基于微分方程的桥梁裂缝检测研究  

Bridge Crack Detection Based on Differential Equations

作　　者：王家伟[1] 杨庭庭[1] 蒋伟[2] 

机构地区：[1]重庆交通大学信息科学与工程学院,重庆400074 [2]重庆交通大学理学院,重庆400074

出　　处：《重庆交通大学学报（自然科学版）》2014年第5期32-36,共5页Journal of Chongqing Jiaotong University(Natural Science)

基　　金：国家自然科学基金项目(61004118;11071266);重庆市科技攻关项目(2010AC2077)

摘　　要：为避免传统人工检测桥梁裂缝耗时、危险、花费高等缺点,将分数阶微分理论和现有的Sobel算子相结合应用于图像边缘检测中,得到了一种基于分数阶微分的桥梁裂缝检测新模型。理论分析和仿真实验结果表明:与现有的边缘检测方法相比,该模型不仅能较好地检测出图像的边缘特征,而且能够检测出图像中更多的纹理细节信息。并且,连接度、错检率和漏检率等客观评价指标说明,它是一种有效的边缘检测方法,能够较好地达到桥梁裂缝检测的目的。Aiming at the shortcomings of time-wasting, high risk and cost in the traditional process of detecting the bridge crack, fractional differential theory and sobel operation were combined and applied into the image edge detection. Therefore, a novel bridge crack detection model based on fractional differential theory was obtained. The theoretical analysis and simulation results show that comparing with the current edge detecting methods, the proposed model not only can detect the edge features of the images better, but also can detect more texture detail information. Furthermore, the objective evaluation indexes, such as connectivity, false detection rate and missing rate, show that the proposed method is an effective edge detection method, which can better achieve the purpose of bridge crack detection.

关 键 词：桥梁工程 桥梁裂缝检测 分数阶微分 SOBEL算子 纹理细节 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]