边界感知SegFormer网络的阵列目标图像分割方法  

Array Target Image Segmentation with Boundary-Aware SegFormer Network

作　　者：吕扬 吴静静 庄祉珊 安聪颖 LYU Yang;WU Jingjing;ZHUANG Zhishan;AN Congying(School of Mechanical Engineering,Jiangnan University,Wuxi,Jiangsu 214122,China;Jiangsu Key Laboratory of Advanced Food Manufacturing Equipment and Technology,Wuxi,Jiangsu 214122,China)

机构地区：[1]江南大学机械工程学院,江苏无锡214122 [2]江苏省食品先进制造装备技术重点实验室,江苏无锡214122

出　　处：《计算机工程与应用》2024年第24期188-199,共12页Computer Engineering and Applications

基　　金：国家自然科学基金(62072416,61873246)。

摘　　要：针对工业场景下阵列目标图像存在非均匀背景、缺陷干扰和弱边缘导致目标分割精度低的问题,提出边界感知SegFormer网络的阵列目标图像分割方法。针对固定种子易受背景和缺陷干扰的问题,提出自适应种子搜索策略。该策略利用种子位置与目标定位精度的相关性构建种子分布热力图,并在热力图的引导下自适应搜索理想种子目标,实现阵列目标的高精度全局分割。设计边界感知SegFormer网络进行局部分割,利用递归门控卷积强调特征的长距离和高阶空间交互,改进的门控残差边界细化模块能够学习更丰富的边缘信息,同时引入混合损失函数加强对区域内部和边缘像素的监督,引导网络更好地学习目标边缘特征,提高边界分割精度。在自建晶粒数据集和语义分割数据集Cityscapes上的验证实验表明,提出的分割方法能在背景不均、缺陷污染、边缘对比度低的高分辨率阵列目标图像中完整精确地分割目标,并具有较高的实时性。Aiming at the problem of low target segmentation accuracy due to the existence of non-uniform background,defect interference and weak edges in array target images in industrial scenarios,an array target image segmentation method with boundary-aware SegFormer network is proposed.Firstly,an adaptive seed searching strategy is proposed for the problem that fixed seeds are susceptible to background and defect interference.This strategy uses the correlation between seed location and target positioning accuracy to construct a heat map of seed distribution,and adaptively searches for ideal seed targets under the guidance of the heat map to achieve high-precision global segmentation of array targets.Secondly,the boundary-aware SegFormer network is designed for local segmentation,using recursive gated convolution to empha-size long-range and higher-order spatial interactions of features,an improved gated residual boundary refinement module to learn richer edge information,and the introduction of a hybrid loss function to enhance the supervision of the region interior and the edge pixels,which guides the network to better learn the target edge features and improve the boundary segmenta-tion accuracy.Validation experiments on the self-built wafer dataset and the semantic segmentation dataset Cityscapes show that the proposed segmentation method is able to completely and accurately segment targets in high-resolution array target images with uneven backgrounds,defective contaminations,and low edge contrast with high real-time performance.

关 键 词：图像分割 阵列目标 自适应种子搜索 边界感知 SegFormer 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]