Object Recognition Device and Object Recognition System

Patent metadata

Field	Value
Publication number	US-2018165828-A1
Application number	US-201515580478-A
Country	US
Kind code	A1
Filing date	Jun 10, 2015
Priority date	Jun 10, 2015
Publication date	Jun 14, 2018
Grant date	—

Abstract

An object recognition device includes a recognition unit, a recognition reliability calculation unit, and a combining unit. A recognition unit recognizes an object by a plurality of functions, based on information which is obtained by measuring an object by a plurality of measurement devices. A recognition reliability calculation unit calculates recognition reliability of recognition results that are obtained by recognizing an object by a recognition unit, for each function. A combining unit combines recognition reliability of the object and recognition results, detects a specified object, and outputs detection results of the specified object.

First claim

1 . An object recognition device comprising:
a recognition unit that recognizes an object by a plurality of functions, based on information which is obtained by measuring the object by a plurality of measurement devices; a recognition reliability calculation unit that calculates recognition reliability of recognition results that are obtained by recognizing the object by the recognition unit, for each function; and a combining unit that combines recognition reliability of the object and the recognition results, detects a specified object, and outputs detection results of the specified object.

2 . The object recognition device according to claim 1 ,
wherein in a case where information that is obtained by measuring the object by the plurality of measurement devices is a captured image which is obtained by capturing an image of the object and a parallax image which includes distance information to the object, the recognition unit is a captured image recognition unit that recognizes the object based on the captured image, and a three-dimensional recognition unit that recognizes the object based on the distance information which is included in the parallax image, the captured image recognition unit calculates a matching degree between a pattern identifier that is generated based on sample data which is read from a database and a feature part of the object that is extracted from the captured image, the three-dimensional recognition unit calculates a matching degree between a template that is read from the database and three-dimensional information of the object in an overhead view image that is obtained by converting a viewpoint of the parallax image, and the recognition reliability calculation unit calculates recognition reliability for recognizing the object in the captured image by the captured image recognition unit and recognition reliability for recognizing the object in the parallax image by the three-dimensional recognition unit.

3 . The object recognition device according to claim 2 , wherein the recognition reliability calculation unit includes
an image information acquisition unit that acquires the captured image and the parallax image from the measurement device, a measurement device information acquisition unit that acquires measurement device information of the measurement device from the database, an area information acquisition unit that acquires an image-capturing range division method and matrix data of the recognition reliability which correspond to the measurement device information from the database for each area, and a recognition reliability determination unit that determines recognition reliability for recognizing the object by the captured image recognition unit and recognition reliability for recognizing the object by the three-dimensional recognition unit for each of the areas which are obtained by dividing the image-capturing range of the captured image in predetermined position and size, based on the image-capturing range division method.

4 . The object recognition device according to claim 3 , wherein in a case where results which are obtained by combining a degree indicating that the object which is recognized by the captured image recognition unit from the captured image is the specified object, a recognition reliability for recognizing the object in the captured image, a degree indicating that the object which is recognized by the three-dimensional recognition unit from the parallax image is the specified object, and recognition reliability for recognizing the object in the parallax image, are equal to or larger than a predetermined value, the combining unit sets the obtained results to detection results of the specified object.

5 . The object recognition device according to claim 4 ,
wherein the measurement device information, the division method, and the recognition reliability can be changed by an input from an operation unit through a graphical user interface (GUI) which is displayed on a display unit, and the measurement device information, the division method, and the recognition reliability which are changed are recorded in the database, and wherein the recognition reliability calculation unit calculates the recognition reliability for each of the areas, based on the measurement device information, the division method, and the recognition reliability which are changed and acquired from the database.

6 . The object recognition device according to claim 5 , further comprising:
a recognition function adjustment unit that adjusts a function of recognizing the object by the captured image recognition unit and a function of recognizing the object by the three-dimensional recognition unit, based on the measurement device information; and a recognition reliability updating unit that updates the matrix data in which the recognition reliability is stored, based on a function of recognizing the adjusted object.

7 . The object recognition device according to claim 6 , wherein the measurement device information includes at least one of an installation environment of the measurement device and internal information of the measurement device.

8 . The object recognition device according to claim 1 ,
wherein in a case where information that is measured by the plurality of measurement devices is a captured image which is obtained by capturing an image of the object and a distance image in which distance information to the object is embedded in the captured image, the recognition unit is a captured image recognition unit that recognizes the object based on the captured image, and a distance image recognition unit that recognizes the object based on the distance information which is included in the distance image, the captured image recognition unit calculates a matching degree of the object between a feature part that is extracted from the captured image and a pattern identifier that is generated based on sample data which is read from the database, the distance image recognition unit calculates a matching degree between the object in an overhead view image that is obtained by converting a viewpoint of the distance image and a template that is read from the database, and the recognition reliability calculation unit calculates recognition reliability for recognizing the object in the captured image by the captured image recognition unit and recognition reliability for recognizing the object in the distance image by the distance image recognition unit.

9 . An object recognition system comprising:
a plurality of measurement devices that measure an object; a recognition unit that recognizes an object by a plurality of functions, based on information which is obtained by measuring the object by the plurality of measurement devices; a recognition reliability calculation unit that calculates recognition reliability of recognition results that are obtained by recognizing the object by the recognition unit, for each function; a combining unit that combines recognition reliability of the object and the recognition results, detects a specified object, and outputs detection results of the specified object; and a display unit that displays the detection results.

Assignees

• Hitachi Ltd

Inventors

• Sasatani So
• Itoh Masaya
• Yumiba Ryou

Classifications

• G06V20/64 Primary

  Three-dimensional [3D] objects · CPC title

• G06T7/70 Primary

  Determining position or orientation of objects or cameras (camera calibration G06T7/80) · CPC title

• G06T2200/04

  involving 3D image data · CPC title

• G06V10/40

  Extraction of image or video features · CPC title

• G06K9/00671

  Physics · mapped topic