In this paper, we introduce G2-Mapping, a novel method to comprehensively support online monocular, RGB-D, and LiDAR-Inertial-Visual systems, employing 3D gaussian points as scene representation. There are several issues when applying 3d gaussian splatting (3DGS) techniques to simultaneous localization and mapping (SLAM) 1) for monocular, the lack of depth information makes scene initialization difficult and large baseline positioning challenging; 2) differentiable rendering with respect to depth and pose has not been implemented in 3DGS, making it difficult to directly apply to the SLAM system; 3) strategy for updating the scene with incoming online frames is not present, which may lead to memory overflow. In order to overcome problems mentioned above, we formulate a mathematical derivation and propose a differentiable rendering approach that leverages both depth and color to optimize the scene and pose. We introduce a simplified odometry that provides a metric depth estimation for monocular and enhance the low-overlap scene availability. A scale consistency and uncertainty weighted optimization is further proposed to eliminates the impact of inaccurate depth prediction. Our proposed scene updating strategy effectively prevents rapid memory growth. Tracking and mapping are performed alternatively to achieve precise localization and synchronous high-fidelity map reconstruction. Extensive experiments demonstrate that our G2-Mapping surpasses feature-based SLAM in localization precision and exceeds state-of-the-art neural SLAM methods in the fidelity of view synthesis. Note to Practitioners—This paper is dedicated to tackling the efficiency challenges in multi-source SLAM and map reconstruction, aiming to generate pose and high-fidelity maps synchronously. We introduce G2-Mapping, a novel framework that leverages the power of 3D Gaussian points for scene representation, offering a universal solution for monocular, RGB-D, and LiDAR-Inertial-Visual systems. By developing a comprehensive differentiable renderer and presenting a strategy for dynamic scene updating, G2-Mapping significantly advances the state-of-the-art in localization precision and view synthesis fidelity. Although the approach is highly promising, it currently relies on the accuracy of depth prediction networks and requires further optimization for handling sparse LiDAR data. Future research will focus on enhancing these aspects, aiming to seamlessly integrate G2-Mapping into practical applications within robotics, autonomous vehicles, and augmented reality, where robust and efficient SLAM solutions are paramount.

Traditional change detectors are ill-equipped for incremental learning (IL). Existing IL methods address the problem of catastrophic forgetting by artificially adding categories and utilizing old labels for learning supervision. Current strategies for change detection (CD) are inadequate as they fail to address a crucial aspect of the task: the constant label space throughout each training step, causing label conflicts between background-class pixels (representing unchanged regions) and changed pixels, which can lead to knowledge confusion. In this work, we revisit classical IL methods and propose an effective framework that explicitly addresses this conflict. Furthermore, we design a hierarchical distillation to ensure adequate retention of the learned features. The proposed architecture and distillation method balance acquiring new knowledge and preserving old knowledge effectively. To address the absence of datasets for IL in CD, we design a multidomain CD dataset that encompasses three distinct environments. Our proposed method demonstrates a significant improvement in the performance of IL, as measured by $\Delta _{\text {IoU}}$ and $\Delta _{\text {F1}}$ . Extensive experiments on this dataset show that our performance is promising compared to state-of-the-art CD methods and incremental methods.

Cross-view image-based geolocalization can provide accurate, drift-free navigation in the absence of external positioning signals, which is particularly crucial in domains such as UAV delivery logistics and disaster relief. While existing research predominantly focuses on matching ground panoramic images with known orientations, real-world scenarios often present images with unknown orientations. Additionally, acquiring panoramic images with a full $360^\circ$ field of view (FoV) is impractical, creating a disconnect between research and practical applications. In this study, we introduce an innovative cross-view geolocalization method that integrates sophisticated orientation estimation. Our goal is to improve geolocalization accuracy in limited FoV and arbitrary orientation scenarios. To achieve this, we design an efficient hard sample training strategy employing a dual-weighted soft-margin triplet loss function to extract discriminative features. Concurrently, we develop an orientation estimation module that leverages a convolution-based sliding window to assess the similarity with satellite-view embeddings and query embeddings. Thus, our method is named OriLoc. We validated the approach's unique outperformance on three challenging cross-view geolocalization datasets, including data from commercial, residential, urban landscapes, and suburban areas across two continents, aligning with the demands of real-world scenarios. Our results demonstrate that a straightforward hard sample mining strategy, when combined with an appropriate learning objective, significantly enhances geolocalization performance for limited FoV and orientation-free cross-view images, as shown in Figure 1. Furthermore, our orientation estimation module, which incorporates classical convolutions, exhibits remarkable accuracy in estimating orientations when integrated with attention embeddings prior to polar transformation. Code and trained models are publicly available on https://github.com/boni-hu/OriLoc.

Large-scale visual geolocation is a meaningful task that involves locating a query image by comparing it with images in a database and predicting the most similar image. However, the widely-used training framework based on contrastive learning cannot fully utilize all data and is difficult to adapt to larger scales. At the same time, the traditional Convolutional Neural Networks (CNNs) and Vector of Locally Aggregated Descriptors (VLAD) using aggregated features cannot fully reflect the relationship between the local features of the image. Therefore, a Graph Neural Network (GNN) is designed as the feature extraction network, and then a training framework based on image classification is constructed. Specifically, a data grouping strategy and special loss function are designed for better training results. After training, we adopt an image retrieval strategy based on kNN for position. In addition, considering that existing datasets cannot be adapted to our requirements, two datasets are constructed for experiments, that one contains large-scale satellite images and the other fuses satellite and Unmanned Aerial Vehicle (UAV) images. Results demonstrate that our method outperforms other common methods on both the two datasets. The results demonstrate the effectiveness of our approach for UAV visual geolocation and provide ideas for future research in this field.

Visual geolocalization is a cost-effective and scalable task that involves matching one or more query images, taken at some unknown location, to a set of geotagged reference images. Existing methods, devoted to semantic features representation, evolving towards robustness to a wide variety between query and reference, including illumination and viewpoint changes, as well as scale and seasonal variations. However, practical visual geolocalization approaches need to be robust in appearance changing and extreme viewpoint variation conditions, while providing accurate global location estimates. Therefore, inspired by curriculum design, human learn general knowledge first and then delve into professional expertise. We first recognize semantic scene and then measure geometric structure. Our approach, termed CurriculumLoc , involves a delicate design of multi-stage refinement pipeline and a novel keypoint detection and description with global semantic awareness and local geometric verification. We rerank candidates and solve a particular cross-domain perspective-n-point (PnP) problem based on these keypoints and corresponding descriptors, position refinement occurs incrementally. The extensive experimental results on our collected dataset, TerraTrack and a benchmark dataset, ALTO , demonstrate that our approach results in the aforementioned desirable characteristics of a practical visual geolocalization solution. Additionally, we achieve new high recall@1 scores of 62.6% and 94.5% on ALTO, with two different distances metrics, respectively. Dataset, code and trained models are publicly available on https://github.com/npupilab/CurriculumLoc.