Object recognition apparatus, objection recognition method, and program

Patent metadata

Field	Value
Publication number	US-10452950-B2
Application number	US-201715591690-A
Country	US
Kind code	B2
Filing date	May 10, 2017
Priority date	May 25, 2016
Publication date	Oct 22, 2019
Grant date	Oct 22, 2019

Abstract

An object recognition apparatus includes image information acquisition means for acquiring image information of an object to be recognized, storage means for storing detection profile information associating an object candidate with a detector capable of detecting the object candidate, and model image information of the object candidate associated with the object candidate and object detection means including detectors defined in the detection profile information, the object detection means detecting the object to be recognized by using the detector from the image information acquired by the image information acquisition means. The detector of the object detection means detects the object candidate by comparing the model image information of the object candidate associated with the detector in the detection profile information with the image information of the object to be recognized acquired by the image information acquisition means, and outputs the detected object candidate as the object to be recognized.

First claim

What is claimed is:

1. An object recognition apparatus comprising:
a processor configured to:
acquire image information of an object to be recognized;
store detection profile information associating an object candidate with a detector capable of detecting the object candidate, and model image information of the object candidate associated with the object candidate, the object candidate being a candidate for the object to be recognized; and
operate as a plurality of detectors defined in the detection profile information, each of the detectors being configured to detect the object to be recognized based on the acquired image information, wherein
each of the detectors is configured to detect the object candidate by comparing the model image information of the object candidate associated with the detector in the detection profile information with the acquired image information of the object to be recognized and output the detected object candidate as the object to be recognized, and
the processor is further configured to:
store determination profile information associating the object candidate with a determination unit capable of determining the object candidate;
operate as a plurality of determination units defined in the determination profile information, each of the determination units configured to determine the object candidate output from the detectors based on the acquired image information; and
select a determination unit corresponding to the object candidate detected by the detectors based on the stored determination profile information.

2. The object recognition apparatus according to claim 1 , the processor is further configured to:
determine whether or not the object candidate has a reliability equal to or larger than a predetermined value by comparing at least one of color information and distance information of the acquired image information of the object to be recognized with at least one of color information and distance information of the model image information of the object candidate output from the detectors, and
output the object candidate that is determined to have reliability equal to or larger than the predetermined value as the object to be recognized.

3. The object recognition apparatus according to claim 1 , wherein the processor is further configured to detect a position and posture candidate of the object candidate based on a feature value of the acquired image information of the object to be recognized and a feature value of the model image information of the detected object candidate.

4. The object recognition apparatus according to claim 3 , wherein the processor is further configured to unify similar candidates among object candidates and position and posture candidates detected by the detectors as an identical candidate.

5. The object recognition apparatus according to claim 3 , wherein
the processor is further configured to perform a tracking process and output a tracking-processed object candidate and its position and posture candidate, and
the tracking process includes
obtaining a geometric quantity between a model of a model image of the object candidate and the position and posture candidate detected by the detectors and the object to be recognized of the image information, and
eliminating an object candidate and a position and posture candidate for which the obtained geometric quantity is equal to or larger than a predetermined quantity from the object candidates and the position and posture candidates detected by the detectors.

6. The object recognition apparatus according to claim 5 , wherein the processor is further configured to:
store information of a nearest neighbor point on the model in the model image from each grid of a three-dimensional voxel enclosing the model to the model, and
perform the tracking process by using the stored information of the nearest neighbor point.

7. The object recognition apparatus according to claim 2 , wherein
the processor is further configured to generate at least one of the detection profile information and the determination profile information,
the detection profile information associates the object candidate with at least one detector based on feature information indicating a feature of the object candidate including at least one of color information, shape information, inscription information, and physical-property information of the object candidate, and
the determination profile information associates the object candidate with at least one determination unit based on feature information indicating a feature of the object candidate including at least one of color information and inscription information of the object candidate.

8. An object recognition method comprising:
acquiring image information of an object to be recognized;
storing detection profile information associating an object candidate with a detector capable of detecting the object candidate, and model image information of the object candidate associated with the object candidate, the object candidate being a candidate for the object to be recognized; and
detecting the object to be recognized by using a plurality of detectors defined in the detection profile information based on the acquired image information, wherein
each of the detectors detects the object candidate by comparing the model image information of the object candidate associated with the detector in the detection profile information with the acquired image information of the object to be recognized and outputs the detected object candidate as the object to be recognized, and
the method further comprises:
storing determination profile information associating the object candidate with a determination unit capable of determining the object candidate;
determining the object candidate output from the detectors based on the acquired image information by using a plurality of determination units defined in the determination profile information; and
selecting a determination unit corresponding to the object candidate detected by the detectors based on the stored determination profile information.

9. A non-transitory computer readable medium storing program for causing a computer to execute:
a process of acquiring image information of an object to be recognized;
a process of storing detection profile information associating an object candidate with a detector capable of detecting the object candidate, and model image information of the object candidate associated with the object candidate, the object candidate being a candidate for the object to be recognized; and
a process of detecting the object to be recognized by using a plurality of detectors defined in the detection profile information based on the acquired image information, wherein
each of the detectors detects the object candidate by comparing the model image information of the object candidate associated with the detector in the detection profile information with the acquired image information of the object to be recognized and outputs the detected object candidate as the object to be recognized, and
the program further causes the computer to execute:
a process of storing determination profile information associating the object candidate with a determination unit capable of determining the object candidate;
a process of determining the object candidate output from the detectors based on the acquired image information by using a plurality of determination units defined in the determination profile information; and
a process of selecting a determination unit corresponding to the object candidate detected by the detectors based on the stored determination profile information.

Assignees

• Toyota Motor Co Ltd

Inventors

• Kobori Norimasa
• Hashimoto Kunimatsu
• Yamauchi Minoru

Classifications

• G06V10/761

  Proximity, similarity or dissimilarity measures · CPC title

• G06V20/64 Primary

  Three-dimensional [3D] objects · CPC title

• G06F18/22

  Matching criteria, e.g. proximity measures · CPC title

• G06F18/24147

  Distances to closest patterns, e.g. nearest neighbour classification · CPC title

• G06T7/248

  involving reference images or patches · CPC title