中文摘要：

基于小波变换与Gamma校正提出一种新的数字图像处理技术,将其应用于堆积体的离散元模拟,可实现数值模型快速、准确的建立。该技术先基于二维小波理论对数字图像进行消噪预处理,并采用自适应Gamma校正方法以消除不同光照和环境对图像的影响,最后进行色彩空间转换以实现二值化处理,针对整个数字图像处理过程自行开发Wavelet-Gamma-Conversion（WGC）自动建模程序,可实现对堆积体图像的信息获取、分析处理并自动生成细观建模文件的全过程,为堆积体力学行为的模拟提供新的建模方法。以澜沧江古水水电站堆积体边坡为例,对实拍数码照片采用WGC建立其细观模型,运用双轴压缩数值模拟实验对其力学特性进行分析和研究。研究表明,堆积体中碎块石与土体相互压密形成共同受力体承担外力作用,由于碎块石的影响,出现明显的应变硬化阶段;随着土颗粒间黏结强度的提高,堆积体的初始弹性模量和峰值强度都有所提高。数值模拟反映出的规律与室内实验结果相一致。

英文摘要：

Based on the wavelet transform and Gamma correction, a new digital image processing technology （DIPT） is proposed. Then it is applied to the discrete element simulation of talus deposit; the numerical model can be established fast and correctly. As for the technology, the two-dimensional wavelet theory is used for the denoising of digital image, then the adaptive Gamma correction method is adopted to eliminate the impact of different lights and environments, and finally, the color space is conversed for the binary treatment. Based on the study of DIPT, a Wavelet-Gamma-Conversion（WGC） auto-modeling program is developed. It is able to realize the whole process： acquiring the information, analyzing and processing, establishing the micro-model files. A new modeling method is developed for the mechanical behavior simulation of talus deposit. As an example, some talus deposit data of Gushui hydropower station is used; and then the micromodel is established by the WGC from the digital photo on-site. The mechanical properties are studied by the numerical simulation of biaxial compression test. The results show that the gravels and soil mass compress with each other to form a corporate load-carrying body for sharing the external forces. The stress-strain curves show a strain-hardening stage evidently due to the impact of gravels. With the increasing of bond strength between soil particles, the initial elastic modulus and peak strength of talus deposit go up at a certain level. The law of the numerical simulation test is in agreement well with the results of laboratory test.