Gaoshuang Huang

Narrow field-of-view (NFoV) image cross-view localization is limited by insufficient information in the ground scene itself and cross-view domain differences, resulting in even more limited information directly corresponding to satellite imagery of the same area. This leads to low accuracy in existing methods and relatively few related studies. To address this issue, this paper innovatively employs cube projection technology to construct an NFoV ground dataset. Compared to traditional panoramic cropping methods, this method converts cylindrical projection images into central projection images, accurately simulating camera imaging geometry and providing high-fidelity data support. Secondly, this paper introduces monocular depth estimation to reduce interference information from ground images during satellite image retrieval, thereby highlighting core localization features. Therefore, a two-branch ground feature extraction algorithm (processing the original image and enhancing the image after filtering out invalid information) is designed to mine rich ground features, supplemented by an independent satellite feature extraction branch, forming a three-branch network to improve the accuracy of cross-view localization in narrow field-of-view images. Experimental results show that on the CVUSA dataset, the TOP-1 localization accuracy with a 90-degree field of view is improved from 55.9% to 89.14%; and on the CVACT dataset, the TOP-1 localization accuracy with a narrow field of view is improved from 40.6% to 84.6%.

Cross-view image geo-localization refers to determining the geographic location of a given query image using an image database with known location information. Existing methods mainly focus on learning discriminative image representations to optimize the distance of image feature representations in feature space without fully considering the positional relation information of the features and the information redundancy in the features themselves. Therefore, we proposed a cross-view image localization method that combines global spatial relation attention with feature aggregation. First, we utilize the lightweight global spatial relation attention to learn the spatial location structure information of features, which fully enhances the perceptual and discriminative capabilities of the model. The proposed attention has little effect on the complexity and memory occupancy of the model and can be generalized to other image-processing tasks. In addition, we introduce the sinkhorn algorithm for locally aggregated descriptors, which represents the aggregation of local features as an optimal transport problem and selectively discards useless information during the clustering and assignment of features, thus enhancing the generalization and robustness of the descriptors. Experimental results on the public University-1652, CVACT, and CVUSA datasets validate the effectiveness and superiority of the proposed method. Our code is available at https://github.com/WenjianGan/LRFR.

Cross-view geo-localization (CVGL), which involves matching and retrieving satellite images to determine the geographic location of a ground image, is crucial in GNSS-constrained scenarios. However, this task faces significant challenges due to substantial viewpoint discrepancies, the complexity of localization scenarios, and the need for global localization. To address these issues, we propose a novel CVGL framework that integrates the vision foundational model DINOv2 with an advanced feature mixer. Our framework introduces the symmetric InfoNCE loss and incorporates near-neighbor sampling and dynamic similarity sampling strategies, significantly enhancing localization accuracy. Experimental results show that our framework surpasses existing methods across multiple public and self-built datasets. To further improve global-scale performance, we have developed CV-Cities, a novel dataset for global CVGL. CV-Cities includes 223 736 ground-satellite image pairs with geolocation data, spanning sixteen cities across six continents and covering a wide range of complex scenarios, providing a challenging benchmark for CVGL. The framework trained with CV-Cities demonstrates high localization accuracy in various test cities, highlighting its strong globalization and generalization capabilities.

Cross-view geo-localization (CVGL), which involves matching and retrieving satellite images to determine the geographic location of a ground image, is crucial in GNSS-constrained scenarios. However, this task faces significant challenges due to substantial viewpoint discrepancies, the complexity of localization scenarios, and the need for global localization. To address these issues, we propose a novel CVGL framework that integrates the vision foundational model DINOv2 with an advanced feature mixer. Our framework introduces the symmetric InfoNCE loss and incorporates near-neighbor sampling and dynamic similarity sampling strategies, significantly enhancing localization accuracy. Experimental results show that our framework surpasses existing methods across multiple public and self-built datasets. To further improve globalscale performance, we have developed CV-Cities, a novel dataset for global CVGL. CV-Cities includes 223,736 ground-satellite image pairs with geolocation data, spanning sixteen cities across six continents and covering a wide range of complex scenarios, providing a challenging benchmark for CVGL. The framework trained with CV-Cities demonstrates high localization accuracy in various test cities, highlighting its strong globalization and generalization capabilities. Our datasets and codes are available at https://github.com/GaoShuang98/CVCities.

Obtaining the geographical location of images through image geo-localization technology is a highly significant task. However, existing image geo-localization methods struggle with accuracy under difficult conditions such as viewpoint changes, illumination variations, seasonal changes, and occlusions. To address these challenges, we proposed an image geo-localization architecture based on a foundation vision model and feature mixing. The architecture involves truncating and fine-tuning the foundation vision model DINOv2 to extract robust image features, which are then aggregated using an MLP-Mixer-based mix module to obtain robust and generalized image global features. This architecture significantly improves the accuracy of image geo-localization under difficult conditions. Experimental results demonstrate that the proposed architecture outperforms state-of-the-art (SOTA) methods in image geo-localization accuracy. Compared to SOTA methods, our architecture achieves accuracy improvements of 6.35%, 4.06%, and 6.30% on the test sets Tokyo 24/7, Nordland, and SF-XL-testv1, respectively, with viewpoint changes, illumination changes, season changes, and occlusions.

Visual place recognition (VPR) involves obtaining robust image descriptors to cope with differences in camera viewpoints and drastic external environment changes. Utilizing multiscale features improves the robustness of image descriptors; however, existing methods neither exploit the multiscale features generated during feature extraction nor consider the feature redundancy problem when fusing multiscale information when image descriptors are enhanced. We propose a novel encoding strategy—convolutional multilayer perceptron orthogonal fusion of multiscale features (ConvMLP-OFMS)—for VPR. A ConvMLP is used to obtain robust and generalized global image descriptors and the multiscale features generated during feature extraction are used to enhance the global descriptors to cope with changes in the environment and viewpoints. Additionally, an attention mechanism is used to eliminate noise and redundant information. Compared to traditional methods that use tensor splicing for feature fusion, we introduced matrix orthogonal decomposition to eliminate redundant information. Experiments demonstrated that the proposed architecture outperformed NetVLAD, CosPlace, ConvAP, and other methods. On the Pittsburgh and MSLS datasets, which contained significant viewpoint and illumination variations, our method achieved 92.5% and 86.5% Recall@1, respectively. We also achieved good performances—80.6% and 43.2%—on the SPED and NordLand datasets, respectively, which have more extreme illumination and appearance variations.

Geolocating a street-view image by matching it with geotagged satellite images is crucial for location assessment. However, the perspective disparity between satellite and street-view images presents significant challenges. To address the issue, the mainstream approach is to convert satellite images to ground-level perspective. So, the reference satellite images not only required the center but also required the coverage to be consistent with the street-view images and sometimes even required consistency in the north direction, which is difficult to achieve in the practical applications. This letter introduces a ground-breaking method for converting ground-level images to satellite images. We effectively couple ground images and satellite images by establishing a hemispheric projection relationship to achieve conversion from ground images to satellite perspectives, thus solving the problem of huge perspective differences in cross-view geolocation (CVG). In addition, we propose the multiscale fusion of a spatial features mechanism to enhance deep feature representations and improve recall and geolocation accuracy. Our proposed methodology markedly enhances its practical utility and performance, attaining remarkable Top-1 accuracy rates of 75.08% on the CVACT_val dataset and 39.92% on the CVACT_test dataset, respectively, following a stringent quantitative appraisal. This advancement constitutes a substantial contribution to the progression of research in the field of CVG.

Utilizing visual place recognition (VPR) technology to ascertain the geographical location of publicly available images is a pressing issue for real-world VPR applications. Although most current VPR methods achieve favorable results under ideal conditions, their performance in complex environments, characterized by lighting variations, seasonal changes, and occlusions caused by moving objects, is generally unsatisfactory. In this study, we utilize the DINOv2 model as the backbone network for trimming and fine-tuning to extract robust image features. We propose a novel VPR architecture called DINO-Mix, which combines a foundational vision model with feature aggregation. This architecture relies on the powerful image feature extraction capabilities of foundational vision models. We employ an MLP-Mixer-based mix module to aggregate image features, resulting in globally robust and generalizable descriptors that enable high-precision VPR. We experimentally demonstrate that the proposed DINO-Mix architecture significantly outperforms current state-of-the-art (SOTA) methods. In test sets having lighting variations, seasonal changes, and occlusions (Tokyo24/7, Nordland, SF-XL-Testv1), our proposed DINO-Mix architecture achieved Top-1 accuracy rates of 91.75%, 80.18%, and 82%, respectively. Compared with SOTA methods, our architecture exhibited an average accuracy improvement of 5.14%.

Retro-reflective targets (RRTs) are important for obtaining the coordinates of the surface points of objects measured using industrial photogrammetry systems. Targeting the low measurement accuracy caused by the poor roundness of the conventional RRT, a novel RRT was designed in this study by covering a die-cut slice with round holes and high roundness on a glass bead reflective film. The roundness of this novel RRT was on average 0.03 mm higher than that of the conventional RRT. The accuracy of the novel RRT was evaluated in terms of both internal and external coincidence accuracies. The measurement repeatability accuracy of the novel RRT was improved by 25.0% compared with that of the conventional RRT. The thinner the die-cut slice thickness, the higher the repeatability in a certain range. Similarly, the plane fitting, flatness measurement, and standard length measurement accuracies were improved by 30.7%, 34.1%, and 28.1%, respectively.