欢迎访问《图学学报》 分享到：
图学学报

图学学报

• 图像与视频处理 • 上一篇    下一篇

基于图像畸变校正的车轮踏面区域提取

  

  1. 石家庄铁道大学电气与电子工程学院，河北 石家庄 050043
  • 出版日期:2018-06-30 发布日期:2018-07-10
  • 基金资助:
    国家自然科学基金项目(11372199，51405313，51208318)；河北省自然科学基金项目(A2014210142)

Wheel Tread Area Extraction Based on Image Distortion Correction

  1. Electrical and Electronics Engineering, Shijiazhuang Railway University, Shijiazhuang Hebei 050043, China
  • Online:2018-06-30 Published:2018-07-10

PDF (PC)

222

摘要： 为了实现踏面缺陷部位面积的精确测量，提出了一种基于图像畸变校正的车轮踏
面区域提取方法。首先采用改进的彩色图像标记分水岭初次分割侧视视角踏面曲面图像，然后
根据踏面曲面图像横向和纵向畸变的不同特点顺次进行平面横纵坐标几何校正并在此过程中完
成踏面区域的二次分割与视角转换，最后得到踏面正视视角矫正图像。实验结果表明，该方法
可以准确分割并校正侧视视角踏面图像，为实现踏面缺陷面积的精确计算打下基础。

关键词: 踏面损伤, 彩色图像标记分水岭算法, 畸变校正, 视角转换, 精确测量

Abstract: In order to realize accurate measurement of the area of tread defect, a new method of wheel
tread area extraction is proposed based on image distortion correction. Firstly, the improved colorful
image labeling watershed algorithm is adopted to segment the surface image of the side view. Then,
the geometric correction is perfomed according to different characteristics of the horizontal and
vertical distortion of tread surface image. In this process, the two segmentations and the perspective
transformation of tread area are completed. Finally, a correction image is obtained from the tread
angle. The experimental results reveal that this method can accurately segment and correct the side
view of the tread image, which lays a solid foundation for the accurate calculation of tread defect
area.

Key words: tread damage, colorful mage marked watershed algorithm, distortion rectifying;
perspective transformation, accurate measurement