基于YOLO-v5的星载SAR图像海洋小目标检测  

Small marine target detection in space-borne SAR image based on YOLO-v5

作　　者：窦其龙 颜明重[1] 朱大奇[1] DOU Qilong;YAN Mingzhong;ZHU Daqi(Shanghai Engineering Research Center of Intelligent Maritime Search&Rescue and Underwater Vehicles,Shanghai Maritime University,Shanghai 201306,China)

机构地区：[1]上海海事大学智能海事搜救与水下机器人上海工程技术研究中心,上海201306

出　　处：《应用科技》2021年第6期1-7,共7页Applied Science and Technology

基　　金：国家自然科学基金项目(U1706224,62033009).

摘　　要：针对星载合成孔径雷达(SAR)图像中的小目标检测,提出了一种基于YOLO-v5的算法。首先根据目标在图像中占比很小的特点,对深度学习网络进行优化;其次通过自适应锚点框算法重新设置锚点框大小,加快模型训练的收敛速度;最后嵌入GDAL模块,使得检测头能够直接读取被检测目标的地理位置信息。实验结果表明本文提出的方法具有较强的泛化能力,能够有效检测出大范围海面上的船舶目标,检测速度较快,漏检率低。An algorithm based on YOLO-v5 is proposed for small target detection in synthetic aperture radar(SAR)images.Firstly,the deep-learning network is optimized according to the small proportion of the target in the image.Secondly,an adaptive anchor box algorithm is used to reset the anchor frame size to accelerate the convergence speed of the model training.Finally,the GDAL module is embedded to enable the detector to directly read the geographic location information of the detected targets.The experimental results show that the proposed method has a relatively strong generalization ability,which can effectively detect ship targets on a wide range of sea surfaces with relatively fast detection speed and low misdetection rate.

关 键 词：合成孔径雷达 星载SAR图像 深度学习 数据增强 自适应锚点框 YOLO-v5 小目标检测 GDAL 

分 类 号：TP389.1[自动化与计算机技术—计算机系统结构]