粉质粘土图像纹理参数的多元线性回归分析及其工程应用  

Multivariate linear regression and applications of textural parameters of silty clay images

作　　者：翟文信[1] 徐金明[1] 张学明 谢建强 

机构地区：[1]上海大学土木工程系,200072 [2]都市置业(中国),上海200001 [3]江苏省江阴市泰昌房地产开发有限公司,214400

出　　处：《水文地质工程地质》2009年第1期90-95,共6页Hydrogeology & Engineering Geology

基　　金：国家自然科学基金(40572162);上海市教委科技创新项目(09YZ39)

摘　　要：由二维小波技术分析粉质粘土图像的纹理特征,获得小波能量参数与粉质粘土工程性质指标的多元线性回归方程。在考虑拍摄条件下(光照、拍摄距离等),现场勘查并拍摄粉质粘土照片。将这些彩色照片转化为灰度图,在二尺度小波分解水平下得到反映粉质粘土图像纹理特征的9个能量参数,并将这些参数与对应土样的11个工程性质指标进行多元线性回归。在此基础上对2个土样的工程性质指标进行了预测。结果表明,本文提出的粉质粘土的小波能量参数与传统工程性质指标具有较好的对应关系,可以为现场快速确定粉质粘土的工程性质指标提供一个新的途径。By two-dimensional wavelet decomposition, the photograph texture of silty clay is analyzed in this paper and the multiple linear regression relations of the wavelet energy parameters and the traditional engineering properties of silty clays are constructed. During the image analysis, considering the requirements （such as light intensity, the distance between the camera lens and the section of silty clays） in taking images, colorful photographs of the clay were captured by a commercial camera in field investigation and transformed into grayscale images. The nine energy parameters, considering the elimination of the unfavorable influences of capturing condition, were obtained from the images. A regression technique was applied to estimate the relations between the nine energy parameters and the eleven traditional engineering properties of silty clays, traditional engineering properties in two samples are calculated out by the multiple linear regression equations. The results show that these relations are much satisfactory. So the method developed in this paper may provide a new technique for quickly examining the engineering properties of silty clay in field.

关 键 词：粉质粘土 纹理 小波变换 能量参数 多元线性回归分析 

分 类 号：TU411.3[建筑科学—岩土工程]