Abstract:Aiming at the problems of complex application station environment, uneven illumination intensity, unclear monitoring video and inaccurate personnel intrusion alarm, a personnel intrusion alarm method based on visual electronic fence is proposed. The adaptive weighted median filter algorithm filters out the noise in the surveillance video, and combines the inter frame difference method and the background difference method to construct the background update difference method. The background image is updated in real time and the image illumination is compensated for the target detection of moving personnel. The mean shift algorithm is used to track the occluded moving personnel target. According to the target recognition results and relevant thresholds, the personnel intrusion is determined, and the alarm or alarm cancellation is made. The experimental results show that the proposed method has ideal video processing effect, high accuracy, recall and F1 value, and less number of parameters and floating-point operations. |