摘要
为了客观评估矿山生态修复效果,设计了一种基于随机森林(RF)算法的评估模型。该模型根据矿山生态用地面积、群落垂直结构、归一化植被指数(NDVI)、与周边地类一致性等17个评价指标,采用信息量法确定各指标的权重,利用随机森林算法对已修复矿山图斑进行分类。通过五折交叉验证(CV),检验了指标的合理性和模型精度。研究结果表明,该模型能够较好地反映重庆市已完成生态重建类矿山的修复成效,契合度可达80%,修复2年后的矿山图斑契合度可高达92%及以上。本文提出的基于随机森林(RF)算法的评估模型构建方法,可为重庆丘陵山区以及其他区域的矿山生态修复评估和治理提供参考。
In order to evaluate the effectiveness of ecological restoration of mines objectively,This article designed an evaluation model based on the Random Forest(RF)algorithm.According to 17 evaluation indicators,such as the area of mine ecological land,the vertical structure of community,the normalized difference vegetation index(NDVI),and the consistency with surrounding land types,the weight of each indicator was determined by the information volume method,and the repaired mine patches were classified by the RF.The reasonableness of the index and the accuracy of the model were tested by the k-fold cross validation(CV).The results show that the model can better reflect the effectiveness of the ecological restoration of mines in Chongqing,and the results are fit up to 80%.The fit rate of the restored mining pattern after 2 years can reach 92%or higher.The construction method of the evaluation model based on the Random Forest(RF)algorithm proposed in this article can provide reference for the ecological restoration assessment and governance of mines in the hilly and mountainous areas of Chongqing and other regions.
作者
王素伟
周川
巫长悦
谭利丽
李春利
王力
杨赟
罗冬
杨欢
文敏
刘凯
WANG Su-wei;ZHOU Chuan;WU Chang-yue;TAN Li-li;LI Chun-li;WANG Li;YANG Yun;LUO Dong;YANG Huan;WEN Min;LIU Kai(Chongqing Institute of Geology and Mineral Resources,Chongqing 401120,China;Field Scientific Observation and Research Station of Ecological Restoration for Chongqing Typical Mining Areas,Ministry of Natural Resources,Chongqing 401120,China;Chongqing Wansheng Mining Area Ecological Environment Protection and Restoration Field Scientific Observation and Research Station,Chongqing 401120,China)
出处
《煤炭科技》
2024年第1期1-5,10,共6页
Coal Science & Technology Magazine
基金
重庆市自然科学基金面上项目(CSTB2022NSCQ-MSX0280)
重庆市科研机构绩效激励引导专项(cstc2021jxjl20001)。
关键词
矿产资源
生态修复
成效评估
模型应用
mineral resources
ecological restoration
effectiveness evaluation
model application
作者简介
王素伟(1987—),男,山西太原人,高级工程师,2011年毕业于西南大学地理信息专业,主要从事矿山生态修复、地理信息系统等方面的研究;通讯作者:周川(1987—),女,重庆人,高级工程师,主要从事矿山生态修复等研究工作。
