Method and device for extracting and identifying fire fighting image blocks

Abstract

The invention relates to a method and device for extracting and identifying fire fighting image blocks. A target image is determined according to the difference between a dynamic background frame anda current frame of a source image, and the target image is determined according to a previous frame and the current frame of the source image. According to the target image, the point-by-point micro-wobble, noise filtering, block micro-wobble and block reset screening are carried out to eliminate various noise disturbances and obtain the target images of a recording body and a changing edge of thecurrent frame. Target image block data of the body and the image changing edge is merged into a file of a picture format. The block data is read from the files of the current frame and the first 1~5frames one by one, a block characteristic parameter set is calculated, a matching value is calculated, and whether the block is the target graph required by the user is judged. The invention also discloses a method and device for extracting and identifying, which automatically adapt to various background changes including day and night alternation and the like, better solve the drift problem in the process of image acquisition and identification, and accurately extract an object map required by the user.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to a method and device for extracting fire image blocks. Background technique [0002] Fire fire image recognition technology, that is, the technology of using computer to replace human eyes to automatically recognize fire fire images. Use visible light cameras to monitor flames, smoke, water flow trajectories, laser pointer points, etc. in any background in real time, and identify and extract (suspected) target image blocks such as flames, smoke, water flow trajectories, and laser pointer points in real time. It is extremely valuable for the research and development of an intelligent fire protection system that integrates real-time monitoring, identification, alarm and fire extinguishing of fires. [0003] At present, there are many computer fire image recognition methods, usually using characteristic parameters such as flame color, brightness, flicker, moving target, edg...

AI Technical Summary

Problems solved by technology

[0004] However, since the image of the flame itself and its background are ever-changing and changing, it is difficult to find a set of consistent general rules. Therefore, various flame models inevitably have some shortcomings

Images