基于数字图像的混凝土桥梁裂缝检测技术

Crack Detection of Concrete Bridges Based Digital Image

应用visual C++6.0语言编制程序,研究了基于数字图像的桥梁裂缝检测方法,深入分析评价了图像灰度化、棋盘格角点求解像素率、滤波除噪、边缘检测等图像处理算法,实现了基于视频(或图像)的桥梁裂缝宽度计算和软件系统,并用15幅桥梁裂缝图像验证了其裂缝检测精度.结果表明:本文提出的裂缝识别方法能较好地用于钢筋混凝土桥梁和B类预应力混凝土桥梁的裂缝检测,当裂缝宽度大于0.3mm时,图像计算出的裂缝宽度值与实测值非常接近,相对误差在6%以内;当裂缝宽度为0.2～0.3mm时,相对误差在10%以内.

A crack detection method of concrete bridges was proposed by using digital image technolo- gy, and some image processing algorithms, including gray processing of image, calculation of pixel rate, image noise filter, and edge detection of crack, were analyzed and evaluated for the crack detection of con- crete surface. Then, a computer program was developed with visual C＋＋6.0 programming language to detect the cracks, which was tested with 15 cases of bridge video images. The results have indicated that this method can effectively detect the width of cracks for reinforced concrete bridges and the B type pres- tressed concrete bridges. The results have also showed that the relative error is within 6% for the more than 0.3 mm cracks and it is less than 10 ~ for the crack width between 0.2 mm to 0.3 mm.