Research On Scrap Steel Evaluation Technology Based On Faster-RCNN

Conference: ICMLCA 2021 - 2nd International Conference on Machine Learning and Computer Application
12/17/2021 - 12/19/2021 at Shenyang, China

Proceedings: ICMLCA 2021

Pages: 4Language: englishTyp: PDF

Personal VDE Members are entitled to a 10% discount on this title

Authors:
Qin, Yuman; Chen, Wangyiyang; Zhang, Peiran; Zhang, Shunxiang; Chen, Sihan; Huang, Liping (Software College, Northeastern University, Shenyang, China)

Abstract:
The identification and evaluation of scrap steel is of great significance to assist steel mills in scrap acceptance. However, at present, China's scrap steel sorting and recycling work is mainly completed manually, with subjectivity and uncertainty. Most of the existing domestic technologies do not take into account the complex actual situation of waste steel mixing such as multiple types and shapes in the actual sorting project. Therefore, a scrap recognition and evaluation technology based on target detection is proposed in this paper. After using vgg16 as the main network for feature extraction, input the feature map into the region proposal network to generate the candidate box, and then use the maximum pooling method to obtain the fixed dimension features, which are sent to the full connection network to predict the classification of each ROI. The experiment on the scrap dataset shows that the accuracy of the scrap evaluation technology based on Faster-RCNN can reach 87.3%, which can provide a better solution for the automatic identification and evaluation of scrap.