Cross-View Image Geo-localization Based on Attention Weight Masks

Cross-view Image Geo-localization is the process of determining the geographic location of a ground-view query image by matching it with geotagged satellite or unmanned aerial vehicle (UAV) captured images. In the context of ground images characterized by a constrained field of view (FoV), the query image exhibits a reduced coverage area, limited scene content, and an unknown imaging direction. Furthermore, reference satellite images from the same location may contain significant feature redundancy. The existing methods are not able to achieve a high degree of accuracy in the positioning of cross-sight terrain images, which is a problem given the importance of this task. The above issues result in the existing methods providing low accuracy for cross-view localization of ground images with a limited FoV. We propose a cross-view image geo-localization method based on attention weight mask alignment. The Coordinate Attention (CA) mechanism, when integrated into a lightweight ResNet18 network to generate weight masks, enables precise alignment between limited FoV ground images and satellite image feature maps. This process aims to remove redundant areas in satellite images and improve localizing accuracy. Given that feature maps at different levels describe images at different granularities, this work also introduces a multi-scale feature fusion strategy. It generates more representative image descriptors by combining features from different convolutional layers. Experimental results on the CVUSA and CVACT_val benchmark datasets demonstrate that when the FoV of ground images to be located is 70° and 90° with a random imaging direction, the proposed method significantly improves location accuracy.