荧光磁粉探伤结果自动识别系统研究  

Study on Automatic Identification System in fluorescent magnetic detection

作　　者：张家栋[1] 张强[1] 霍凯[1] 

机构地区：[1]北京交通大学机械与电子控制学院,北京100044

出　　处：《铁路计算机应用》2009年第10期12-15,共4页Railway Computer Application

基　　金：高速动车组修程修制评价方法研究及评价系统的开发校基金(KMJ06012532)

摘　　要：针对目前荧光磁粉探伤人工观察方法存在的问题,对荧光磁粉探伤结果自动识别系统进行研究。搭建基于ARM的嵌入式Linux系统,应用USB摄像头进行图像采集。采用选定色彩范围的方法获取图像可疑成分,使用二值图像来表示图像的可疑成分与背景,主要使用数学形态学的方法进行图像处理,通过计算图像连通区域的大小和圆形度并与设定值比较来确认裂纹,识别结果有效。Aim at the existing problems of the manual inspecting in fluorescent magnetic detection, it was researched on the Automatic Identification System in fluorescent magnetic detection, built Embedded Linux Systems based on ARM, used USB camera for image acquisition, gained image suspicious ingredient by designates the color scope, used binary image to express the suspicious composition and background, mainly used mathematical morphology method to carry on the image processing, calculated the regional image＇s size and roundness then confirm the crack by compare them with the specified value. The result was effective.

关 键 词：荧光磁粉探伤 嵌入式LINUX 图像处理 数学形态学 

分 类 号：TP216[自动化与计算机技术—检测技术与自动化装置] TP39[自动化与计算机技术—控制科学与工程]