摘要
黄土的颜色与其含水率关系密切,而含水率可与其强度、变形参数建立经验计算模型。为此,本文以伊犁黄土为研究对象,通过制取原状样,配置不同含水率,并获取标准光源下的图像,通过ALV各向异性扩散滤波算法进行图像处理,提取土样的RGB值,研究了RGB值随含水率的变化规律。研究结果表明:ALV算法在孔隙、杂质滤噪与土样边缘保持方面效果显著;滤波后RGB值标准差降低70%以上,滤波后获取的RGB值更精确;RGB值均随含水率的增加而降低,降低过程存在显著的三阶段特征,不同阶段RGB值变化速率的土水浸润物理机制不同。本研究为建立黄土含水率图像识别模型提供了理论基础。
The color of loess is closely related to its moisture content,which can be used to establish an empirical calculation model for its strength and deformation parameters.Therefore,in this paper,Yili loess is taken as the research object,through the preparation of undisturbed samples,configuration of different moisture content,and obtain the image under the standard light source,through the ALV anisotropic diffusion filtering algorithm for image processing,extract the RGB value of soil sample,and study the change rule of RGB value with water content.The results show that:ALV algorithm has a significant effect on pore,impurity noise filtering and soil sample edge preserving;the standard deviation of RGB value after filtering is reduced by more than 70%,and the RGB value obtained after filtering is more accurate;the RGB value decreases with the increase of water content,and there are significant three-stage characteristics in the reduction process,and the physical mechanism of soil water infiltration is different in different stages of RGB value change rate.This study provides a theoretical basis for establishing image recognition model of loess moisture content.
作者
宋华杰
Song Huajie(Xi'an Railway Vocational and Technical Institute,Xi'an,Shaanxi 710026,China)
出处
《西安轨道交通职业教育研究》
2020年第4期28-33,共6页
Xi'an Rail Transit Vocational Education Research
关键词
各向异性扩散滤波算法
伊犁黄土
含水率
图像滤波
Anisotropic Diffusion Filtering Algorithm
Yili Loess
Water Content
Image Filtering
