AutoCrop: An Automatic Approach to Cropping Panoramic Images

During the process of image acquisition, factors such as camera motion and lighting changes often affect the usability and viewing ability of panoramic images, resulting in irregular boundaries in the panoramic images calculated by image stitching algorithms. In response to the above issues, in this paper, we propose an automatic cropping method for panoramic images, called AutoCrop which can crop irregular regions and make the edges of the panoramic images more regular, thereby enhancing the usability and viewing ability of the panoramic images. The core idea of the AutoCrop algorithm is as follows: firstly, calculate the binary image for the input panoramic image; Secondly, calculate the maximum inscribed rectangle contained in the calculated binary image; Then, calculate the region of interest of the panoramic image based on the maximum inscribed rectangle; Finally, build the regular panoramic image by the extracted image information within the region of interest of the input panoramic images. Based on this method, we developed an automatic panoramic image cropping software called AutoCrop, which can crop stitched panoramic images into regular panoramic images. The AutoCrop has been tested on several open-source datasets, and experimental results have all proven the effectiveness of this method.