中文摘要：

在分析桥梁结构表面裂缝图像特征基础上,结合改进Snake主动轮廓模型图像分割算法,融合距离传感器信息,提出一种综合高分辨率图像采集、裂缝敏感区域截取、孤立噪声去除、裂缝旋转和标记的裂缝提取算法,实现了交互式远距离混凝土桥梁结构表面裂缝的精确提取。分析结果表明：改进Snake主动轮廓模型图像分割算法误分率为3.89%,运算时间为153ms,试验对比裂缝宽度显微镜观测值,裂缝提取绝对误差小于0.05mm,该算法可满足工程检测需求。

英文摘要：

Based on the analysis of crack image features of bridge structure surface,with image segmentation algorithm of modified Snake active contour model combined with distant sensor information for accurate crack extraction,the authors proposed an integrated crack extraction algorithm combining with high resolution image acquisition,crack sensitive area clip,isolated noise removing,crack rotation and labeling.With the proposed algorithm,interactive and distant accurate crack extraction of concrete bridge structure surface was realized.The results indicate that the misclassification rate of the modified Snake model is 3.89% and the operation time is 153 ms;compared with observed values by microscope,the absolute error of crack extraction is less than 0.05 mm.The proposed algorithm can meet the engineering detection need.