P-tile与直方图模糊C-均值聚类相结合的路面图像阈值分割

Threshold Segmentation of Road Images Based on the Combination of P-tile and Histogram-Based Fuzzy C-Means Clustering

路面图像裂缝自动检测技术是公路养护技术的重要方向,路面图像的分割是路面图像处理的关键步骤。由于噪声等干扰因素的影响,使得利用传统的模糊C-均值聚类(FCM)算法进行路面图像分割得不到满意的结果。本文采用P-tile算法和直方图模糊C-均值聚类算法对路面图像进行分割,一方面克服了传统FCM运算量大、计算速度慢的缺点,另一方面减少分割算法分析的范围,增强了分割的效果。实验证明,本文算法能较好地分割出路面图像的裂缝。

The technology of road cracks＇ automatic detection is an important development trend of the road protection technology. The image segmentation of road is the key to the road image processing steps. Because of the influence of noise, the traditional fuzzy C-means clustering （FCM） algorithm for image segmentation has no satisfied results. In this paper, we combine the P-tile algorithm and the Histogram-based fuzzy C-means clustering algorithm to segment the images of road. On the one hand,the algorithms overcome the characteristics of large amount of computing and the shortcomings of slow speed of the traditional FCM; on the other hand, they reduce the scope of segmentation, and enhances the segmentation effectiveness. The experimental results show that using the proposed algorithm in this paper can better segment the cracks in the road images.