map data for a point corresponding to the detected position is read. A scene at a location can be generated. The scene generated by the scene generation unit 72 is provided to the traveling path recognition unit 20.
The traveling road recognizing unit 20 performs accurate recognition of the absolute position of the own vehicle by comparing the candidate point set W given as a feature amount from the feature amount extracting unit 10 with the scene given from the scene generating unit 72. be able to.

As a simple example, as shown in FIG. 6, it is assumed that map data is prepared in the map database 70 for five points A to E. The scene generation unit 72 calculates
Scenes a to e can be generated respectively.
For example, the scene a corresponds to an external environment when a specific direction is turned at the point A, that is, a two-dimensional image of a scene.

As described above, the feature of this device is that the vehicle position Y and the vehicle attitude Z can be accurately recognized. In order to perform such accurate recognition, two methods can be used. The first method is as follows. First,
The absolute position detector 71 detects the absolute position and attitude of the host vehicle. Since this detection accuracy is not so high, it includes an error. Now, it is assumed that it is recognized from the detection result by the absolute position detection unit 71 that the vehicle is traveling toward the point A shown in FIG. In this case, the scene generation unit 72 reads the map data for the point A from the map database 70, and creates a scene a when the vehicle reaches the point A in the detected vehicle posture based on the map data. The travel path recognizing means 20 includes the created scene a and the feature amount extracting means 1
The image is compared with an image indicating a feature amount (candidate point set W) given from 0. Then, at the moment when they match, it is recognized that the vehicle has passed the actual point A. At this time, a position correction signal is provided from the traveling path recognition means 20 to the absolute position detection unit 71, and correction is performed so that the detection value of the absolute position detection unit 71 indicates the point A. By performing such a method, the position correction for the absolute position detection unit 71 can be performed each time the vehicle passes a point prepared in the map database 70.

On the other hand, the second method is as follows. Now, it is assumed that information indicating that the vehicle is traveling near points B, D, and E shown in FIG. 6 is obtained from the detection result by the absolute position detection unit 71. Since the detection accuracy of the absolute position detection unit 71 is low, it is assumed that it is not possible to specify where the points B, D, and E are traveling. In such a case,
The scene generation unit 72 reads map data for the points B, D, and E from the map database 70, and based on these map data, scenes b, d, and e of the points B, D, and E in the detected vehicle posture. Create The travel path recognizing unit 20 determines, from among the created scenes b, d, and e, an image that is most similar to the image indicating the feature amount (candidate point set W) provided from the feature amount extracting unit 10. select. For example, when it is determined that the position is closest to the scene b, the position of the own vehicle can be specified as the point B. Thereafter, if the above-described first method is performed, accurate position recognition can be performed.

[0025]

As described above, according to the traveling path recognition apparatus for a mobile vehicle according to the present invention, the knowledge in the knowledge base is extracted during traveling.
Perform various processing using the feature amount of the running road
As a result, higher-speed recognition processing can be performed.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a basic configuration of a traveling road recognition device for a mobile vehicle according to an embodiment of the present invention.

FIG. 2 is a block diagram showing a basic configuration of a traveling road recognition device using a map database.

FIG. 3 is a diagram showing an example of the definition of a road segment in the device shown in FIG. 2;

4 is a map database 60 in the device shown in FIG.
FIG. 4 is a diagram showing an example of the contents of the above.

FIG. 5 is a block diagram showing a basic configuration of another traveling road recognition device using a map database.

6 is a map database 70 in the device shown in FIG.
FIG.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 10 ... Feature extraction means 11 ... Road image input part 12 ... Line candidate point extraction part 20 ... Travel road recognition means 30 ... Global knowledge base 40 ... Local knowledge base 50 ... Knowledge base 60 ... Map database 61 ... Segment counting part 62 ... Environment generation unit 70: Map database 71: Absolute position detection unit 72: Scene generation unit

Continuation of front page (58) Fields investigated (Int. Cl. ⁷ , DB name) G08G 1/16 G01C 21/00 B60R 21/00 G05D 1/02 G06F 15/18

Claims (4)

(57) [Claims] 1. A feature amount extracting means for inputting an image indicating the environment of the outside of a moving vehicle, and extracting a feature amount indicating a feature of the traveling path from the image ; a travel path recognizing means for recognizing a first knowledge base which holds the knowledge necessary to perform the recognition of the shape of the road, the recognition results relating to the shape of the road corresponding to a particular feature quantity <br/> And a second knowledge base in which is registered. The travel path recognizing means first searches the second knowledge base when performing the work of recognizing the travel path, and is similar to the extracted feature amount. When the feature amount is registered, the recognition result of the traveling path corresponding to the similar feature amount is output as it is, and when the feature amount similar to the extracted feature amount is not registered, the first knowledge base is output. Refer to the knowledge held in And obtains the recognition result, the recognition result and outputs, said recognition result transport vehicle of the travel path recognizing device and performs processing for newly registering the second knowledge base. 2. The travel route recognition device according to claim 1 , wherein the feature amount extracting unit includes a road image input unit that inputs a road image of the travel road, and a road lane line based on the input road image. And a line candidate point extracting unit for extracting the line candidate points that constitutes the above-mentioned, and a set of the extracted line candidate points is used as a feature amount. 3. The travel path recognition device according to claim 1, wherein
In place The travel path recognition means determines the vehicle position along with the travel path shape.
Further has the function of performing The first knowledge base contains information about the vehicle position.
Knowledge necessary for In the second knowledge base, the recognition result regarding the vehicle position is also registered.
An apparatus for recognizing a traveling path of a mobile vehicle, comprising: 4. The traveling path according to claim 1,
In the recognition device, The travel path recognition means determines the vehicle attitude along with the travel path shape.
Further has the function of performing The first knowledge base contains information about the vehicle attitude.
Knowledge necessary for In the second knowledge base, the recognition result on the vehicle attitude is also registered.
An apparatus for recognizing a traveling path of a mobile vehicle, comprising: