A Fine-Grained Authentication Model Based on Perceptual Hashing and Grid Descriptor for Remote Sensing Image

Abstract

In this paper, a fine-grained authentication model for remote sensing image based on perceptual hashing and grid descriptor is proposed. Most perceptual hashing algorithms generate the hash value from an image’s global features, while remote sensing images are generally of huge amount and large size, so they are not suitable for remote sensing images authentication applications with high security demand. In this work, we apply grid descriptor to divide a remote sensing image, then generate the perceptual hash value of each region, and organize these hash values by embedding them into the corresponding region with watermarking technique. The grid descriptor is applied to detect and represent the tamper of the image. Compared with other authentication algorithms, the model can authentically remote sense image with different granularity.