摘要点击次数: 167  全文下载次数: 0    查看全文  查看/发表评论  下载PDF阅读器
基金项目:湖南省自然资源科技计划项目“基于遥感地球化学方法的地表成矿元素丰度定量反演研究”(2020-04)
作者单位
孔华,黄慧坤,成功 中南大学地球科学与信息物理学院 
摘要:
      地球化学勘探找矿效果较好,但成本高、周期长;遥感地球化学利用遥感技术解决地球化学问题,遥感技术与地球化学方法相结合,可以降低找矿成本,提高效率。以新疆齐石滩金矿为研究区,将金矿成矿相关元素的化探数据与该区的WorldView-2遥感数据相结合进行遥感地球化学研究。采用多元线性回归(MLR)法建立各元素的反演模型,但反演结果精度不高,且异常位置和异常形态也存在一定的偏差,可能是混合像元致使含量与遥感影像呈非线性关系。因此,引入BP神经网络方法建模,反演精度明显提高,各元素反演异常与化探异常吻合很好,异常位置和异常形态更加精确。可见,BP神经网络方法更适合于遥感地球化学建模。
关键词:遥感地球化学  WorldView-2  定量遥感反演  多元线性回归  BP神经网络  金矿  新疆齐石滩
Abstract:
      Geochemical exploration method is a good method for ore prospecting, but it has high cost and long period. Remote sensing geochemistry is the combination of remote sensing technology and geochemistrical method to solve geochemical problems, which can reduce prospecting cost and improve efficiency. Taking the Qishitan gold deposit in Xinjiang as the study target, the geochemical data of elements related to gold mineralization was combined with WorldView-2 remote sensing data to carry out remote sensing geochemical research. Multiple linear regression (MLR) method was used to establish the inversion model of each element, but the accuracy of the inversion results was not high, and there were some deviations in the abnormal position and shape, which maybe resulted from the nonlinear relationship between the content and the remote sensing image caused by mixed pixels. Therefore, the BP neural network modeling method was introduced, which significantly improved the inversion accuracy. The inversion anomaly of each element is well consistent with the geochemical anomaly, and the anomaly location and shape are more accurate. Therefore, BP neural network method is more suitable for remote sensing geochemical modeling.
Keywords:remote sensing geochemistry  WorldView-2  quantitative remote sensing inversion  multiple linear regression  BP neural network  gold deposit  Qishitan in Xinjiang
孔华,黄慧坤,成功.结合化探数据的齐石滩金矿区遥感地球化学建模[J].地质学刊,2022,46(2):111-119