The Research of Forest Fire Smoke Moving Target Detection Methods Based on Video Streaming

摘要

Moving target detection is the premise to classify and recognize the video images correctly. The smoke is a marked characteristic in the early forest fire. Through the analysis of the characteristics of forest fire smoke image, this paper studied several common moving target detection methods, analyzed their realization processes, compared their advantages and disadvantages, and sought the best method of video forest fire smoke moving target detection. Experiments showed that the improved background estimation method combined with color judgment criteria not only has better ability of smoke capture, but also has stronger anti-interference ability, which greatly eases the pressure on the subsequent image recognition.