A Survey of Deep Learning Target Detection Algorithms under Candidate Regions

Abstract

Target detection is a research hotspot in the field of computer vision. The target detection algorithm based on candidate regions has high accuracy, but the structure is complex, and the detection speed is slow. The regression-based target detection algorithm has simple structure and fast detection speed, and it is in real-time target detection The field has high application value, but the detection accuracy is relatively low. Based on this research background, this paper first analyses the mechanism of shadow generation, understands the characteristics of shadows and human visual characteristics, and summarizes and classifies the current existing features. Based on the similar shadow detection algorithm, a shadow detection algorithm based on the RGB colour model is proposed. The algorithm in this paper is verified through experiments. The article proves that the algorithm can detect and remove the shadows of moving targets well.