一种基于自适应阈值的弹壳表面缺陷分割方法  

Method of Bullet Surface Defect Segmentation Based on Adaptive Threshold Algorithm

作　　者：苟文韬 诸洪[1] 刘彬[1] 谢蔚卿[1] 熊长江[1] 

机构地区：[1]中国兵器工业第五八研究所,四川绵阳621000

出　　处：《四川兵工学报》2015年第1期122-123,127,共3页Journal of Sichuan Ordnance

基　　金：国防基础科研(C1020110001)

摘　　要：由于材料表面不均匀和基于线阵相机采集图像的方式,图像的目标和背景的成像效果是动态变化的,不能用单一固定的阈值进行弹壳表面缺陷分割。针对该问题,结合弹壳表面图像的特点和灰度直方图峰值特性,提出了一种自适应确定阈值的方法,试验结果表明该方法对能准确对弹壳表面缺陷进行分割。Because of the uniform surface of the material and the mode of image acquisition based on linear CCD,the imaging results of objet and background in image are dynamicly changed,and the single fixed value could not segment the defects well. An adaptive threshold segmentation method was proposed based on the image characteristic and the peak value of gray histogram characteristic. The experiment results show the method is feasible and effective.

关 键 词：枪弹 表面缺陷 自适应阈值 灰度直方图 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]