基于自组织神经网络和K-means算法的地下空间地质环境质量三维分类及评价  

3D Classification and Evaluation of Underground Space Geological Environment Quality Based on SOM and K-means Algorithm

作　　者：熊芸莹 李晓晖[1,2] 袁峰[1,2] 卢志堂[1] 吴少元 窦帆帆 XIONG Yun-ying;LI Xiao-hui;YUAN Feng;LU Zhi-tang;WU Shao-yuan;DOU Fan-fan(School of Resources and Environment Engineering,Hefei University of Technology,Hefei 230009,Anhui,China;Anhui Province Engineering Research Center for Mineral Resources and Mine Environments,Hefei University of Technology,Hefei 230009,Anhui,China;Xiamen Institute of Geological Engineering,Xiamen 361008,Fujian,China)

机构地区：[1]合肥工业大学资源与环境工程学院,安徽合肥230009 [2]合肥工业大学安徽省矿产资源与矿山环境工程技术研究中心,安徽合肥230009 [3]厦门地质工程勘察院,福建厦门361008

出　　处：《地球科学与环境学报》2023年第4期929-940,共12页Journal of Earth Sciences and Environment

基　　金：安徽省公益性地质工作项目(2023-g-1-18);安徽省自然科学基金项目(1808085QD116)。

摘　　要：针对地下空间地质环境质量,前人运用三维地质信息化技术已开展了大量三维综合评价研究,但其评价结果对于规划和施工建议略显不足。其原因主要是评价过程主观性较强,综合评价结果难以充分表达地质环境的真实类别,难以关注更需受到重视的不良地质环境条件等。针对上述问题,利用自组织神经网络(SOM)和K-means算法对地下空间地质环境质量三维评价信息进行分类研究;以福建省厦门市马銮湾新城南岸片区为实例,基于三维空间分析方法提取三维评价指标因子,开展基于自组织神经网络和K-means算法的地下空间地质环境质量三维评价,最后利用评价获得的地质环境类别与主导因子进一步提出规划和施工建议。结果表明:基于自组织神经网络和K-means算法的评价方法能够有效挖掘多维多源地质数据中的隐含信息,识别出关键区分因子,为地下空间地质环境质量评价提供了新的思路和方法。With the vigorous development of 3D geological information technology,many 3D comprehensive evaluation studies have been conducted on underground space geological environment quality,but the evaluation results are slightly insufficient for planning and construction suggestions.The main reason is that the evaluation process is subjective,comprehensive evaluation results are difficult to fully express the true category of geological environment,and it is difficult to pay attention to unfavorable geological environment conditions that require more attention.In response to the above issues,the self-organizing map(SOM)and K-means algorithm were used to classify and study 3D evaluation information of underground space geological environment quality.Taking the south bank of Maluan bay new city in Xiamen city,Fujian province as an example,3D evaluation index factors were extracted based on 3D spatial analysis method,and 3D evaluation of underground space geological environment quality was carried out based on SOM and K-means algorithm.Finally,the geological environment category and dominant factors obtained from the evaluation were used to propose planning and construction suggestions.The results show that the evaluation method based on SOM and K-means algorithm can effectively mine hidden information in multi-dimensional and multi-source geological data,identify key distinguishing factors,and provide new ideas and methods for the evaluation of underground space geological environment quality.

关 键 词：地质环境质量评价 地下空间 自组织神经网络 K-MEANS算法 聚类分析 地质建模 福建 

分 类 号：X141[环境科学与工程—环境科学] P642[天文地球—工程地质学]