基于VC++路面裂缝图像处理系统研究

PAVEMENT CRACK IMAGE PROCESSING SYSTEM RESEARCH BASED ON VC++

传统的基于人工视觉检测裂缝的方法愈来愈不能适应高速公路发展的要求。为了满足这种需求,给出了一种基于图像处理的路面裂缝自动检测方法,使用一种快速的针对灰度不均图像校正方法,对降质路面图像进行增强,然后进行阈值分割,提取出图像上的线性特征,最后再根据线性特征从原来灰度图像上识别出裂缝目标。整个软件系统采用面向对象的编程技术,设计界面友好,且易于操作。将计算机图像处理结果与实际公路现场检测结果进行了对比分析,表明了该方法的可行性、准确性、有效性,并且有比较满意的检测效果。

Traditional pavement crack detection method on artificial vision basis can hardly fit the demand of highway development increasingly. In order to meet the demand, in this thesis it presented an automatic pavement crack detection method based on image processing. It enhances the degraded pavement image by using a fast rectification method against non-uniform greyscale image, then segments threshold and extracts linear feature of the image, and at last identifies the crack objective from original gray image according to the linear feature. Object-oriented programming technique is adopted in whole software system which makes it with friendly design interface and easy to operate. Computer im- age processing results were compared with actual highway on-site detection results, the analysis has shown the feasibility, accuracy and effectiveness of this method, and detection results were quite satisfied.