基于 OpenCV距离变换的物体计数系统研究  

Research on object counting system based on openCV distance transform

作　　者：张黎黎 王军民[1] 陈璨 Zhang Lili;Wang Junmin;Chen Can(Institute of Geophysics and Petroleum Resources,Yangtze University,WuHan 430100,China)

机构地区：[1]长江大学地球物理与石油资源学院

出　　处：《电子测量技术》2019年第21期97-100,共4页Electronic Measurement Technology

摘　　要：针对大量物体计数依靠人工会漏数,或者重复,生产效率相对较低的问题,系统提出将OpenCV距离变换与Blob算法相结合实现对物体计数,首先以一定数量硬币的图片为实验对象,对图像进行滤波,阈值处理,采用数学形态学中的方法开运算,来调整分割区域的形状,其次基于欧几里得算法距离变换等分割出物体,后采用Blob算法识别目标物体,进而进行数量统计。此次硬币测试目标共24个,最终本系统计数结果也是24个。系统计数准确,有良好的图形界面,具有一定的使用价值。Aiming at the problem that a large number of object counting depends on the leak number of artificial union,or repetition,and the production efficiency is relatively low,the system proposes to combine OpenCV distance transformation with Blob algorithm to realize object counting.Firstly,the image is filtered and thresholded,and the shape of the segmentation area is adjusted by mathematical morphology.Secondly,objects are segmented based on Euclidean distance transform,and then the Blob algorithm is used to identify the target object,and then the number of statistics is carried out.This coin test target totals 24,and the final counting result of this system is 24.The system counts accurately,has a good graphical interface,and has a certain use value.

关 键 词：OPENCV 模式识别 BLOB算法 距离变换 图像处理 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]