基于灰度极小值的路面裂缝检测算法

Pavement Crack Detection Algorithm Based on Gray Minimum Value

为提高路面裂缝检测的识别准确率、提高检测效率和降低检测成本,提出一种基于灰度极小值的路面裂缝检测算法。算法在不对路面灰度图像做任何预处理和标注训练的前提下,依据路面裂缝灰度极小值直方图提取不同对比度的裂缝特征;根据不同对比度裂缝的灰度空间分布特点,设计不同对比度裂缝多方向激活重构算法;依据裂缝的线性特点设计伪裂缝去除算法,在保留裂缝细节信息的前提下去除伪裂缝。在公共数据集上进行测试,结果表明算法能够检测不同类型的路面裂缝,具有一定的实际应用价值。

In order to improve the accuracy of pavement crack detection,improve the detection effi-ciency and reduce the detection cost,a pavement crack detection algorithm based on gray minimum value is proposed.The algorithm extracts crack features with different contrast according to the histogram of the minimum gray level of pavement cracks without any preprocessing and labeling training.According to the gray spatial distribution characteristics of cracks with different contrast,a multi-directional activation re-construction algorithm for cracks with different contrast is designed.According to the linear characteristics of cracks,a pseudo-crack removal algorithm is designed to remove pseudo-cracks while retaining the detailed information of cracks.Tests on public data sets show that the algorithm can detect different types of pavement cracks and has certain practical application value.