Generalized Category Discovery for Remote Sensing Image Scene Classification

Deep neural networks have achieved promising progress in remote sensing (RS) image classification. However, the training process requires abundant samples for each class, and it is unrealistic to annotate labels for each RS category, especially considering that the RS target database is increasing dynamically. Therefore, we introduce an innovative prototype network tailored for Generalized Category Discovery (GCD) in remote sensing scene classification. This network consists of two essential modules: one dedicated to representation learning and the other to prototype learning. Through extensive experiments conducted on three benchmark datasets, i.e., RSS-DIVCS, NWPU-RESISC45, and AID, we demonstrate that the proposed model achieves remarkable performance gain up to 20%, effectively addressing the challenges inherent in classifying dynamically varying remote sensing images.

1. INTRODUCTION

Remote sensing (RS) image scene classification has captured growing interest owing to its invaluable data support across various domains, encompassing environmental monitoring, resource management, disaster response, etc [1], [2]. At the same time, remarkable advancements have been made in the application of deep learning algorithms for RS image scene classification, progressing from initial usage of CNN [3], vision transformer [4], to vision-language Models [1] and multi-modal learning [5]. These algorithms continuously enhance the accuracy and efficiency of classifying RS images and can even surpass human capabilities. Nonetheless, their high accuracy predominantly pertains to familiar scene categories. Additionally, the application of deep learning in the field of remote sensing [6], [7], [8] is widespread, reflecting its significant impact and utility in various aspects of this area.