Land cover classification based on SAR and optical images using ensemble machine learning

Abstract

Land cover classification is an important remote sensing application. Satellite optical and radar images are two of the most widely used remote sensing images with respective advantages and disadvantages. Using both radar and optical images and machine learning, this project is to explore automatic land cover classification traditionally done manually or semi-manually.

The objective is to study, develop and compare different ensemble/non-ensemble machine learning algorithms for the PolSAR-based crop identification task, and learn the pre-processing procedure and general flow of any land-cover classification based on the remote sensing SAR image data.

By analyzing different classification results with various machine learning algorithms, we can get a deep understanding of discriminative machine learning problems.