Research And Improvement Of Small Subdomain Filtering Of Gravity And Magnetic Anomaly

Gravity and magnetic survey is traditional geophysical exploration method, it hasunique advantage in the research of the geological target body edge position-such asdetermination of the fracture location, rock boundaries, etc. variation ratio of gravityand magnetic anomaly is larger near the edge of the geological body. Subdomaindivision methods use the principle that anomaly changes fast near the edge to detectedge. That’s why we can recognize and enhance based on it. However, undergroundsource is often complicated and the actual measurement data is always with kinds ofinterference. Therefore, people will filter out interference, and try to enhance the targetinformation characteristics such as gravity anomaly zone. In numerous methods, subdomain decomposition method (called small subdomain filtering in some documents)identify purpose by using the edge features. Due to its principle is simple and thecalculation is fast, and effective, the method becomes a research hotspot in recent years.This work is mainly trying to deepen the small subdomain filtering and try to solve itsdefect to make it better applied to the need of gravity and magnetic exploration.Subdomain division can be low pass filter as well as enhance abnormal resolution.Its effect is obvious in recognizing and enhancing gravity anomaly zone, but there arealso some defects such as polyline aliasing in contour map and this is our research focalpoint. In order to improve polyline aliasing, we have done a large numbers ofexperiment. The reasons of polyline are as follows, the first is data format, what thealgorithm process is discrete grid data. For the discrete data, when the grid spacing istoo large, often embodied in the form of polyline on the grid. The second is thealgorithm itself, the algorithm makes data in flat area compress to gravity anomaly zoneso that it can be enhanced. Finally, data after encryption concentrates between twoadjacent data points and the width of gravity anomaly zone filtered is one grid distance.The general trend is consistent with the original trend but will be fixed in the grid. For the two reasons, we propose several kinds of methods to improve the defects. Thefirst one is interpolating the original data to reduce the width of the grid. The second oneis post-processing of the data after filtering, here we use the gradient retention algorithm.These two methods together can achieve better results.The third method is improving thealgorithm, the idea is to locate the polyline data and reducing the changing degree of data.The forth method changes the data points’ locations to change contours’ density. All thefour methods can solve the anomaly problem well. What’s more, there are other methodswhose effect is not very ideal, they are also presented briefly in the paper just for reference.