Answers are grounded in this paper and cite their sources.
Demonstrates improved defect detection accuracy in strip steel production, suggesting better surface quality assurance.
With the continuous improvement of the requirements for surface quality assurance in steel strip industrial production, the automatic defect detection based on computer vision has been widely concerned. Currently, the algorithm based on convolutional neural network can achieve satisfactory accuracy in the task of surface defect detection. However, these high-precision deep convolution networks need a lot of computing power, which makes it difficult for these network models to be directly applied to factories. In this paper, a lightweight object detection framework based on MobileNetv3 backbone model is proposed, and three improvements have been made to increase the detection accuracy on strip steel surface defect, i.e., additional output layer for multi-scale detection, spatial pyramid pooling module for richer deep features, and feature pyramid and bottom-up path augmentation module for feature fusion. Experimental results demonstrate that the performance of the network model proposed in this paper is better than that of network models such as YOLOv3, YOLOv4, SSD and faster R-CNN.
No takes yet. Share an insight, caveat, or question.
Hao et al. (2021) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: