An efficient automobile assembly state monitoring system based on channel-pruned YOLOv4 algorithm

Abstract

ABSTRACT An efficient automobile assembly state monitoring system in industrial environment is presented in this paper. The system only needs to input a video that contains the whole detected parts and manually label it in the first frame. By finding the best point for tracking and tracking the point, the dataset can be automatically generated which saves time spent on manufacturing the dataset and makes the assembly state monitoring system easy to deploy into a practical industrial environment. The target detection algorithm uses the channel-pruned YOLOv4 neural network. The experimental result shows the algorithm balances speed and accuracy. Compared to original YOLOv4, the proposed method is two times faster and the performance is nearly equal to it. Comparative experiments show that the proposed algorithm performs better and is faster than other lightweight models which demonstrates that the channel pruning process dynamically improves the speed of the forward propagation without sacrificing accuracy. Additionally, the algorithms are deployed on two common embedded systems. The results show that in the industrial environment, the speed can fully meet real-time requirements.