基于遥感和机器学习的矿山资源储量预测模型研究  

Research on Mining Resource Reserve Prediction Model Based on Remote Sensing and Machine Learning

作　　者：王恩宇 WANG En-yu(The First Hydrogeological Engineering Geological Survey Institute of Anhui Provincial Geological Survey Bureau,Bengbu 233000,China)

机构地区：[1]安徽省地勘局第一水文工程地质勘察院,安徽蚌埠233000

出　　处：《世界有色金属》2025年第3期148-150,共3页World Nonferrous Metals

摘　　要：精准推算矿山资源储量对资源的管理与开采非常重要。传统地质寻找方式成本较高且准确率相对较低,本文提出一种新的矿山资源预测模型。这种模型将遥感技术与机器学习算法相结合,以此获得预估矿山资源的更高精确度与更低的成本。首先是利用遥感技术搜集矿山地表的各元素数据,接着使数据标准化,最后使用计算机的学习方法,如支持向量机(SVM)、神经网络(NN)等,对调整后的遥感数据执行建模与训练工作。通过构建训练集和测试集,及时优化模型参数,提升模型的预测能力。研究结果显示,基于遥感和机器学习的预测模型在预测精度上均优于传统的地质勘探方法。尤其是在资源量较大的矿山预测中,其误差降低了20%,表现出较强的适用性和优越性。本文为矿山资源储量的预估提供了新思路,同时此种方法还有望推广到其它自然资源的预测和管理之中,为我国的资源管理和可持续开发提供参考。Accurately calculating the reserves of mining resources explains its particularity for resource management and compliant mining.Examining traditional geological search methods,it was found that their cost is relatively high and accuracy is relatively low.Therefore,this article presents an innovative mining resource prediction model.This model combines remote sensing technology with machine learning algorithms,aiming to achieve higher accuracy and lower cost in estimating mining resources.In terms of research methods,the first step is to use remote sensing technology to collect data on various elements of the mining surface,and then make the data standardized.The second step is to use widely accepted computer learning methods,such as support vector machines(SVM),neural networks(NN),etc.,to model and train the adjusted remote sensing data.At the same time,by constructing training and testing sets,the model parameters can be optimized in a timely manner to improve the predictive ability of the model.The research results show that prediction models based on remote sensing and machine learning have better prediction accuracy than traditional geological exploration methods.Especially in mines with large resource reserves,the error is reduced by 20%,demonstrating strong applicability and superiority.This study provides a new perspective and approach,making it possible to estimate the reserves of mining resources.Meanwhile,this method is also expected to be extended to the prediction and management of other natural resources,providing reference value for resource management and sustainable development in China.

关 键 词：遥感技术 机器学习 矿山资源储量预测 SVM 神经网络 

分 类 号：TD167[矿业工程—矿山地质测量]