Norbert Haala

We present HI-SLAM2, a geometry-aware Gaussian SLAM system that achieves fast and accurate monocular scene reconstruction using only RGB input. Existing Neural SLAM or 3DGS-based SLAM methods often trade off between rendering quality and geometry accuracy, our research demonstrates that both can be achieved simultaneously with RGB input alone. The key idea of our approach is to enhance the ability for geometry estimation by combining easy-to-obtain monocular priors with learning-based dense SLAM, and then using 3D Gaussian splatting as our core map representation to efficiently model the scene. Upon loop closure, our method ensures on-the-fly global consistency through efficient pose graph bundle adjustment and instant map updates by explicitly deforming the 3D Gaussian units based on anchored keyframe updates. Furthermore, we introduce a grid-based scale alignment strategy to maintain improved scale consistency in prior depths for finer depth details. Through extensive experiments on Replica, ScanNet, Waymo Open, ETH3D SLAM and ScanNet++ datasets, we demonstrate significant improvements over existing Neural SLAM methods and even surpass RGB-D-based methods in both reconstruction and rendering quality.

The creation of highly accurate and collaborative mapping algorithms is crucial for the progress of SLAM technology, as it greatly improve the efficiency of building detailed maps. In the area of mapping based on single moving trajectory, DROID-SLAM (Differentiable Recurrent Optimization-Inspired Design) by (Teed and Deng, 2021) stands out as an innovative method based on deep learning, providing a visual-only solution that works with various types of camera, such as monocular, stereo, and RGB-D. Its ability to create maps with excellent accuracy makes it superior to well-known methods like ORB-SLAM3 by (Campos et al., 2021). Despite its impressive individual mapping performance, DROID-SLAM does not account for scenarios involving multisession data or the collaborative map creation by multiple agents. To address this problem, we propose two collaborative map construction algorithms built upon DROID-SLAM. Compared to prior methods that compute explicit relative transformations for loop closures, our algorithm leverages the power of deep learning-based bundle adjustment, using dense per-pixel correspondence, to merge into a globally consistent state. These algorithms have been thoroughly tested with stereo and RGB-D models. we validated the effectiveness of our proposed algorithms on both public and self-collected datasets, showing higher accuracy than prior methods. By leveraging the strengths of DROID-SLAM while addressing its limitations with our novel algorithms, we extend the application scenarios of this method and provide a new way of thinking about collaborative mapping.

While originally developed for novel view synthesis, Neural Radiance Fields (NeRFs) have recently emerged as an alternative to multi-view stereo (MVS). Triggered by a manifold of research activities, promising results have been gained especially for texture-less, transparent, and reflecting surfaces, while such scenarios remain challenging for traditional MVS-based approaches. However, most of these investigations focus on close-range scenarios, with studies for airborne scenarios still missing. For this task, NeRFs face potential difficulties at areas of low image redundancy and weak data evidence, as often found in street canyons, facades or building shadows. Furthermore, training such networks is computationally expensive. Thus, the aim of our work is twofold: First, we investigate the applicability of NeRFs for aerial image blocks representing different characteristics like nadir-only, oblique and high-resolution imagery. Second, during these investigations we demonstrate the benefit of integrating depth priors from tie-point measures, which are provided during presupposed Bundle Block Adjustment. Our work is based on the state-of-the-art framework VolSDF, which models 3D scenes by signed distance functions (SDFs), since this is more applicable for surface reconstruction compared to the standard volumetric representation in vanilla NeRFs. For evaluation, the NeRF-based reconstructions are compared to results of a publicly available benchmark dataset for airborne images.

In this letter, we present a neural field-based real-time monocular mapping framework for accurate and dense Simultaneous Localization and Mapping (SLAM). Recent neural mapping frameworks show promising results, but rely on RGB-D or pose inputs, or cannot run in real-time. To address these limitations, our approach integrates dense-SLAM with neural implicit fields. Specifically, our dense SLAM approach runs parallel tracking and global optimization, while a neural field-based map is constructed incrementally based on the latest SLAM estimates. For the efficient construction of neural fields, we employ multi-resolution grid encoding and signed distance function (SDF) representation. This allows us to keep the map always up-to-date and adapt instantly to global updates via loop closing. For global consistency, we propose an efficient Sim(3)-based pose graph bundle adjustment (PGBA) approach to run online loop closing and mitigate the pose and scale drift. To enhance depth accuracy further, we incorporate learned monocular depth priors. We propose a novel joint depth and scale adjustment (JDSA) module to solve the scale ambiguity inherent in depth priors. Extensive evaluations across synthetic and real-world datasets validate that our approach outperforms existing methods in accuracy and map completeness while preserving real-time performance.

In recent decades, the geospatial domain has benefitted from technological advances in sensors, methodologies, and processing tools to expand capabilities in mapping applications. Airborne techniques (LiDAR and aerial photogrammetry) generally provide most of the data used for this purpose. However, despite the relevant accuracy of these technologies and the high spatial resolution of airborne data, updates are not sufficiently regular due to significant flight costs and logistics. New possibilities to fill this information gap have emerged with the advent of Very High Resolution (VHR) optical satellite images in the early 2000s. In addition to the high temporal resolution of the cost-effective datasets and their sub-meter geometric resolutions, the synoptic coverage is an unprecedented opportunity for mapping remote areas, multi-temporal analyses, updating datasets and disaster management. For all these reasons, VHR satellite imagery is clearly a relevant study for National Mapping and Cadastral Agencies (NMCAs). This work, supported by EuroSDR, summarises a series of experimental analyses carried out over diverse landscapes to explore the potential of VHR imagery for large-scale mapping.

In this paper, we present BAMF-SLAM, a novel multi-fisheye visual-inertial SLAM system that utilizes Bundle Adjustment (BA) and recurrent field transforms (RFT) to achieve accurate and robust state estimation in challenging scenarios. First, our system directly operates on raw fisheye images, enabling us to fully exploit the wide Field-of-View (FoV) of fisheye cameras. Second, to overcome the low-texture challenge, we explore the tightly-coupled integration of multi-camera inputs and complementary inertial measurements via a unified factor graph and jointly optimize the poses and dense depth maps. Third, for global consistency, the wide FoV of the fisheye camera allows the system to find more potential loop closures, and powered by the broad convergence basin of RFT, our system can perform very wide baseline loop closing with little overlap. Furthermore, we introduce a semi-pose-graph BA method to avoid the expensive full global BA. By combining relative pose factors with loop closure factors, the global states can be adjusted efficiently with modest memory footprint while maintaining high accuracy. Evaluations on TUM-VI, Hilti-Oxford and Newer College datasets show the superior performance of the proposed system over prior works. In the Hilti SLAM Challenge 2022, our VIO version achieves second place. In a subsequent submission, our complete system, including the global BA backend, outperforms the winning approach.

Radargrammetry is a useful approach to generate Digital Surface Models (DSMs) and an alternative to InSAR techniques that are subject to temporal or atmospheric decorrelation. Stereo image matching in radargrammetry refers to the process of determining homologous points in two images. The performance of image matching influences the final quality of DSM used for spatial-temporal analysis of landscapes and terrain. In SAR image matching, local matching methods are commonly used but usually produce sparse and inaccurate homologous points adding ambiguity to final products; global or semi-global matching methods are seldom applied even though more accurate and dense homologous points can be yielded. To fill this gap, we propose a hierarchical semi-global matching (SGM) pipeline to reconstruct DSMs in forested and mountainous regions using stereo TerraSAR-X images. In addition, three penalty functions were implemented in the pipeline and evaluated for effectiveness. To make accuracy and efficiency comparisons between our SGM dense matching method and the local matching method, the normalized cross-correlation (NCC) local matching method was also applied to generate DSMs using the same test data. The accuracy of radargrammetric DSMs was validated against an airborne photogrammetric reference DSM and compared with the accuracy of NASA’s 30 m SRTM DEM. The results show the SGM pipeline produces DSMs with height accuracy and computing efficiency that exceeds the SRTM DEM and NCC-derived DSMs. The penalty function adopting the Canny edge detector yields a higher vertical precision than the other two evaluated penalty functions. SGM is a powerful and efficient tool to produce high-quality DSMs using stereo Spaceborne SAR images.