C. Zhang and Y.-h. Cheng (PRC)
Tread wear, template match, wavelet analysis, threshold segmentation
During the driving, tyre tread will be gradually marked with abraded print in different shapes, also original or deformed texture as well. It is more effective and convenient for state evaluation to adopt graphic analysis on tyre texture image directly snapped from reality. However wear progress is accumulated, in early time its feature is not clear, and recognized till later. Considered that wavelet is sensitive to singular variance on pixel’s distribution, while threshold segmentation makes a quick, stable dealing with detecting texture, based on template match, the former two are incorporated to modify segmentation on dominant contour. Among them template match is deployed by similarity comparison. The instance shows that it could give more characteristics, e.g. mass, circumference, especially with a better adaptability to various resolutions.
Important Links:
Go Back
