<p>In this paper, we propose a copyright protection method for panoramic works based on feature recognition. The method can be used to compare original and plagiarism suspicious panoramic images to identify and authenticate whether there is any infringement. If feature extraction procedure is directly applied on the sphere will cause a large amount of feature information loss. We use equirectangular projection to project the panoramic image of sphere to a 2D plane, which is convenient for feature extraction and reduces the loss of feature information. A scheme of panorama feature extraction is explored to achieve copyright protection. Swin-transformer is used for feature extraction, which can split the image into windows and layers, exchanging information between different windows to improve the integrity of feature information. Feature fusion and feature enhancement are performed on three feature layers through the feature pyramid network to output the final feature information. The proposed method can recognize multi-angle feature information by means of splitting windows and layers, which has good practicality. The experimental results show that the proposed method has good effect on the infringement determination of panoramic images.</p>