The high-frequency content of high-resolution aeromagnetic data is of particular interest to geophysic ists to identify mineral deposits, shallow faults, and dikes. However, high-resolution aeromagnetic data is contaminated by cultural noise generated from aircraft and man-made features. The culture noise must be removed before starting the interpretation process. Manual techniques are more selective of the noise; however slower and more expensive because they require considerable hands-on interaction. The present study develops a novel method for detecting and removing the culture noise from aeromagnetic data based on an artificial neural network (ANN) in an automatic way and comparing the results with a conventional algorithm using the nonlinear filter. The proposed method is tested using a theoretical example that combine a magnetic anomaly due to a dyke with three sources of cultural noise, besides using a practical example to increase the number of training pattern. The network is trained based on the backpropagation training function, where the algorithm updates the weight and bias states as per the Levenberg–Marquardt optimization. The optimization is reached during the training and validation process after 3,000 iterations. The correlation coefficient (R) is utilized along with the mean squared error (MSE) as performance indices of the ANN. The ANN demonstrates the capability to detect the spiky data based on the optimal weights, thus allowing for removing and replacing them with clean data using the piecewise cubic Hermite interpolating polynomial (PCHIP) function. The practical utility of the two-method is discussed using high-resolution aeromagnetic data from the Tushka area located in the southwestern desert of Egypt. Comparing the denoising results using the two methods shows that the current approach is more effective in processing and more closely recovering the original magnetic data.

You do not have access to this content, please speak to your institutional administrator if you feel you should have access.