基于实时语义分割的红外小目标检测算法  

Infrared Small Target Detection Algorithm Based on Real-Time Semantic Segmentation

作　　者：邵斌 杨华 朱斌[1,2,3] 陈熠 邹融平[1,2,3] Shao Bin;Yang Hua;Zhu Bin;Chen Yi;Zou Rongping(College of Electronic Engineering,National University of Defense Technology,Hefei 230037,Anhui,China;State Key Laboratory of Pulsed Power Laser Technology,Hefei 230037,Anhui,China;Key Laboratory of Infrared and Low Temperature Plasma of Anhui Province,Hefei 230037,Anhui,China)

机构地区：[1]国防科技大学电子对抗学院,安徽合肥230037 [2]脉冲功率激光技术国家重点实验室,安徽合肥230037 [3]红外与低温等离子体安徽省重点实验室,安徽合肥230037

出　　处：《激光与光电子学进展》2023年第14期57-64,共8页Laser & Optoelectronics Progress

基　　金：国家自然科学基金(61307025)。

摘　　要：语义分割网络对图像进行像素级分类,相较于目标检测其对于目标的精准定位更有优势,因此在红外小目标检测中发挥着重要作用。针对红外小目标的特点,提出一种基于实时语义分割的红外小目标检测网络。该网络基于双分支特征提取结构,采用渐进式特征融合模块和改进的Dice损失函数,使红外小目标分割的速度与效果达到良好的平衡。实验结果表明,该算法在较小参数量和计算量的情况下相较于FCN、ICNet、BiSeNet V2、STDCNet、TopFormer等5种算法达到较高的精度,在实际采集的红外小目标数据集上,其推理帧率相较于传统的FCN提升44%,达到117 frame/s,且红外小目标的交并比相较于与其推理帧率相近的TopFormer提升49%,有利于语义分割在红外小目标检测的实际应用。Semantic segmentation network classifies images at the pixel level,which has more advantages for accurate target location than target identification,thus playing an essential role in infrared small target detection.According to the characteristics of an infrared small target,a novel infrared small target detection network based on real-time semantic segmentation is proposed.A good compromise between the speed and impact of infrared tiny target segmentation is achieved by the network,based on the dual branch feature extraction structure,using the progressive feature fusion module and enhanced Dice loss function.The experimental results demonstrate that the algorithm achieves high accuracy compared with five algorithms,namely FCN,ICNet,BiSeNet V2,STDCNet,and TopFormer for small parameters and calculation.The proposed algorithem is advantageous for the practical application of semantic segmentation in infrared small target detection because its reasoning frame rate on the actual collected infrared small target dataset is 44%higher than that of traditional FCN,reaching 117 frame/s,and the intersection and merging of infrared small targets are 49%higher than that of TopFormer with the similar reasoning frame rate.

关 键 词：图像处理 红外小目标 实时语义分割 双分支特征提取 渐进式特征融合 

分 类 号：P391.9[天文地球—地球物理学]