摘要
路面裂缝检测是道路维护工作的重要内容之一。分维数可以很好地表现裂缝的特征。在差分计盒方法计算图像分维数的基础上,提出了一种改进算法;并将其应用到路面裂缝图像的阈值化。再结合数学形态学中的膨胀、腐蚀、开和闭运算以及底帽变换,采用多个合适的结构元素,在阈值化的裂缝图像基础上进行裂缝骨架提取,并去除噪声和毛刺,将利用该方法得到的检测结果与利用最大类间方差法和Canny算子得到的结果比较,表明该方法能够很好的实现路面裂缝的分割和提取。分割提取的裂缝位置准确,抗噪性能改善,更具实用性。
Pavement crack detection is an important part of the road maintenance works.Fractal dimension as a significant feature can express cracks well.The differential box-counting(DBC) algorithm is widely used to calculate the dimension of image.To make it more practical,a modified DBC approach is presented and applied on pavement crack images.Then combined with four basic operations and bottom-hat transform of mathematical morphology,they were used to extract crack skeleton and eliminate noise with multiple structure elements.Compared to Otsu method and Canny operator,the results of the new approach are more accurate,much more practical and with good anti-noise performance.
出处
《科学技术与工程》
北大核心
2013年第23期6746-6750,共5页
Science Technology and Engineering
基金
国家自然科学基金(60873186)
西安市自然科学基金合同项目(CX125218)资助
关键词
路面裂缝
差分计盒维数
数学形态学
底帽变换
骨架提取
pavement cracks differential box-counting mathematical morphology bottom-hat transform skeleton extraction