Novel Method for Correcting Land-Cover Misclassification in Mountainous Paddy Fields: a Case Study of the Honghe Hani Rice Terraces in Yunnan, China

In mountainous areas, radiance decreases locally owing to the shade created by the topographic slope. This reduction causes paddy fields to be misclassified using satellite images. The characteristics of the radiance of paddy fields are close to those of shaded forest and grassland because paddy fields are covered with paddy and water that absorb radiance. In this study, we propose a novel method to correct the classification results using the terrain characteristics of terraced paddy fields in mountainous areas. We then applied this method to the Honghe Hani Rice Terraces in Yunnan, China. The results show that this correction method significantly improves the accuracy of paddy field classification. The kappa coefficient of the proposed method is equal to or greater than that of existing model.