Tractor wheel tracks detection, delineation, and characterization in very high spatial resolution SAR imagery
Résumé
Compacted tractor wheel tracks have been recognized as key factors controlling runoff and erosion processes in agricultural landscapes. In this context, the availability of an automatic tool for wheel tracks detection and characterization would be very useful. In this letter, an original algorithm based on Radon Transform is proposed for automatic wheel tracks detection on Synthetic aperture radar images with a spatial resolution of 1 m. Compared to on-screen measurements, wheel tracks orientations and widths were accurately estimated for the images acquired with shallow incidence angles whereas poorer results were observed for sharp incidence angles.