岩石颗粒与孔隙系统数字图像识别方法及应用  

Digital image recognition method of rock particle and pore system and its application

作　　者：刘春[1,3] 许强 施斌[1] 顾颖凡[1] LIU Chun;XU Qiang;SHI Bin;GU Ying-fan(School of Earth Sciences and Engineering, Nanjing University, Nanjing 210023, China;State Key Laboratory of Geo-hazard Prevention and G-co-environment Protection, Chengdu University of Technology, Chengdu 610059, China;Nanjing University （Suzhou） High-tech Institute, Suzhou 215123, China)

机构地区：[1]南京大学地球科学与工程学院,江苏南京210023 [2]地质灾害防治与地质环境保护国家重点实验室(成都理工大学),四川成都610059 [3]南京大学(苏州)高新技术研究院,江苏苏州215123

出　　处：《岩土工程学报》2018年第5期925-931,共7页Chinese Journal of Geotechnical Engineering

基　　金：国家自然科学基金项目(41230636,41302216);国家杰出青年科学基金项目(41225011);江苏省自然科学基金青年项目(BK20170393)

摘　　要：以砂岩薄片微观图像为例,研究了岩石颗粒与孔隙系统数字图像识别、定量化和统计分析方法。通过多颜色分割和去杂等操作获得二值图像;提出改进的种子算法来封闭特定直径的孔喉,并自动分割和识别不同的孔隙和颗粒;引入了概率统计的方法,实现了由二维颗粒面积计算颗粒系统的三维分选系数;使用概率熵和分形维数分别来描述颗粒和孔隙的定向性和形状复杂度的变化等。在理论研究基础上,研发了孔隙（颗粒）及裂隙图像识别与分析系统（PCAS）,通过简单操作即可得到颗粒和孔隙各种几何参数和统计参数。实例中,将PCAS应用于高孔隙度砂岩压密带微观形成机制研究。图像分析结果表明：（1）颗粒和孔隙得到了准确地识别量化,统计参数有效地描述和区分了不同的砂岩微观结构;（2）与V型压密带相比,产生平直型压密带的砂岩具有较大的孔隙度和平均孔隙面积,以及较好的颗粒分选性。研究表明砂岩压密带形态与其微观结构紧密联系,所采用的微观结构识别和分析方法可行。The thin-section micro images of sandstone are used as an example to investigate the image recognition, quantification and statistical analysis methods for rock particle and pore system. A binary image is obtained by using multi-color segmentation and spot removal operations. An improved seed algorithm is proposed, by which the pore throats with certain diameter can be closed automatically so as to divide and identify different pores and particles. A probability statistical method is introduced to calculate a three-dimensional sorting coefficient based on two-dimensional particle areas. The probability entropy and fractal dimension are used to describe the directionality and the variation of shape complexity, respectively. On the basis of theoretical researches, the software ＂Pore（Particle） and cracks analysis system＂（PCAS） is developed. In the example, PCAS is applied in the investigation of formation mechanism of compaction bands in porous sandstone. The image processing results show that：（1） The particles and pores are recognized and quantified accurately, and micro structures of different sandstone can be described and distinguised effectively by using the statistical parameters;（2） In comparison with that with chevron compaction bands, the host rock with straight compaction bands has greater porosity, average pore area and better grain sorting. The research indicates the formation of compaction bands in sandstone is closely related to the micro structures, and the recognition and the analysis methods of micro structures are reliable.

关 键 词：微观结构 图像处理 自动识别 PCAS 砂岩 

分 类 号：TU43[建筑科学—岩土工程]