基于HALCON图像处理的粘连零件颗粒计数方法研究  

Research on the method of particle counting for adhesive parts based on HALCON image processing

作　　者：黄明鑫[1] 杨龙兴[1] 庄立东[1] 梁栋[1] 

机构地区：[1]江苏理工学院机械工程学院,江苏常州213001

出　　处：《机械设计与制造工程》2016年第2期85-89,共5页Machine Design and Manufacturing Engineering

基　　金：江苏省科学技术厅产学研联合创新基金--前瞻性联合研究项目资助(BY2014038-01)

摘　　要：为了解决传统的粘连零件颗粒计数方法存在的效率低、准确性不高等问题,基于HALCON图像识别和处理技术,提出了一种基于欧式距离变换及高斯滤波的分水岭分割算法。首先,利用工业相机采集零件颗粒图像,然后对图像进行预处理,接着对预处理后的图像进行数学形态学处理,再对数学形态学处理后的图像进行距离变换和高斯滤波,最后进行分水岭分割,得到分割图像。实验表明,该算法能有效减少过分割,为后续零件颗粒的准确计数提供了保障。In order to solve the problem of low efficiency and low accuracy in the traditional method of particle counting for adhesive parts,it describes Euclidean distance transform and Gaussian filter,presents the watershed segmentation based on HALCON image recognition and processing technologies. It applies industrial camera to obtain the particle image. Based on the pre- processing of image,it realizes the image mathematical morphology processing,then uses Euclidean distance transform and Gaussian filter to process these data. The last step is the watershed segmentation and gets the segmentation image. The experimental results show that the algorithm can reduce the over- segmentation phenomena effectively,provide the guarantee for accurate count of subsequent parts particle.

关 键 词：粘连 HALCON 距离变换 高斯滤波 分水岭分割 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]