Efficient Detection of Grab Based on Improved Lightweight MC-YOLOv5s
Automated bulk terminals can enhance the efficiency and safety of bulk unloading operations. Hence, improving the accuracy and speed of grab detection is important for realizing the automation of bulk terminals. To achieve efficient detection of grabs, the YOLOv5s backbone network was reconstructed through the use of the lightweight network MobileNetV3 to improve the feature extraction speed. By introducing the CASPP structure, an improved lightweight detection method named MC-YOLOv5s was proposed. The experimental results indicated that the accuracy of the improved model was comparable to that of YOLOv5s, but the parameters and computation volume were reduced by 44% and 64% respectively. In addition, MC-YOLOv5s performed better on the extended dataset.
Authors
- Peng Ji (ORCID: https://orcid.org/0000-0003-1868-5004)
- Xiang Bian
- Cong Zhou
- Rongji Cui
Institutions
- Suzhou University of Science and Technology (CN)
- Jiangsu University of Science and Technology (CN)
- Shazhou Professional Institute Of Technology (CN)
- Institute of Microelectronics (CN)
Publication Details
- Published
- 2025-06-06
- DOI
- https://doi.org/10.1109/raiie65740.2025.11140559
- Primary Topic
- Advanced Neural Network Applications
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- Natural Science Foundation of Jiangsu Province