基于分形几何理论的公路路面初生裂纹辨识策略与定量评价机理

Road pavement crack is one of the most important factors that affect road traffic safety, ride comfort and economics. It has important theoretical significance and wide-range application prospects to solve the increasingly serious road foundation facility maintenance issues with the fractal geometry identification and quantitative evaluation mechanism study of the road pavement newborn cracks. The research contents include:① In terms of crack inspection scale, the optimal objective function is constructed with the constraint of newborn crack scale and with the optimization goals of the system parameters of structure, position and posture. The matching principle of system parameters is explored with the newborn crack scale. The optimal image correction transformation matrix is obtained. The optimal mathematical model of the pavement newborn crack inspection is established. ② In terms of crack identification method, the self-similarity characteristics of the pavement newborn cracks are studied deeply at different scales. The characterization and identification strategies of the self-similarity characteristics of pavement newborn cracks in road image are explored based on fractal geometry theory. ③ In terms of crack quantitative evaluation, the box dimension model and the Brown motion dimension model of the road newborn crack characteristics are built. The new accurate measurement methods of road pavement cracks with fractal information dimension are disclosed to realize the pavement crack shape quantitative evaluation at different spatial scales. The project research will lay the theoretical foundation for achieving accurate identification and quantitative evaluation of road pavement newborn cracks with high recognition rate, high reliability as well as providing the research proof with reference value.

公路路面裂纹是影响道路交通安全性、平顺性和经济性的重要因素，研究公路路面初生裂纹分形几何辨识与定量评价机理，对解决日益严峻的道路基础设施维护问题具有重要的理论意义和广泛的应用前景。研究内容包括:①裂纹检测尺度方面，构造以初生裂纹尺度为约束，以系统结构参数、位置参数和姿态参数为优化目标的最优目标函数，探明初生裂纹尺度的系统参数匹配规律，获得最佳图像校正转换矩阵，建立最优路面初生裂纹检测数学模型；②裂纹辨识方法方面，深入研究路面初生裂纹在不同尺度的自相似性特征，探明基于分形几何理论的路面初生裂纹自相似特征的表征与辨识策略；③裂纹定量评价方面，构建路面初生裂纹特征盒维数模型和布朗运动维数模型，探求全新的路面裂纹分形信息维数精确计量方法，实现不同空间尺度的路面裂纹形貌特征定量评价。项目研究将为实现高识别率、高可靠性的公路路面初生裂纹精确辨识与定量评价奠定理论基础并提供具有参考价值的研究依据。

公路路面裂纹是影响道路交通安全性、平顺性的重要因素之一。研究公路路面初生裂纹分形几何辨识与定量评价机理，可以为公路路面预维护提供充分的先进技术保证和理论支撑。①公路路面初生裂纹最优视觉测量模型的建立。首先，建立了基于结构光主动视觉的车载路面裂纹检测系统。其次，建立了对路面裂纹检测系统的摄像机及线结构光标定的闭合解解算模型和优化解解算模型。再次，根据线结构光投影完成了路面平面及路面坐标系在车载摄像机坐标系的重建。最后，采用重投影误差对基于车载主动视觉的路面检测系统的精度进行了试验评价。②基于分形几何理论的路面图像裂纹自相似性特征的辨识研究。首先，完成了对路面图像的滤波去噪、图像增强等预处理。然后，基于分形布朗运动模型对路面裂纹图像进行了分形维数的计算并根据分形维数完成了路面裂纹分割、形态学处理及骨架提取。最后，完成了采用分形维数分割方法辨识路面裂纹骨架的对比试验。③基于分形几何理论的路面裂纹形貌特征定量评价。首先，研究了采用分形方法对路面裂纹进行定量评价的方法。其次，基于计盒维数的分形方法对原始路面裂纹骨架和真实尺度路面裂纹骨架分别进行了计量。最后，基于计盒维数对单一路面裂纹、带分叉的路面裂纹与网状路面裂纹等进行了理论分析与试验研究，得到了不同形状路面裂纹分形维数的分布规律。研究对对延长公路的使用寿命、提高道路的通行能力并降低公路的养护成本具有重要的理论意义和广泛的应用前景。