面向自主水下机器人的深海污染物检测

打开文本图片集

Deep-sea pollutant detection for autonomous underwater robots

ZHANG Biao*， ZHU Zhenyang，XU Jiazhong （Heilongjiang Provincial Key Laboratory of Complex Intelligent System and Integration， Harbin University ofScience and Technology，Harbin15oo8O，China） * Corresponding author，E-mail： zhangbiao@hrbust. edu. cn.

Abstract：Marine debris constitutes a significant global environmental challenge.Autonomous underwater robots offer a potential solution for the removal ofdeep-sea debris.To enable rapid and accurate detection of such debris，this study presents a lightweight detection model，termed Deep-sea Debris YOLO （DebrisYOLO），developed using deep learning techniques. An initial deep-sea debris dataset was constructed based on the Deep-sea Debris Database provided by the Global Oceanographic Data Centre （GODAC）. Subsequently，an enhanced BiFPN feature fusion network was employed to reduce model parameters while enhancing background discrimination capabilities. A lightweight detection head was designed to decrease computational complexity，thereby improving the model’ s practicality and deployability for deepsea debris detection.Furthermore，the Wise-DIoU（Wise Distance Intersection over Union） loss function was introduced to mitigate the influence of low-quality samples，enabling more precise localization of deepsea debris.Data augmentation and adaptive color-balanced underwater restoration were applied to enrich the training dataset.Experimental results demonstrate that the proposed Debris-YOLO model achieves improvements of 1.3% and 1.6% in mAP0.5 and mAP0.5：0.95 ， respectively， compared to YOLOv8n，while reducing the number of parameters and GFLOPS by 48.2% and 36% ，respectively. Key words： computer vision； target detection；underwater robot；deep-sea debris

1引言

海洋垃圾的累积对海洋生态构成了相当大的威胁，已经成为了全球性的问题[1-2]。（剩余18127字）