基于证据权法的矿山地质环境质量评价模型研究

发布时间：2018-05-12 20:18

本文选题：矿山地质环境 + 评价模型　；参考：《中国地质大学(北京)》2015年博士论文

【摘要】：矿山地质环境评价是实施矿山地质环境监管、防治规划的主要依据,也是矿山地质环境研究中的重要内容。客观、定量、结果可重现的评价模型有助于提高评价结果的可信度和准确性。许多学者在该领域进行了较为深入的研究,但现有评价模型仍存在一定的局限性:一是忽视评价的尺度效应,将某一尺度下的评价结果直接外推到其他尺度,缺乏多尺度及最优评价尺度研究;二是评价指标选取及其权重的计算不同程度地依赖主观知识;三是评价结果的分级受主观因素干扰,降低了评价的可信度。为了解决上述问题,本文以赣南稀土矿山为研究区,综合应用遥感和GIS技术,针对研究区存在的主要地质环境问题进行矿山地质要素信息提取,以此建立地质环境评价因子体系。开展地质环境多尺度评价,基于研究区地表要素量化空间变异特性指标,提出一种最优评价尺度选择模型来客观量化地确定最优评价尺度。在此基础上,将证据权法引入矿山地质环境评价领域,发展一种基于数据空间位置关系来描述和理解评价指标与矿山地质环境之间相互作用,以贝叶斯定律为数学基础的评价方法。避免主观因素干扰的同时,定量计算指标权值和后验概率(地质环境质量定量评价值)。综合先验概率和后验概率——面积频率曲线的分级特征对评价结果进行分级。以实测数据为基础,计算证据权模型的正确率和预测率来完成模型验证与评价。以此形成一套完备的客观、定量、结果可重现的矿山区域地质环境质量评价模式。通过分析研究,获得以下几点主要认识和结论:(1)高分辨率遥感影像丰富的纹理细节信息,为精确提取矿山地表要素奠定了基础。在线性评价模型的前提下,完成赣南稀土矿山地质环境质量的多尺度评价研究,发现其尺度效应主要来自矿山地表要素和评价指标权重的空间分布异质性两个方面。基于地统计学理论,根据矿山地表要素的空间位置和属性计算多尺度区域平均标准差和全局Moran指数,分别作为单元格内均质性和单元格间异质性的度量参数,并构建最优评价尺度模型。结果表明,202米是评价单元格内均质性和单元格间异质性同时达到最佳的均衡距离,也是开展研究区地质环境评价的最佳尺度。(2)根据各证据层的对比度值和条件独立性检验结果,优选证据层,计算后验概率。结果表明,矿产开发活动和植被覆盖度是对研究区地质环境质量影响最大的两个因素。基于后验概率——面积频率曲线和先验概率,将研究区地质环境质量划分为三个等级(好、一般和差),面积比例分别为:4.94%、7.11%和87.95%。(3)采用效率曲线法,用实测数据对模型进行定量验证和评价。正确率和预测率分别为88.7%和92.2%,表明本文所建立的模型在矿山地质环境质量评价方面具有较好的精度和适用性。
[Abstract]:The evaluation of mine geological environment is not only the main basis for the implementation of mine geological environment supervision and prevention planning, but also an important content in the study of mine geological environment. Objective, quantitative and reproducible evaluation models can improve the reliability and accuracy of evaluation results. Many scholars have done more in-depth research in this field, but the existing evaluation models still have some limitations: first, ignore the scale effect of evaluation, and extrapolate the evaluation results directly to other scales. Lack of multi-scale and optimal evaluation scale research; second, the selection of evaluation indicators and the calculation of their weights depend on subjective knowledge to varying degrees; third, evaluation results are affected by subjective factors, which reduces the credibility of evaluation. In order to solve the above problems, this paper takes Gannan rare earth mine as the research area, synthetically applies remote sensing and GIS technology to extract the information of mine geological elements in view of the main geological environment problems in the study area. Based on this, the factor system of geological environment evaluation is established. The multi-scale evaluation of geological environment is carried out. Based on the quantization of spatial variation characteristics of surface elements in the study area, an optimal evaluation scale selection model is proposed to determine the optimal evaluation scale objectively and quantitatively. On this basis, the paper introduces the evidence right method into the field of mine geological environment evaluation, and develops a kind of data spatial position relation to describe and understand the interaction between the evaluation index and the mine geological environment. An evaluation method based on Bayesian law. While avoiding the interference of subjective factors, the weights and posterior probabilities of the indexes are calculated quantitatively (the quantitative evaluation value of geological environment quality). The evaluation results are classified by combining the classification characteristics of the priori probability and the posterior probabilistic-area frequency curve. Based on the measured data, the correct rate and prediction rate of the weight of evidence model are calculated to complete the verification and evaluation of the model. Thus, a set of objective, quantitative and reproducible evaluation models of regional geological environment quality of mines are formed. Based on the analysis and research, the following main conclusions are obtained: 1) High resolution remote sensing images are rich in texture details, which lays a foundation for accurate extraction of surface elements in mines. On the premise of linear evaluation model, the multi-scale evaluation of geological environment quality of rare earth mines in south Jiangxi is completed. It is found that the scale effect mainly comes from the spatial distribution heterogeneity of surface elements and evaluation index weights of mines. Based on geostatistics theory, the average standard deviation of multi-scale region and global Moran exponent are calculated according to the spatial position and attributes of mine surface elements, which are used as measurement parameters of cell homogeneity and cell heterogeneity, respectively. The optimal evaluation scale model is constructed. The results show that 202m is the best equilibrium distance to evaluate the homogeneity and heterogeneity of the cells, and is the best scale to evaluate the geological environment of the study area.) according to the contrast value of each evidence layer and the test result of conditional independence, Select the evidence layer and calculate the posterior probability. The results show that the mineral exploitation activity and vegetation coverage are the two most important factors affecting the geological environment quality in the study area. Based on the posterior probability-area-frequency curve and priori probability, the geological environment quality in the study area is divided into three grades (good, general and poor, the area ratio is 7.11% and 87.95.3, respectively) and the efficiency curve method is used. The model was quantitatively validated and evaluated with measured data. The accuracy and prediction rate are 88. 7% and 92. 2% respectively, which shows that the model established in this paper has good accuracy and applicability in the evaluation of mine geological environment quality.
【学位授予单位】：中国地质大学(北京)
【学位级别】：博士
【学位授予年份】：2015
【分类号】：X82;TD167

【相似文献】