基于机器视觉的种子质量检测新型算法

在分析已有区域标记算法的基础上,提出了一种新的二值图像连通区域准确标记算法。顺序扫描和标记二值图像的各个像素点,准确判断标记过程中出现的标记冲突,并建立标记冲突的模型,在算法中增加回溯扫描算法,消除标记冲突引起的标记误差。实验证明:该算法可以准确地标记出各种形状的连通区域,与已有算法相比,扫描重复率低,运行准确,速度快,具有很好的应用前景。

A New Arithmetic of Seed Quality Checking Based on Machinery Vision

The basic characteristic of current labeling algorithms were analyzed and a new algorithm for exact labeling of connected regions in binary images was put forward.First,scan and label all the pixel points in turn and judge the labeling conflicts exactly occurring during the labeling source and build the models of them,moreover add the rescan algorithm in order to eliminate the labeling errors aroused by label conflicts.Experimental result shows that the algorithm can label connected regions of any shape exactly and compared with the current labeling algorithms,this one has lower rate of repeated scan and higher efficiency and more exact veracity in speed,so it has wider application.