钢桥桥面细小裂缝图像特征动态识别仿真

Dynamic Recognition Simulation of Image Features of Small Cracks on Steel Bridge Deck

钢桥桥面颗粒纹理及自然环境等外界因素,严重影响了对细小裂缝的识别,为了能够有效解决以上问题,提出钢桥桥面图像细小裂缝最优化动态识别方法。通过Mask匀光法和灰度级腐蚀法预处理钢桥桥面图像,利用模糊C均值聚类算法粗提取钢桥桥面图像中的细小裂缝,采用非线性中值滤波、Sobel算子和Otsu法改进Canny算子,提取细小裂缝边缘。引入区域生长法识别细小裂缝,采用形态学方法去除细小裂缝图像毛刺和空洞,依据像素位置信息拼接细小裂缝,分析、识别不断更新的钢桥桥面图像,实现钢桥桥面图像细小裂缝最优化动态识别。实验结果表明,所提方法的裂缝识别准确性高、平均绝对误差在0.015以下、且运行时间在0.5s以内。

Some external factors such as particle texture and natural environment on the deck of steel bridge seriously affect the recognition for small cracks.In order to effectively solve this problem,an optimized dynamic identification method for small cracks on the deck of steel bridge was proposed.First of all,the image of steel bridge deck was preprocessed by Mask dodging algorithm and gray-scale corrosion method.Then,fine cracks in the image of steel bridge deck were roughly extracted by fuzzy C-means clustering algorithm.Moreover,Canny operator was improved by using the nonlinear median filter,Sobel operator and Otsu method.Meanwhile,the edge of fine crack was extracted.After that,the regional growth method was adopted to identify the small cracks,and the morphological method was used to remove the burrs and cavities in the image.Furthermore,the small cracks were spliced according to the position information of pixel.Finally,the constantly updated steel bridge deck image was analyzed and recognized,thus achieving the optimal dynamic identification for small cracks in the image of steel bridge deck.Experimental results prove that the proposed method has high accuracy in crack identification.Meanwhile,the mean absolute error is less than 0.015,and the running time is less than 0.5s.