The problem of using segmentation data for change detection in surveyed areas in Thailand

Abstract

The study of spatial changes is crucial in the field of remote sensing. To determine which areas undergo changes each year, we use change detection techniques to identify them changes over different periods. However, a limitation we faced with the data from Thailand was the lack of labeled data for change detection. To address this, we utilized segmentation data provided by experts who created new labels for change detection. Despite this effort, several issues arose during testing. This report details those challenges and the outcomes of our study