Video Surveillance System and Video Surveillance Device

Patent metadata

Field	Value
Publication number	US-2016127692-A1
Application number	US-201314897492-A
Country	US
Kind code	A1
Filing date	Jun 14, 2013
Priority date	Jun 14, 2013
Publication date	May 5, 2016
Grant date	—

Abstract

A video surveillance technology capable of including motion information from a plurality of surveillance cameras even when the installation conditions of the cameras are different is provided. This video surveillance system is characterized by being equipped with an extraction unit which, upon receiving video images from the plurality of cameras, extracts motion feature quantities from a plurality of frames which constitute the video images, a storage unit for storing extraction results from the extraction unit, a conversion unit for converting the extracted motion feature quantities, and an analysis unit for analyzing the converted motion feature quantities, wherein the conversion unit retains virtual coordinate axes which are different from the coordinate axes of the plurality of cameras and converts the feature quantities by calculating virtual viewpoints on the virtual coordinate axes.

First claim

1 . A video surveillance system comprising:
an extraction unit which receives video images from a plurality of cameras and extracts motion feature quantities from a plurality of frames constituting the video images; a storage unit for accumulating extraction results from the extraction unit; a conversion unit for converting the extracted motion feature quantities; and an analysis unit for analyzing the converted motion feature quantities, wherein the conversion unit retains virtual coordinate axes different from coordinate axes of the plurality of cameras and calculates virtual viewpoints on a virtual coordinate system to convert the feature quantities.

2 . The video surveillance system according to claim 1 , wherein
the conversion unit calculates a position of a crossing point between a sight line axis of each of the cameras and the ground level on the basis of installation information of the cameras, and generates, of positions of the cameras each having a sight line the center of which is the crossing point, a position at which a virtual installation state of the camera is matched with an actual installation state of the camera as the virtual viewpoint.

3 . The video surveillance system according to claim 1 , wherein
the conversion unit further includes: a three-dimensional position estimation unit for estimating a three-dimensional position of the motion feature quantity extracted by the extraction unit in a real space; a conversion method determination unit for determining a conversion method used in conversion of the motion feature quantity by using the virtual viewpoint; and a coordinate conversion unit for performing coordinating conversion of the motion feature quantity by using the estimated three-dimensional position and the determined conversion method.

4 . The video surveillance system according to claim 3 , further comprising:
the plurality of conversion units; and a presentation unit for presenting an analysis result from the analysis unit, and wherein the analysis result presented by the presentation unit is switched by a control signal.

5 . The video surveillance system according to claim 2 , wherein
a position of a white line is extracted from input road information, a point on the white line closest to the crossing point is calculated, and, of positions of the cameras each sight line the center of which is the point on the white line, a position at which a virtual installation state of a camera is matched with an actual installation state of the camera is generated as a virtual viewpoint.

6 . The video surveillance system according to claim 1 , further comprising
a user interface for setting the virtual viewpoint used in the conversion unit.

7 . A video surveillance method comprising:
the extraction step of receiving video images from a plurality of cameras and extracting motion feature quantities from a plurality of frames constituting the video images; the storage step of accumulating extraction results from the extraction step; the conversion step of converting the extracted motion feature quantities; and the analysis step of analyzing the converted motion feature quantities, wherein the conversion step retains virtual coordinate axes different from coordinate axes of the plurality of cameras and calculates virtual viewpoints on a virtual coordinate system to convert the feature quantities.

8 . The video surveillance method according to claim 7 , wherein
the conversion step
calculates a position of a crossing point between a sight line axis of each of the cameras and the ground level on the basis of installation information of the cameras, and generates, of positions of the cameras each having a sight line the center of which is the crossing point, a position at which a virtual installation state of the camera is matched with an actual installation state of the camera as the virtual viewpoint.

9 . The video surveillance method according to claim 7 , wherein
the conversion step further includes: the three-dimensional position estimation step of estimating a three-dimensional position of the motion feature quantity extracted in the extraction step in a real space; the conversion method determination step of determining a conversion method used in conversion of the motion feature quantity by using the virtual viewpoint; and the coordinate conversion step of performing coordinate conversion of the motion feature quantity by using the estimated three-dimensional position and the determined conversion method.

10 . The video surveillance method according to claim 9 , further comprising:
the plurality of conversion steps; and the presentation step of presenting an analysis result from the analysis unit, and wherein the analysis result presented in the presentation step is switched by a control signal.

11 . A computer readable information recording medium on which a program to cause a computer to achieve the following steps:
the extraction step of receiving video images from a plurality of cameras and extracting motion feature quantities from a plurality of frames constituting the video images; the storage step of accumulating extraction results from the extraction step; the conversion step of converting the extracted motion feature quantities; and the analysis step of analyzing the converted motion feature quantities, wherein the conversion step retains virtual coordinate axes different from coordinate axes of the plurality of cameras and calculates virtual viewpoints on a virtual coordinate system to convert the feature quantities.

12 . The information recording medium according to claim 11 , wherein
the conversion step
calculates a position of a crossing point between a sight line axis of each of the cameras and the ground level on the basis of installation information of the cameras, and generates, of positions of the cameras each having a sight line the center of which is the crossing point, a position at which a virtual installation state of the camera is matched with an actual installation state of the camera as the virtual viewpoint.

13 . The information recording medium according to claim 11 , wherein
the conversion step further includes: the three-dimensional position estimation step of estimating a three-dimensional position of the motion feature quantity extracted in the extraction step in a real space; the conversion method determination step of determining a conversion method used in conversion of the motion feature quantity by using the virtual viewpoint; and the coordinate conversion step of performing coordinate conversion of the motion feature quantity by using the estimated three-dimensional position and the determined conversion method.

Assignees

• Hitachi Ltd

Inventors

• Yoneji Kenichi
• Matsubara Daisuke

Classifications

• G06T7/292

  Multi-camera tracking · CPC title

• G06T2207/10016

  Video; Image sequence · CPC title

• G06T2207/30232

  Surveillance · CPC title

• H04N7/181 Primary

  for receiving images from a plurality of remote sources · CPC title

• G06K9/00664

  Physics · mapped topic