摘要
Blob分析是机器视觉系统中的一个重要组成部分,为满足工业应用的实时性要求需研究快速高效标记算法.该文提出一种基于游程链的快速Blob分析法,使用游程链表和动态数组,建立相邻两行游程信息的动态链表,通过分析游程节点间连通性标记游程节点.该方法只需扫描图像一次,并且无需建立等价表和进行等价标记合并,克服了现有算法中的标记冗余,另外还能标记连通区域中的内洞.实验结果表明该方法对于任意复杂形状、任意数目的Blob区域都能正确检测并计算参数,并且具有很快的速度和很好的稳定性.该方法已成功应用到印刷缺陷在线检测系统中.
Blob analysis is an important part of machine vision system. Fast and efficient Blob analysis is needed to meet real-time demands in industrial applications. This paper presents a fast Blob analysis algorithm based on runlist. The algorithm uses the method of run-lists and dynamic array to build dynamic chain tables for storing two neighbor rows run-coding information. The run list node can be labeled by analyzing its connectivity. This algorithm requires only a single pass over the image, without the need to build an equivalence table and to unite equivalent labels. It can avoid label redundancies present in conventional algorithms. In addition, it can also label holes in blobs. Experimental results show that it can correctly label any Blob regions with complicated shapes and random numbers, and compute blobs features. With faster speed and good stability, the proposed labeling algorithm has been successfully applied in an on-line defect detection system for printed matter.
出处
《应用科学学报》
CAS
CSCD
北大核心
2008年第5期536-540,共5页
Journal of Applied Sciences
基金
陕西省自然科学基金(No.2006F26)
陕西省教育厅自然科学研究计划(No.07JK356)资助项目
