轮胎气泡边缘检测算法的研究

Research on algorithm for tire bubbles edge detection

针对轮胎激光剪切散斑干涉图像背景中存在大量条纹的特点,为避免其对气泡边缘检测产生妨碍,主要对其边缘检测算法进行了研究。给出了对常用的几种算子进行边缘检测后的轮胎激光剪切散斑干涉图像,并分析了其在处理图像时的不适用性,设计出基于模糊理论的小波多尺度边缘检测算法。在多尺度边缘检测过程中,设计了适合的模糊算子,改善了轮胎气泡边缘检测的效果。由仿真结果可知:本文设计的算法与常用边缘检测算法相比,更好地标记出了轮胎气泡的边缘图像,滤除图像中的伪边缘。

Aiming at the characteristic of the tire laser shearography image, there are many stripes exist in the background. In order to avoid the influence of stripes on bubble edge detection, this paper is to research on its bubble edge detection algorithm. First give the tire laser shearography images after using several kinds of commonly edge detection operators, and analyze the unworthiness when used these operators to process the image. Then design wavelet multi-scale edge detection algorithm based on fuzzy theory. In the course of multi-scale edge detection has designed the suitable fuzzy operator, improved the effect of the tire bubbles edge detection. By the simulation results we know that the algorithm can mark the edge of bubble better than common ones, and it can remove the false edges of the image.