Year: 2024
Type: conference-paper
Abstract: With the rapid development of computer vision technology, the application of deep learning models in real-time object detection is becoming increasingly widespread. This article proposes an improved YOLOv5 model for ship detection at sea. By introducing the C3_SCConv module and SKAttention mechanism, the feature extraction and fusion capabilities of the model have been enhanced. Extensive experiments on both daytime and nighttime ship datasets have shown that the improved model outperforms the original model in terms of detection accuracy, recall, and mean accuracy (mAP), especially when dealing with multi-scale targets. The research results of this article provide an efficient and reliable solution for maritime vessel inspection, which has important practical application value.
Authors: Hanzhi Cui, Qiuxue Ouyang, Hui Deng, Conghan Zhong, Shu Yan, Xiangdi Wei
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
OpenAlex ID: W4409095264