The emergence of cloud storage model has brought new opportunities to the storage and management of massive ocean remote sensing images, and more and more users choose to transfer their image data into the cloud, which brings challenges for image data security and usability due to the openness of cloud. We propose an image authentication scheme to protect the ocean remote sensing image confidentiality in cloud environment, which combines the Hash function with the (k, n) threshold secret sharing method to detect the changes of the image data in sensitive regions and verify the consistency between encrypted and recovered images. The sensitive-region image cannot be restored when more than n-k out of n sub-secrets get lost or tampered. To avoid this, we divide the sensitive region image into blocks, and implement secret sharing for each sub block, thus guaranteeing the lossless recovery of partial images and enhancing the availability of image data. Analysis on comparison experiments shows that the algorithm can effectively prevent fraud in the process of secret image recovery, and obtain higher cloud storage security for remote sensing images than traditional methods.