An Adaptive Image Segmentation Approach for Metallographic Analysis

摘要：

Metallographic analysis is widely used in industrial production detection in recently years. Accurate segmentation of metallographic images is utmost important for metallographic analysis. In this paper, we develop an adaptive metallographic image segmentation approach combining some methods and technology, which achieve good result and can be well used in practical metallographic images analysis projects. Firstly, original color metallographic images are pre-processed for later image segmentation, containing convert it to gray images and then extract object blocks from the gray images using an adaptive threshold method, although some of them may be connected. After that we segment the connected object blocks to independent object blocks respectively using some methods and technology. Lastly, post-processing is used to achieve prefect result. Experiments show that the approach we developed can extract and segment object blocks accurately and have good robustness, real-time property and universality.

关键词： metatlographic analysis image segm-entation adaptive threshold

作者: Chuntao Li Bo Kang

作者单位: School of Automation Engineering, University of Electronic Science and Technology of China Chengdu, China

会议类型: 国际会议

会议名称: 2010 International Conference on Measurement and Control Engineering(2010年IEEE测量与控制工程国际会议 ICMCE2010)

会议地点: 成都

会议语种:英文

页码: 267-269

在线出版日期: 2010-11-16（万方平台首次上网日期，不代表论文的发表时间）