INVESTIGATION OF A FUZZY-NEURAL NETWORK APPLICATION IN CLASSIFICATION OF SOILS USING GROUND-PENETRATING RADAR IMAGERY

Errors associated with visual inspection and interpretations of radargrams often inhibit the intensive surveying of widespread areas using ground-penetrating radar (GPR). To automate the interpretive process, this article presents an application of a fuzzy-neural network (F-NN) classifier for unsupervised clustering and classification of soil profiles using GPR imagery. The classifier clusters and classifies soil profile strips along a traverse based on common pattern similarities that can relate to physical features of the soil (e.g., number of horizons; depth, texture, and structure of the horizons; and relative arrangement of the horizons, etc.). This article illustrates this classification procedure by its application on GPR data, both simulated and actual. Results show that the procedure is able to classify the profile into zones that corresponded with the classifications obtained by visual inspection and interpretation of radar grams. Application of F-NN to a study site in southwest Tennessee gave soil groupings that are in close correspondence with the groupings obtained in a previous study, which used the traditional methods of complete soil morphological, chemical, and physical characterization. At a crossover value of 3.0, the F-NN soil grouping boundary locations fall within a range of +/-2.7 m from the soil groupings determined by the traditional methods. These results indicate that F-NN can supply accurate real-time soil profile clustering and classification during field surveys.