基于流形学习的路面破损图像多特征融合与可视化  

Multi-feature Fusion and Visualization of Pavement Distress Images Based on Manifold Learning

作　　者：石陆魁[1,2] 周浩[1] 刘文浩[1] 

机构地区：[1]河北工业大学计算机科学与软件学院,天津300401 [2]河北省大数据计算重点实验室,天津300401

出　　处：《公路交通科技》2016年第11期26-33,共8页Journal of Highway and Transportation Research and Development

基　　金：天津市应用基础与前沿技术研究计划重点项目(14JCZDJC31600);河北省自然科学基金项目(F2016202144)

摘　　要：针对路面破损图像自动识别中的多特征融合问题,提出了一种基于流形学习的多特征融合方法,利用流形学习方法将组合投影、混合密度因子和二阶不变矩3种特征的高维数据映射到低维空间中,提取出路面破损图像的本质特征,实现了多特征融合和路面破损图像的可视化。在试验中,将这种多特征融合方法应用在路面破损图像的检测中,首先从8个组合特征中提取出二维特征,然后比较ELM、KNN、SVM、BP神经网络在二维特征上的识别效果。试验结果表明:利用特征融合方法有效提高了路面破损图像的识别精度;通过可视化得到了二维特征的物理含义,一个特征初步表明了路面裂缝的复杂程度和破损程度,另一个特征给出了裂缝的方向。For multi-feature fusion in automatic recognition of pavement distress images, we proposed a multi-feature fusion method based on manifold learning. In this method, the intrinsic features of pavement distress images are extracted through mapping the high dimensional data combing projection, mixture density factor and second order moment invariant into the low dimensional space. The multiple features are fused and the visualization of pavement distress images is implemented. In the experiments, we applied the multi-feature fusion method in the detection of pavement distress images. Two-dimensional features are firstly extracted from the 8 combining features, then the recognition effects on the 2D features of 4 methods including ELM, KNN, SVM and BP network are compared. The experiment result shows that （ 1 ） the proposed method effectively improved the detection accuracy of pavement distress images;（2） the physical meaning of the 2D features is obtained through visualizing, one feature preliminary denotes the complexity and damaged extent of the cracks in images, the other describes the direction of the cracks.

关 键 词：道路工程 多特征融合 流形学习 路面破损图像 可视化 

分 类 号：U416.2[交通运输工程—道路与铁道工程] TP181[自动化与计算机技术—控制理论与控制工程]