Construction Crane Detection under Transmission Line Based on Improved Vibe Algorithm

Abstract

In this paper, the moving object detection part of crane’s illegal construction detection under transmission line is studied, and two scenes are simulated. The Visual background extraction (Vibe) algorithm is selected as the moving object detection algorithm used in this paper. For the problems of the algorithm, including ghost phenomenon in detection results and the the presence of falsely detected foreground points under dynamic background, the Vibe algorithm is improved by the reconstructed background model, dynamic radius threshold and connected domain analysis. Results of the experiment show that ghost area in the detection results of the improved Vibe algorithm can quickly disappear, and the number of falsely detected foreground points is significantly reduced. Compared with the original Vibe algorithm, the average value of the comprehensive evaluation index F1 is increased by 5.8%, and the detection effect of moving object is significantly improved.