Research On Method Of Magnetic Anomaly Detection Based On Gradient Tensor

The magnetic anomaly detection technology is becoming a hotspot in the researchof magnetic. Along with the improvements of the magnetic sensor technology andmagnetic measurements, this technology has made great progress. The theory ofmagnetic anomaly detection is based on the fact that a ferromagnetic material couldgenerate magnetic anomaly by the effect of geomagnetic field. The basic idea of themagnetic anomaly detection is as the following steps: First, measure the magnetic fieldsto obtain the information of magnetic anomaly. Then, process the data by using certainalgorithm. The detail of magnetic anomaly target, such as the location, shape and stateof motion, could be calculated. As a mathematical tool, tensor reflects a uniqueadvantage in the analysis of the physical field. Compared with the total amount ofmagnetic field, the magnetic field component and a magnetic field gradient, magneticgradient tensor contains more information of magnetic field, so it is suitable for theapplications in magnetic anomaly detection. The method of magnetic anomaly detectionbased on the gradient tensor is studied in this subject, its main content and innovationsare as following:Based on tensor theory, the properties of the magnetic gradient tensor are discusseddetailedly, especially about its advantage in measuring the magnetic anomaly that it isable to overcome the infection of geomagnetic field and obtain the information ofmagnetic anomaly effectively. Combined with the characteristics of the magneticgradient tensor, three different types of magnetic gradient tensor measurement array,which is constituted by different sensors and different structure, are introduced. Aplanar magnetic gradient tensor measurement array is designed by using four triaxialfluxgate sensors.In order to eliminate the measurement error and improve the measurementaccuracy of the measurement array, a suitable method is proposed for calibration of themagnetic gradient tensor measurement array. The error of a single triaxial fluxgatesensor is mainly caused by triaxial non-orthogonal error, zero error bias and scale factorerror. After discussing the calibration model for the single triaxial fluxgate sensor, amethod based on linear parameter is given. The results of experiment show that, thecalibration method is able to reduce standard deviation of the error to only about3nT,the measurement of a single three-axis fluxgate sensor has been improved significantly.The error of the measurement array is mainly caused by the non-alignmentbetween coordinates of different sensors. The calibration model is based on the Eulerrotation angles and the calibration method is to estimate the parameters by solving thenonlinear equations. The results of experiment show that, the calibration method is ableto reduce the error of components to only about15nT, the accuracy of the measurement array has been improved significantly.On the condition of magnetic dipole model, two different magnetic anomalylocalization algorithms are discussed, especially, the idea of data process is shown andthe localization formula is deduced. The characteristics of the two methods arecompared, and one of them is chosen as the localization algorithm for the research ofthis subject.Some exploratory experiments are carried out. Firstly, the software simulation iscarried out in order to verify the effectiveness of the localization algorithm. The resultindicates that the localization algorithm has very high localization accuracy andexcellent real-time performance. In the real geomagnetic field environment, magneticanomaly detection experiments are carried out by using the magnetic gradient tensormeasurement array. The results show that the localization algorithm discussed in thissubject is able to realize the magnetic anomaly detection. As for the magnetic gradienttensor measurement array designed in this subject, the range of effective localization is2m, the error is less than15%; the range of maximal detection is3m, the error is lessthan50%.Finally, a comprehensive summary of the subject is given, the prospect of furtherresearch is described.