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

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.