| Railroad Track Flaws Detection is one of the most important approaches to ensure the safety of railway. In China, railroad track flaws are tested by manual work till now. Therefore, developing an automated system based on image processing and analyses for railroad track flaws detection has great meanings. Compared with manual test, computer automatic test is more accurate and more feasible.In this paper, the requirements of this system and the real images are firstly analyzed, then, based on the analysis of many real railroad track images and the research work in edge detection field that have been done in recent years by Image Processing Institute of Xi' an Jiaotong University and Dalian Maritime University, the algorithms are brought forward as following:1.Railroad track region is segmented from the other regions by use of conversion, difference, projection.2.Discriminating circular flaws from lathy flaws by zero-crossings of ridge edges.3.Using different operators respectively to inspect two kinds of flaws, the LOG operator for circular flaws and the steerable filter for lathy flaws.4.Using step edge detection method to detect the crackles in the sidepiece of rail.5.Research of fastener test algorithm based on structure light.
|
PDF Full Text Request