Crop classification with multi-temporal and dual-polarimetric TerraSAR-X data

Abstract

Crop maps are useful for agricultural field management and synthetic aperture radar data are attractive for generating crop classification because of their all-weather, all-day imaging capability. Additionally, classification algorithms are essential for generating accurate and multi-Grained Cascade Forest, which is also called ‘deep forest', was developed and its high performances have been shown for pattern recognition, voice recognition and so on. In this study, the capability of TerraSAR-X (including TanDEM-X) dual-polarimetric data for crop classification in the Tokachi Plain, Japan was investigated and the comparison of three different classification algorithms including classification and regression tree, random forests and deep forest was conducted.