摘要
对公路摩擦力的属性进行图像特征分析,有利于智能化的解决公路老化等问题。本文通过对公路路面不同摩擦参数路面图像纹理特征的分析,运用形态学高低帽转换对公路路面的图像进行预处理,剔除噪声图像,得到清晰的纹理特征。运用不完全小波树形结构剖析摩擦力属性对应的公路路面图像纹理等差异特征;分析得到不同的公路路面摩擦力与不同图像纹理特征的对应情况。实验结果表明,该方法特征索取和识别速率较传统的识别方法更快。尤其是对不清晰、场景环境复杂的路面图像纹理识别效果明显,很好的保证了路面图像纹理特征与对应摩擦力的识别准确性。
Texture Feature Model Highway ground friction properties correctly identify the problem.Factors that affect the image texture features based on the traditional identification methods ignore theenvironment in which the road surface (light, noise, size) result in recognition of road surface informationis not accurate, the paper on the road surface friction characteristics of different surface textureparameters were analyzed, test complete contrast. Using morphological level cap conversion of roadsurface image preprocessing to eliminate image noise. By getting clear texture characteristics, usingincomplete analysis of different wavelet tree structure corresponding to the road surface frictionproperties of texture image orientation and other differences in texture features; get a different analysisand different textures road surface friction characteristics corresponding to the situation. Experimentalresults show that the method is characterized obtain recognition rate and faster than other methods.Especially for unclear, complex environmental attributes scene texture image recognition better ensurethe accuracy of the identification image texture features.
作者
桑园
王莉丽
Sang Yuan;Wang Lili(Sias International University School of Electronics and Information Engineering,Xinzheng 451150,China)
出处
《科技通报》
北大核心
2016年第7期188-192,共5页
Bulletin of Science and Technology
基金
河南省教育厅科学技术研究项目(14A520059)
河南省科技攻关项目(152102310368)
