摘要
面向火车轮对磁粉探伤图像的典型裂纹图像,基于试验分析比较,提出了一套以Wiener滤波、迭代法全局阈值分割算法和采用不变矩特征描述方法为核心算法,火车轮对荧光磁粉探伤图像的裂纹识别流程,实现了典型裂纹的有效识别。
This paper puts forward a set of image recognition process which designs for the train wheel typical crack image of magnetic particle inspection images based on experiment analysis and comparison. The core algorithm of the process is Wiener filtering, iterative method global threshold segmentation algorithm and using the moment invariant feature description method. Thus, the process realized the effective identification of typical crack.
出处
《机械与电子》
2013年第12期44-46,共3页
Machinery & Electronics
基金
长安大学公路养护装备国家工程实验室开放基金项目(2013G1502048)
关键词
火车轮对
荧光磁粉探伤
图像识别
WIENER滤波
阈值分割
不变矩
train wheel
fluorescent magnetic
image recognition
Wiener filtering
threshold segmentatlon
moment invarlantn
