Complex Scene Tracking Algorithm Based on Multi-feature Fusion

Abstract

At present, natural disasters of coalmine in China are still serious, and coalmine safety guarantee capabilities are still relatively low. In recent years, a lot of manpower and material resources have been implemented to upgrade the coalmine monitoring system, but manual operation has limited their applications and effects, the efficiency is low and the error rate is high. The application of computers for intelligent tracking and monitoring of dynamic object can greatly reduce the waste of manpower, improve monitoring efficiency, and provide more reliable and concise safety early warning and system linkage than manual operation. In addition, object tracking can provide information support for behavior understanding, event detection, object classification, its importance is self-evident. The object tracking algorithm has achieved good results in open scenes with rich features and sufficient illumination. However, in confined spaces such as coalmine and tunnels, due to unfavorable factors such as coal dust, lack of illumination, scale changes, and image blur-ring, etc. The stability and robustness of tracking are greatly affected. In this paper, the deep network model is used to extract the depth features of the tracking target, the depth features and the results of the HOG feature location filter are weighted and merged to obtain the final target location. The algorithm has better tracking stability and robustness in the case of insufficient illumination and scarce feature points in coalmines. The dynamic object tracking experiment is carried out on the coalmine monitoring video. The experimental results show that the algorithm is more robust than the comparison algorithm and can meet the needs of object tracking in complex scenes in coalmine.