Homography estimation is a common image alignment method. Existing methods of learning focused principal plane masks through deep neural networks lack explicit guidance. In this paper, we propose a novel unsupervised method to explicitly model anomaly descriptor removal and mask generation. Specifically, we select reliable feature descriptors from a novel perspective, and regard the features that are not responsible for alignment as outlier detection. The pixel-level support vector data description (PL-SVDD) module is designed. This module not only learns the feature representation of image pixels, but also fits a hypersphere to exclude the feature redundancy information that is not responsible for alignment from the hypersphere, thereby optimizing the feature descriptor. Based on the optimized image features, a correlation learning (CL) module is designed, which displays a generated mask through mathematical modeling to select reliable areas for homography estimation. Specifically, the feature descriptor of one of the unaligned images is modeled as a multivariate Gaussian distribution by Gaussian Density Estimation (GDE). Then, combined with the multivariate Gaussian distribution of the model and the feature descriptor of another image, the Mahalanobis distance is used to generate the mask. Experiments show that our method achieves good performance compared with the previous methods.