Research of moving object detection algorithm in transmission lines under complex background
Ye Zhang, Xinbo Huang, Juqing Li, Xinhui Liu, Huiying Zhang, Xiaoqiang Xing · 2016
In allusion to the deficiency of existing methods in moving target detection for transmission lines under complex background, we proposed a method for realizing automatic moving target detection. Adopting the field images with moving target which were taken from the transmission lines anti-theft warning system of a power company in Xinjiang, and choosing the background difference as the object detection algorithm, Gaussian mixture model (GMM) was used for real-time update of background model, through subtracting the updated background image from the current image, we can obtain the foreground moving target. Then, the OTSU algorithm was adopted to complete the image segmentation of moving target area, and marked the foreground moving target in transmission lines filed image to realize the automatic detection for moving target. The results show that the combination of background difference and Gaussian Mixture Background Modeling is sensitive to the change of complex background image, so the complex background can be timely updated and all the foreground moving targets can be automatically extracted from transmission lines images with complex background, showing that the proposed method is applicable for automatically detection of moving target in transmission lines under complex background.