Studies on Machine Vision-Based Dynamic Multi-Object Recognition

Abstract

Dynamic multi-object acknowledgment in light of machine vision is a vital exploration course in the field of example acknowledgment. In this paper, the multi-object recognizable proof of mechanical fishes in water is considered and A HSV variety space model is proposed. By re-demonstrating the first variety space, the vitally light obstruction in the picture can be successfully taken out, and the shadow of the automated fish can be eliminated, making the foundation division more precise. The improved algorithm outperforms the initial matching tracking algorithm in terms of robustness.