Date of this Version
Written for presentation at the 2004 ASAE/CSAE Annual International Meeting Sponsored by ASAE Fairmont Chateau Laurier, The Westin, Government Centre Ottawa, Ontario, Canada 1 - 4 August 2004
Soils can be non-intrusively mapped by observing similar patterns within ground-penetrating radar (GPR) profiles. We observed that the intricate and often indiscernible textural variability found within a complex GPR image possesses important parameters that help delineate regions of similar soil characteristics. Therefore, in this study, we examined the feasibility of using textural features extracted from GPR data to automate soil characterizations. The textural features were matched to a "fingerprint" database of previous soil classifications of GPR textural features and the corresponding ground truths of soil conditions. Four textural features (energy, contrast, entropy, and homogeneity) were selected for inputs into a neural-network classifier. This classifier was tested and verified using GPR data obtained from two distinctly different field sites. The first data set contained features that indicate the presence or lack of sandstone bedrock in the upper 2 m of a shallow soil profile of fine sandy loan and loam. The second data set contained columnar patterns that correspond to the presence or the lack of vertical preferential-flow paths within a deep loess soil. The classifier automatically grouped each of these data sets into one of the two categories. Comparing the results of classification using extracted textural features to the results obtained by visual interpretation found 93.6% of the sections that lack sandstone bedrock correctly classified in the first set of data, and 90% of the sections that contain pronounced columnar patterns correctly classified in the second set of data. The classified profile sections were mapped using integrated GPR and GPS data to show surface boundaries of different soil categories. These results indicate that extracted textural features can be utilized for automatic characterization of soils using GPR data.