A Method of Video Fire Smoke Image Recognition Based on Computer Automatic Location

Fire smoke needs early detection and accurate identification, so as to protect people's lives and property, while manual control method has problems such as large time consumption, subjective misjudgment, so an efficient, accurate and real-time fire smoke image recognition method is needed. In this paper, a method of video fire smoke image recognition is proposed based on computer automatic location algorithm, computer vision and image processing technology, combined with optimized algorithms and high performance hardware. Through real-time processing and analysis of video frames, the method can quickly and accurately locate the fire smoke area, and realize real-time monitoring and alarm. After experimental testing, it is found that the accuracy of the proposed method is between 94% and 99%, maintaining a high accuracy. The method has high-speed processing capability and automatic characteristics, and can quickly analyze a large number of video image data and accurately locate the fire smoke area.