Leveraging Permuted Image Restoration for Improved Interpretation of Remote Sensing Images
Awen Bai, Jie Chen, Wei Yang, Zhirong Men, Shengming Zhang, Hongcheng Zeng, Weichen Xu, Jian Cao · IEEE Transactions on Geoscience and Remote Sensing · 2024
In this study, we introduce a novel self-supervised learning adapter based on permutated image restoration (PIR) for effectively transferring pretrained weights from natural images to remote sensing object detection tasks. The adapter’s unique methodology encompasses a three-phase process: segmenting and permuting image blocks, estimating permutation matrices for sequence reconstruction, and applying specialized loss functions for accurate block positioning. The use of our approach results in the maintenance of fidelity in both absolute and relative block positions as demonstrated by the evaluation of block similarities. The empirical results indicate significant performance enhancements for diverse datasets spanning optical and SAR data types, including HRSC2016, SODA-A, and RSDD, while effectively avoiding overfitting.