Smoke Diffusion Feature Extraction and Concentration Prediction Method Based on Canny Edge Detection

In the dynamic intersection scene between laser detector and smoke interference, it is difficult to quantitatively describe the smoke state parameter information of laser detection optical path at different times due to the fluidity and randomness of smoke. This paper proposes a smoke diffusion video feature extraction and concentration prediction method based on Canny edge detection. This method extracts high-speed video of smoke diffusion, detects the edge of the video image and segments the smoke area, analyzes and obtains the corresponding relationship between the shielding effect of smoke diffusion, the number of image pixels in the transmission path and the smoke concentration, By extracting the shielding effect of 1, 2, and 4 typical smoke scenes during diffusion, and analyzing the number of image pixels of a certain transmission path of smoke diffusion, the smoke impact in the dynamic intersection scene of the laser detector and smoke interference is quantified. Subsequently, the laser detector dynamic intersection scene and detection beam parameters can be combined to quantitatively describe the relative position of the laser beam in the smoke transmission path and the concentration state estimation in different forms of smoke scenes, providing important support for the time-correspondence of the echo signal and the optimization of the smoke interference algorithm.