| 基于重磁数据梯度比值的深度学习技术实现场源位置反演方法 |
| |
| 引用本文： | 马国庆,吴琪,熊盛青,李丽丽.基于重磁数据梯度比值的深度学习技术实现场源位置反演方法[J].地球科学,2021,46(9):3365-3375. |
| |
| 作者姓名： | 马国庆 吴琪 熊盛青 李丽丽 |
| |
| 作者单位： | 1.吉林大学地球探测科学与技术学院, 吉林长春 130026 |
| |
| 基金项目： | "十三五"国家重点研发计划项目2017YFC0601606 |
| |
| 摘 要： | 场源中心位置的计算是重磁数据反演的主要任务之一，现主要通过异常与场源位置之间的数学物理方程来估算地质体的位置.为了快速、准确获得地质体的位置信息，提出基于重磁梯度比值的深度学习技术实现场源位置的获取；其利用深度学习技术所建立的重磁梯度比值水平分布与地质体埋深、构造指数的关系，快速实现异常场源位置计算，且提出利用多个值的相互关系来更加准确、稳定地计算出地质体的信息.该方法可以计算复杂地质体的中心位置，且避免了以往线性方程反演方法需对结果进行筛选的复杂过程，对于存在剩磁的磁异常则采用解析信号的深度学习方法来进行位置反演.理论模型试验证明利用梯度比值的深度学习方法可以准确获得地质体的深度，且通过对比更多点的深度学习计算结果发现，采用多个不同比例极值点可以减弱噪声带来的干扰，从而得到更加准确的位置.最后将该方法应用于实测磁异常的反演工作，获得了地下磁性物体的中心位置，且计算结果与欧拉反褶积法相接近，因此该方法具有良好的实用性.
|
| 关 键 词： | 重磁场源 中心位置 比值 深度学习 地球物理 |
| 收稿时间： | 2020-08-13 |
Ratio Method for Calculating the Source Location of Gravity and Magnetic Anomalies Based on Deep Learning |
| |
| Ma Guoqing,Wu Qi,Xiong Shengqing,Li Lili.Ratio Method for Calculating the Source Location of Gravity and Magnetic Anomalies Based on Deep Learning[J].Earth Science-Journal of China University of Geosciences,2021,46(9):3365-3375. |
| |
| Authors: | Ma Guoqing Wu Qi Xiong Shengqing Li Lili |
| |
| Abstract: | The location of field source's center is one important purpose in the inversion of gravity and magnetic data, and the location of geological body is estimated mainly through the linear equation between the anomaly and the location of the field source. In order to obtain the location accurately and quickly, a deep learning technique based on the ratio of gravity and magnetic gradient is proposed to achieve the acquisition of the field source location in this paper, which can calculate the field source location quickly by using the deep learning technique to learn the relationship between the horizontal distribution of the gravity and magnetic gradient ratio, the buried depth, and the index. It is also proposed to use the mutual relationship of multiple values to calculate the information of the geological body more accurately and stably. This method can calculate the center location of complex geological bodies, and avoid the complicated process of screening the results of the previous methods, using the deep learning method of analytic signal for magnetic anomaly with remanent magnetism to achieve the location inversion. The application effect of the method is tested by theoretical models, which shows that the proposed method can obtain the depth information of the geological body accurately. Comparing the calculation results of the deep learning of more points, it is found that the use of multiple extreme points of different proportions can reduce the interference by noise and can get a location more accurately, so the method has good practicability. |
| |
| Keywords: | |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
| 点击此处可从《地球科学》浏览原始摘要信息 |
|
点击此处可从《地球科学》下载全文 |
|
