Research On Magnetic Interference Compensation Technology For Aeromagnetic Measurement System Based On Rotor UAV

Aeromagnetic survey,as a fast and efficient method of aerial geophysical exploration,plays a significant role in structural and mineral resource exploration.Traditional aeromagnetic measurement systems are mostly composed of fixed wings,helicopters,and other manned aerial vehicles.With the development of drone technology,its application in fields such as geological and mineral exploration is increasingly widespread due to its high flexibility,low flight altitude,and low cost.Rotor unmanned aerial vehicles(UAVs)have become an ideal platform in the field of aeromagnetic detection systems due to their excellent ability to fly anywhere and perform vertical takeoff and landing.However,the rotor unmanned aerial vehicles aeromagnetic system is also subject to the same magnetic interference as the manned aircraft aeromagnetic measurement system.Magnetic interference will affect the data quality of aeromagnetic measurement.The high-frequency magnetic interference sources within the system can be eliminated through filtering or other specific nonfiltering methods,but low-frequency maneuvering magnetic interference needs to be processed and eliminated through aeromagnetic interference compensation technology.Therefore,studying an efficient magnetic interference compensation algorithm for rotor unmanned aerial vehicle aeromagnetic measurement systems is of great significance for improving the quality of aeromagnetic measurement data.This article conducts research on two representative aeromagnetic measurement systems,single rotor unmanned aerial vehicles and multi rotor unmanned aerial vehicles.By analyzing the sources and properties of magnetic interference in the aeromagnetic system,a magnetic interference compensation model is established for the main low-frequency magnetic interference in the system,and an aeromagnetic compensation model and evaluation system based on Tolles Lawson linear regression are introduced.Due to the differences in flight control algorithms and control methods between single and multi rotor unmanned aerial vehicles,targeted research on magnetic interference compensation methods is of great significance for improving compensation accuracy.This article focuses on the research of magnetic interference compensation for rotor unmanned aerial vehicle aeromagnetic measurement systems,and proposes two methods of aeromagnetic compensation to obtain high-quality unmanned aerial vehicles aeromagnetic measurement data.This article decouples the T-L compensation model of the rotor unmanned aerial vehicle’s magnetic measurement system,and obtains that multicollinearity variables have differences in different heading directions.Based on this characteristic,an improved adaptive modeling method for aeromagnetic compensation is proposed.By adaptively allocating effective variables with high significance and low multicollinearity,a compensation model for compensating flight is constructed based on heading.Compared with non-adaptive compensation models,the ill posed problem of multi heading models is further refined to improve compensation accuracy.This article focuses on the gradient vanishing problem in neural network aeromagnetic compensation models.From the perspective of improving the neural network aeromagnetic compensation model,a high-performance residual neural network aeromagnetic compensation model is proposed by combining residual connected and fully connected neural networks.Due to its identity mapping structure,residual connections can achieve cross layer propagation of gradients,effectively solving the problem of gradient vanishing in the backpropagation process of fully connected neural networks,and improving the performance of neural network aeromagnetic compensation models.This article validates the two methods on single and multi rotor unmanned aerial vehicle detection platforms.Compensation flight experiments show that the proposed method has higher compensation accuracy compared to traditional methods,which verifies the effectiveness of the model.The application on survey line data shows that the compensated magnetic anomaly range is clearer and smoother,which can provide better data basis for subsequent inversion interpretation work.On the basis of various models and algorithms,this article has integrated and developed an aeromagnetic compensation software system for practical exploration and production of aeromagnetic data compensation processing.