一种基于方向特征的裂缝提取算法

A Cracks Extraction Algorithum Based on Directional Characteristics

针对传统算法提取精度不高和稳定性较差的问题,提出一种基于裂缝方向特征的提取算法.该算法由方向线的提取和修正两部分组成.(1)通过L(1)～L(8)共8个方向模板匹配,提取出8*10单位区域像素的裂缝方向线;(2)依据裂缝的强度和连续性,对该区域及相邻区域的方向线进行筛选和修复.实验结果表明,该算法能在照明阴暗,对比度较差的场合下稳定有效的提取出清晰裂缝,效果满意.

Aiming at the traditional algorithm for the extraction accuracy is not high and poor stability, this paper proposes a method for cracks extraction based on directional features. The algorithm is composed of the direction extraction and orientation correction. （1） Through the total of 8 direction template （L （1） - L （8）） matching, the crack is extracted direction lines in the 8 ~ 10 regional pixels. （2） The direction of line in the region and its adjacent areas is filtered and repaired, according to the strength and continuity of cracks. The experimental results show that the algorithm can stabily and effectively extract clear cracks in the dark and poor occasions. Finally, the effect is satisfied.