算法
计算机科学
碎片
遥感
计算机视觉
人工智能
地质学
特征(语言学)
算法设计
事件(粒子物理)
出处
期刊:
日期:2026-04-24
卷期号:: 245-248
标识
DOI:10.1109/icsi70302.2026.11544931
摘要
To address the challenges of detecting debris on riverbeds and the insufficient detection accuracy in urban river dredging operations, this paper proposes PSF-YOLO, a debris detection algorithm for side-scan sonar images based on an improved YOLOv11n model. To address the limitations of the SPPF module at the end of the YOLOv11n backbone network in capturing spatial dependencies at long distances, the LSKA attention mechanism is introduced to enhance the model’s ability to detect slender, low-texture objects and suppress interference from complex backgrounds. Experimental results show that PSFYOLO achieves a 2.8% increase in precision and a 3.5% increase in mAP compared to the original YOLOv11n. Compared to current mainstream algorithms, it achieves optimal performance in terms of precision, mAP, and inference speed, validating the effectiveness of the proposed approach.
科研通智能强力驱动
Strongly Powered by AbleSci AI