Feature information collecting apparatus and feature information collecting method

• US 8,229,169 B2
• Filed: 03/13/2008
• Issued: 07/24/2012
• Est. Priority Date: 03/30/2007
• Status: Expired due to Fees

• Current Assignee: Aisin AW Company Limited (Aisin Corp.)
• Original Assignee: Aisin AW Company Limited (Aisin Corp.)
• Inventors: Nakamura, Masaki, Nakao, Koichi, Kanba, Motoki
• Primary Examiner(s): Koziol, Stephen R
• Application Number:
  US12/076,068
• Publication Number:
  US 20080240573A1
• Time in Patent Office: 1,594 Days
• Field of Search: 382/100, 382/104, 382/190
• US Class Current: 382/104
• CPC Class Codes: G01C 21/26 specially adapted for navigation in a road network
• CPC Class Codes: G06V 20/588 Recognition of the road, e.g. of lane markings; Recognition of the vehicle driving pattern in relation to the road
• CPC Class Codes: G06V 20/80 Recognising image objects characterised by unique random patterns

Abstract

Feature information collecting apparatuses, methods, and programs acquire vehicle position information that represents the current position of a vehicle, acquire image information for a vicinity of the vehicle, and carry out image recognition processing on recognition target objects that are included in the image information. The apparatuses, methods, and programs store recognition information in a memory that represents a result of the image recognition of the recognition target objects in association with information for the recognition position of the recognition target objects, the recognition position determined based on the vehicle position information. The apparatuses, methods, and programs extract, as learned features, recognition target objects that can be repeatedly recognized by image recognition based on a plurality of sets of recognition information related to the same position, the plurality of sets of recognition information being stored due to the image information for the same position being recognized a plurality of times by image recognition.

Claims

1. A feature information collecting apparatus, comprising:

a first memory; and

a controller that;

acquires vehicle position information that represents a current position of a vehicle;

acquires image information for a vicinity of the vehicle;

carries out image recognition processing on recognition target objects that are included in the image information;

stores recognition information in the first memory that represents a result of the image recognition of the recognition target objects in association with information for the recognition position of the recognition target objects, the recognition position determined based on the vehicle position information; and

extracts, as learned features, recognition target objects that can be repeatedly recognized by image recognition based on a plurality of sets of recognition information related to the same position, the plurality of sets of recognition information being stored due to the image information for the same position being recognized a plurality of times by image recognition;

wherein the recognition information is a characteristic amount represented by at least one of a noise amount and an amount of extractions of edge included in the image information.

2. The feature information collecting apparatus according to claim 1, wherein:

the controller stores the recognition results for the recognition target objects in association with information about the recognition position in the first memory;

the recognition information includes a distribution of the recognition results for each target object; and

the controller extracts recognition target objects at a positions where a peak in the distribution is present as the learned features, each peak representing the position information for each extracted learned feature.

3. The feature information collecting apparatus according to claim 1, wherein the controller:

adds a learning value to each of a predetermined position range each time the recognition results for the recognition target objects satisfy learning conditions, the learning value allocated due to the recognition results for recognition target objects satisfying predetermined learning conditions;

extracts as the learned features the recognition target objects in a position range in which the learning value has become equal to or greater than a predetermined learning threshold value; and

uses a position associated with each position range as the position information for the learned features.

4. The feature information collecting apparatus according to claim 1, wherein:

the recognition information is stored in association with predetermined conditions including information about at least one of recognition time and weather during the recognition; and

the controller extracts recognition target objects that can be repeatedly recognized by image recognition during the predetermined conditions as limited learned features that are valid only when the conditions are satisfied.

5. The feature information collecting apparatus according to claim 1, wherein the controller:

determines a feature-to-feature distance that represents a distance between two learned features based on the position information for the two learned features.

6. The feature information collecting apparatus according to claim 1, wherein:

the apparatus is connected so as to be able to communicate with a plurality of vehicles; and

the first memory stores the recognition information that has been obtained by the plurality of vehicles.

7. The feature information collecting apparatus according to claim 1, wherein:

the learned features include at least one of road markings provided on a surface of a road, stains on the surface of a road, grime on the surface of a road, cracks in the surface of a road, and shadows on the surface of a road.

8. The feature information collecting apparatus according to claim 1, wherein:

the first memory stores feature information for the target features including;

learned feature information that is associated with attribute information related to the recognition information for the learned features and position information for the learned features;

or

both the learned feature information and initial information that includes position information and attribute information that has been prepared in advance for a plurality of features.

9. The feature information collecting apparatus according to claim 8, wherein the controller:

refers to the feature information based on the vehicle position information and determines whether the target features that have been stored as feature information for the vicinity of the vehicle are present; and

carries out the storing of the recognition information when the target features are not present in the vicinity of the vehicle.

10. A vehicle position recognizing apparatus comprising:

the feature information collecting apparatus according to claim 8; and

a controller that;

corrects the vehicle position information based on the image recognition result for the target features;

carries out image recognition processing for the target features represented by the feature information that is included in the image information, based on the feature information for the target features that are present in the vicinity of the vehicle that is extracted based on the vehicle position information; and

corrects the vehicle position information, based on the image recognition results for the target features and the position information that is included in the feature information for the target features.

11. A navigation apparatus comprising:

the vehicle position recognizing apparatus according to claim 10;

a second memory that stores map information; and

a controller that provides outputs guidance information based on the map information.

12. A feature information collecting method, comprising:

acquiring vehicle position information that represents a current position of a vehicle;

acquiring image information for a vicinity of the vehicle;

carrying out image recognition processing on recognition target objects that are included in the image information;

storing recognition information that represents a result of the image recognition of the recognition target objects in association with information for the recognition position of the recognition target objects, the recognition position determined based on the vehicle position information; and

extracting, as learned features, recognition target objects that can be repeatedly recognized by image recognition based on a plurality of sets of recognition information related to the same position, the plurality of sets of recognition information being stored due to the image information for the same position being recognized a plurality of times by image recognition;

wherein the recognition information is a characteristic amount represented by at least one of a noise amount and an amount of extractions of edge included in the image information.

13. The feature information collecting method according to claim 12, further comprising:

storing the recognition results for the recognition target objects in association with information about the recognition position, the recognition information including a distribution of the recognition results for each target object; and

extracting recognition target objects at a positions where a peak in the distribution is present as the learned features, each peak representing the position information for each extracted learned feature.

14. The feature information collecting method according to claim 12, further comprising:

adding a learning value to each of a predetermined position range each time the recognition results for the recognition target objects satisfy learning conditions, the learning value allocated due to the recognition results for recognition target objects satisfying predetermined learning conditions;

extracting as the learned features the recognition target objects in a position range in which the learning value has become equal to or greater than a predetermined learning threshold value; and

using a position associated with each position range as the position information for the learned features.

15. The feature information collecting method according to claim 12, wherein:

the recognition information is stored in association with predetermined conditions including information about at least one of recognition time and weather during the recognition; and

the method further comprises extracting recognition target objects that can be repeatedly recognized by image recognition during the predetermined conditions as limited learned features that are valid only when the conditions are satisfied.

16. The feature information collecting method according to claim 12, further comprising:

determining a feature-to-feature distance that represents a distance between two learned features based on the position information for the two learned features.

17. The feature information collecting method according to claim 12, further comprising:

communicating with a plurality of vehicles; and

storing the recognition information that has been obtained by the plurality of vehicles.

18. The feature information collecting method according to claim 12, wherein:

the learned features include at least one of road markings provided on a surface of a road, stains on the surface of a road, grime on the surface of a road, cracks in the surface of a road, and shadows on the surface of a road.