Multisource Data Reconstruction-Based Deep Unsupervised Hashing for Unisource Remote Sensing Image Retrieval

Unsupervised hashing for remote sensing (RS) image retrieval first extracts image features and then uses these features to construct supervised information (e.g., pseudolabels) to train hashing networks. Existing methods usually regard RS images as natural images to extract unisource features. However, these features only contain partial information about ground objects and cannot produce reliable pseudolabels. In addition, existing methods only generate a pseudo-single-label to annotate each RS image, which cannot accurately represent multiple scenes in an RS image. To address these drawbacks, this article proposes a new Multisource data reconstruction-based deep unsupervised Hashing method, called MrHash, which explores the characteristics of RS images to construct reliable pseudolabels. In particular, we first use geographic coordinates to obtain different satellite images and develop a novel autoencoder network to extract multisource features from these images. Then, pseudo-multilabels are designed to deal with the coexistence of multiple scenes in a single image. These labels are generated by a custom probability function with extracted multisource features. Finally, we propose a novel multisemantic hash loss by using the Kullback-Leibler (KL) divergence to preserve the semantic similarity of these pseudo-multilabels in Hamming space. Our newly developed MrHash only uses multisource images to construct supervised information, and hash code generation still relies on a unisource input image. Experiments on benchmark datasets clearly show the superiority of the proposed method over state-of-the-art baselines. We have added detailed descriptions about our source code. Please check them by accessing https://github.com/sunyuxi/MrHash.