连通域泛水填充融合k-means裂缝图像滤波算法

Crack Image Filtering Algorithm Based on Connected Domain Flooding and K-Means Fusion

鉴于道路和隧道表面裂缝图像普遍存在光照不均匀、对比度低且混有大量噪声,不利于特征提取和后续分类等问题,提出了一种基于连通域泛水填充和K-means融合的裂缝图像滤波去噪方法。方法在对原始图像二值化处理的基础上,通过计算连通域面积大小,并对连通域面积进行K-means自适应聚类分析获取裂缝和噪声的特征分布,最后利用泛水填充滤波进行裂缝图像去噪。对比实验结果表明,上述方法在图像滤波去噪的同时能有效保留裂缝细节特征,算法性能明显优于传统算法。

In view of the common problems of uneven illumination,low contrast,and a large amount of noise in road and tunnel surface crack images,which are not conducive to feature extraction and subsequent classification,a crack image filtering algorithm based on the flooding of the connected domain and K-means fusion is proposed.Based on the binarization of the original image,this method calculates the area of the connected domain and performs K-means adaptive clustering analysis on the area of the connected domain to obtain the characteristic distribution of cracks and noise.Experimental results show that the proposed method can effectively retain the details of the cracks while filtering and denoising the image,and the performance of the algorithm is significantly better than that of the traditional algorithms.