Buried Objects Detection and Segmentation Using a Semi-Automatic Technique Called Selected Peaks of Summing up Traces (SPST)

Abstract

Ground Penetrating Radar (GPR) is a nondestructive geophysical technique that uses electromagnetic waves to evaluate subsurface information. This research presents package software for GPR data analysis and image processing with an algorithm for semi-automatic object detection, displaying, filtering, edges detection, histogram equalization, segmentation such as quantization and Fuzzy C-means clustering (FCM). After applying preprocessing on the image, a new technique depends on taking the Selected Peaks of Summing up Traces (SPST) have been adopted to give the very good results of buried objects detection compares with other segmentation methods. MSE and RMSE criteria were used to perform the comparison between the results. The fundamentals of GPR and software are illustrated inside with data taken by (MALÅ GPR ProEx System) at the University of Baghdad campus -college of science.