SYSTEM AND METHOD FOR VIDEO DETECTION OF SMOKE AND FLAME

• US 20110064264A1
• Filed: 05/08/2008
• Published: 03/17/2011
• Est. Priority Date: 05/08/2008
• Status: Active Grant

• Current Assignee: UTC Fire & Security (Rtx Corporation)
• Original Assignee: UTC Fire & Security (Rtx Corporation)
• Inventors: Caballero, Rodrigo E., Finn, Alan Matthew, Peng, Pei-Yuan, Wang, Hongcheng, Xiong, Ziyou
• Granted Patent: US 8,462,980 B2
• Time in Patent Office: Days
• US Class Current: 382/100
• CPC Class Codes: G06V 20/52 Surveillance or monitoring of activities, e.g. for recognising suspicious objects recognising microscopic objects G06V20/69
• CPC Class Codes: G08B 17/125 by using a video camera to detect fire or smoke

Abstract

A video recognition system detects the presence of fire based on video data provided by one or more video detectors. The video recognition system divides each individual frame of video data into a plurality of frame elements and calculates with respect to each frame element a structural feature that the algorithm uses to identify a core region of potential fire. The video recognition system defines a region surrounding the core region and calculates with respect to each frame element within the surrounding region a temporal feature. Based on the calculated temporal feature, the video recognition system associates a representative temporal value with the core region. Based on the structural and temporal features associated with the core region, video recognition system determines whether the video data indicates the presence of fire.

Claims

1. A method of detecting fire using video analysis, the method comprising:

acquiring video data comprised of individual frames;

dividing each of the individual frames into a plurality of frame elements;

calculating a structural feature associated with each frame element;

identifying a core region of fire based on the calculated structural features;

defining a region surrounding the core region;

calculating a temporal feature associated with each frame element within the surrounding region;

associating the temporal features calculated with respect to frame elements within the surrounding region with the identified core region by assigning a representative temporal feature value to the core region; and

generating an output indicating whether the analyzed video data indicates the presence of fire based, at least in part, on the structural features and assigned temporal features associated with the identified core region.

2. The method of claim 1, wherein identifying a core region of fire includes:

identifying a core region of flame based on one or more of the following structural features associated with each frame element;

a color feature, a color intensity feature, a geometric feature, texture feature, and a morphological feature.

3. The method of claim 1, wherein identifying a core region of fire includes:

identifying a core region of smoke based on one or more of the following structural features associated with each frame element;

a color feature, an edge data feature, a geometric feature, a texture feature, and a morphological feature.

4. The method of claim 1, wherein the surrounding region is defined as including the core region and a defined number of frame elements surrounding a periphery of the core region.

5. The method of claim 1, wherein the surrounding region is defined based on geometric properties of the identified core region.

6. The method of claim 1, wherein defining the surrounding region includes:

defining a bounding box that includes all frame elements identified as part of the core region; and

defining the surrounding region as including the bounding box and a number of frame elements surrounding a periphery of the bounding box.

7. The method of claim 1, wherein associating temporal features calculated with respect to frame elements within the surrounding region with the identified core region includes:

associating a temporal feature associated with flame to the identified core region.

8. The method of claim 1, wherein associating temporal features calculated with respect to frame elements within the surrounding region with the identified core region includes:

associating a temporal feature associated with smoke to the identified core region.

9. The method of claim 1, wherein associating the temporal features calculated with respect to frame elements within the surrounding region with the identified core region includes:

assigning a maximum temporal feature calculated with respect to the frame elements in the surrounding region to all frame elements within the core region.

10. The method of claim 1, wherein associating the temporal features calculated with respect to frame elements within the surrounding region with the identified core region includes:

calculating an average flicker value based on the flicker values generated with respect to each of the frame elements in the surrounding region; and

assigning the average flicker value to all frame elements within the core region.

11. A system for detecting the presence of fire, the system comprising:

a frame operably connectable to receive video data comprised of a plurality of individual frames and to store the received video data;

a frame element divider that divides each of the individual frames into a plurality of frame elements;

a core region identifier that calculates with respect to each frame element a structural feature that is used to identify a core region of potential fire;

a surrounding region identifier that defines a region surrounding the core region;

a temporal feature calculator that calculates with respect to each frame element within the surrounding region a temporal feature;

an assignor that associates a representative temporal feature calculated with respect to the surrounding region to the core region; and

decisional logic that determines based on the structural and temporal features associated with the core region whether the video data indicates the presence of fire and generates the output based on this determination.

12. The system of claim 11, wherein the surrounding region identifier defines a size of the surrounding region based on a size of the core region.

13. The system of claim 11, wherein the surrounding region identifier defines a bounding box that includes all frame elements identified as part of the core region and defines the surrounding region as including the bounding box and a defined number of frame elements surrounding the periphery of the bounding box.

14. The system of claim 11, wherein the assignor associates a temporal feature having a frequency indicative of flame to the identified core region.

15. The system of claim 11, wherein the assignor associates a temporal feature having a frequency indicative of smoke to the identified core region.

16. A video recognition system comprising:

means for acquiring video data comprised of individual frames;

means for dividing each of the individual frames into a plurality of frame elements;

means for calculating a structural feature associated with each frame element;

means for identifying a core region of fire based on the calculated structural features;

means for defining a region surrounding the core region;

means for calculating a temporal feature associated with each frame element within the surrounding region;

means for associating the temporal features calculated with respect to frame elements within the surrounding region with the identified core region by assigning a representative temporal feature value to the core region; and

means for generating an output indicating whether the analyzed video data indicates the presence of fire based, at least in part, on the structural features and assigned temporal features associated with the identified core region.

17. The video recognition system of claim 16, wherein the means for identifying a core region includes:

means for identifying a core region of flame based on one or more structural features; and

means for identifying a core region of smoke based on one or more structural features.

18. The video recognition system of claim 16, wherein the means for associating a temporal feature calculated with respect to frame elements within the surrounding region includes association a temporal feature having a frequency indicating the presence of flame to the core region.

19. The video recognition system of claim 16, wherein the means for associating a temporal feature calculated with respect to frame elements within the surrounding region includes association a temporal feature having a frequency indicating the presence of smoke to the core region.

20. A computer readable storage medium encoded with a machine-readable computer program code for generating a fire detection output, the computer readable storage medium including instructions for causing a controller to implement a method comprising:

acquiring video data comprised of individual frames;

dividing each of the individual frames into a plurality of frame elements;

calculating a structural feature associated with each frame element;

identifying a core region of fire based on the calculated structural features;

defining a region surrounding the core region;

calculating a temporal feature associated with each frame element within the surrounding region;

associating the temporal features calculated with respect to frame elements within the surrounding region with the identified core region by assigning a representative temporal feature value to the core region; and

generating an output indicating whether the analyzed video data indicates the presence of fire based, at least in part, on the structural features and assigned temporal features associated with the identified core region.