Pedestrian motion predicting device

Patent metadata

Field	Value
Publication number	US-9507998-B2
Application number	US-201114125862-A
Country	US
Kind code	B2
Filing date	Jun 13, 2011
Priority date	Jun 13, 2011
Publication date	Nov 29, 2016
Grant date	Nov 29, 2016

Abstract

A subject is to provide a pedestrian motion predicting device capable of accurately predicting a possibility of a rush out before a pedestrian actually begins to rush out. According to the embodiments, the pedestrian is detected from input image data, a portion in which the detected pedestrian is imaged is cut out from the image data, a shape of the pedestrian imaged in the cut-out partial image data is classified by collating the shape with a learning-finished identifier group or a pedestrian recognition template group, and the rush out of the pedestrian is predicted based on a result of the acquired classification.

First claim

The invention claimed is:

1. A pedestrian motion predicting device comprising:
a photographing unit mounted on a vehicle, the photographing unit configured to photograph a periphery of the vehicle and to create image data of the periphery of the vehicle;
a controller; and
a data storage, wherein
the controller receives the image data from the photographing unit, and the controller is configured to detect a pedestrian from the image data;
the controller is configured to acquire leg width values of the pedestrian from the image, each of the leg width values being a space between legs of the pedestrian;
the data storage configured to store the leg width values of the pedestrian in time series acquired by the controller;
the controller is configured to analyze a periodicity of the leg width values based on the leg width values of the pedestrian in time series stored in the data storage; and
the controller is configured to detect a break in the periodicity of the leg width values and to predict, based on a leg width change amount, whether the pedestrian will rush out at a speed according to the leg width change amount, the leg width change amount being a difference between a leg width value before the break in periodicity and a leg width value after the break in periodicity.

2. The pedestrian motion predicting device according to claim 1 , wherein
the data storage is configured to store learning-finished identifier group data or pedestrian recognition template group data, and
the controller detects shapes of the pedestrian from the image data and collates the shapes of the pedestrian with the learning-finished identifier group data or the pedestrian recognition template group data stored in the data storage, thereby classifying the shapes of the pedestrian.

3. The pedestrian motion predicting device according to claim 2 , wherein
the data storage is configured to store in time series a result of a classification of the shapes of the pedestrian,
the controller is configured to determine, based on the result of the classification stored in time series, whether a moving direction and a speed of the pedestrian are continuous, and
the controller is configured to predict a rush out direction and a rush out speed of the pedestrian based on a result of determination.

4. A pedestrian motion predicting method, comprising;
photographing a periphery of a vehicle and creating image data of the periphery of the vehicle by a photographing unit mounted on the vehicle;
receiving the image data from the photographing unit by a controller;
acquiring by the controller leg width values of the pedestrian from the image data, each of the leg width values being a space between legs of the pedestrian;
storing the leg width values of the pedestrian in time series in a data storage;
analyzing by the controller a periodicity of the leg width values based on the leg width values of the pedestrian in time series stored in the data storage; and
detecting by the controller a break in the periodicity of the leg width values and predicting by the controller, based on a leg width change amount, whether the pedestrian will rush out at a speed according to the leg width change amount, the leg width change amount being a difference between a leg width value before the break in periodicity and a leg width value after the break in periodicity.

5. The pedestrian motion predicting method according to claim 4 , further comprising:
detecting by the controller shapes of the pedestrian from the image data; and
collating by the controller the shapes of the pedestrian with learning-finished identifier group data or pedestrian recognition template group data stored in the data storage, thereby classifying the shapes of the pedestrian.

6. The pedestrian motion predicting method according to claim 5 , further comprising:
storing a result of a classification of the shapes of the pedestrian in time series in the data storage,
determining by the controller, based on the result of the classification stored in time series, whether a moving direction and a speed of the pedestrian are continuous, and
predicting by the controller a rush out direction and a rush out speed of the pedestrian based on a result of determination.

Assignees

• Ogawa Masahiro
• Kindo Toshiki
• Funayama Ryuji
• Sakamoto Tadaaki
• Toyota Motor Co Ltd

Inventors

• Ogawa Masahiro
• Kindo Toshiki
• Funayama Ryuji
• Sakamoto Tadaaki

Classifications

• G08G1/166 Primary

  for active traffic, e.g. moving vehicles, pedestrians, bikes · CPC title

• G06T7/20

  Analysis of motion (motion estimation for coding, decoding, compressing or decompressing digital video signals H04N19/43, H04N19/51) · CPC title

• G06K9/00348 Primary

  Physics · mapped topic

• G06V40/25 Primary

  Recognition of walking or running movements, e.g. gait recognition · CPC title