In low-illumination environments, visible light images significantly degrade, and the simultaneous localization and mapping (SLAM) system based on visible cameras fails. In contrast, thermal sensors are not affected by scene lighting conditions and have become a promising alternative. However, due to the problems of low contrast, high noise, and data interruption in thermal images, there are still many challenges in implementing thermal sensor SLAM. Therefore, this paper proposes an edge-aided monocular thermal sensor SLAM system, called EAMT-SLAM, suitable for low illumination environments. Firstly, the edge provides more feature associations to assist the SLAM system in initializing quickly and accurately. Subsequently, the saliency map is calculated based on the strong edge in the thermal image, and the weights are reasonably assigned to the feature points in the optimization process of pose estimation to obtain more accurate results. Finally, this paper proposes an edge-guided loop detection method, which achieves higher accuracy and real-time performance compared with the appearance-based method, effectively ensuring the global consistency of SLAM. In addition, this paper proposes a new thermal image processing method that can enhance image details and remove complex noise in real time, significantly improve the quality of thermal images, and broaden the application scope of the EAMT-SLAM system. The experimental results show that our method can obtain the highest quality thermal image under quantitative and qualitative evaluation, complete accurate loop detection, and finally realize accurate pose estimation in low-illumination environments.

Visual Simultaneous Localization and Mapping (VSLAM) is significant in unmanned driving, being is used to locate vehicles and create environmental maps, and provides a basis for navigation and decision making. However, in inevitable dark night environments, the SLAM system still suffers from a decline in robustness and accuracy. In this regard, this paper proposes a VSLAM pipeline called DarkSLAM. The pipeline comprises three modules: Camera Attribute Adjustment (CAA), Image Quality Enhancement (IQE), and Pose Estimation (PE). The CAA module carefully studies the strategies used for setting the camera parameters in low-illumination environments, thus improving the quality of the original images. The IQE module performs noise-suppressed image enhancement for the purpose of improving image contrast and texture details. In the PE module, a lightweight feature extraction network is constructed and performs pseudo-supervised training on low-light datasets to achieve efficient and robust data association to obtain the pose. Through experiments on low-light public datasets and real-world experiments in the dark, the necessity of the CAA and IQE modules and the parameter coupling between these modules are verified, and the feasibility of DarkSLAM is finally verified. In particular, the scene in the experiment NEU-4am has no artificial light (the illumination in this scene is between 0.01 and 0.08 lux) and the DarkSLAM achieved an accuracy of 5.2729 m at a distance of 1794.33 m.

Visual place recognition (VPR) is the process of identifying previously visited places using visual information. It is crucial for a robot to achieve fast and accurate VPR under appearance and viewpoint changes. Inspired by human perception intuition, this paper proposes a lightweight semantic image descriptor called Forest for robust VPR. Specifically, for each semantic object in the image, three most recognizable reference objects are first found and ranked in descending order according to a customized measure. And then, categories, position and area of the semantic object itself, along with the categories of the three recognized reference objects, are used to encode the local descriptor of the object. As a result, the set of local descriptors is used as the global descriptor of the image. Performance of the proposed descriptor is evaluated and compared with six state-of-the-art VPR methods, Ibow-LCD, CoHOG, NetVLAD, LOST-X, Region VLAD, Patch-NetVLAD, over four public datasets of urban scenes, Extended-CMU Season, RobotCar Seasons v2, SYNTHIA and Kitti, using three Precision-Recall based metrics, AUC(Area Under Curve), Recall@100%Precision and Precision@100%Recall. The results show that our method maintains competitive accuracy in scenes with appearance and viewpoint changes and exhibits high robustness. At the same time, our method is robust to the noise of semantic segmentation and exhibits high real-time performance.

Monocular SLAM has gained widespread popularity due to its numerous advantages. However, this approach is susceptible to appearance changes and often experiences problems such as sensitivity of dynamic objects and scale drift. Using semantics as an advanced visual feature is highly effective since it remains unaffected by appearance and viewpoint changes. This makes it easier to establish a more robust data association. In this paper, we represent the semantic information of features as the semantic probability that they belong to each category, thus reducing the adverse effects of semantic segmentation noise and increasing the ability of semantic representation. Furthermore, we propose a method that combines the semantic probability with geometric features to enhance the accuracy and robustness of the SLAM system. Specifically, the odometry module chooses geometric-semantic feature point descriptors, which help achieve higher accuracy in feature matching and eliminate the influence of dynamic features. It also provides an initial estimation of the real scale. The back-end optimization module utilizes semantic probability to establish medium-term constraints, effectively mitigating the cumulative drift of attitude and scale. The loop detection module constructs global and local descriptors using semantic information and incorporates the DBOW2 algorithm based on geometric features, enabling accurate and robust loop detection. The experimental results demonstrate that the proposed algorithm achieves high-performance pose estimation on the test sequences compared to other advanced monocular SLAM methods. Compared with DBOW2 based on appearance, the AUC (area under the curve) of our loop detection method is increased by 28.1% while remaining lightweight.

In practical applications, mobile robots often return to the same place from opposite directions at different times, resulting in opposite viewpoints and appearance changes. In this intricate environment, the successful generation of visual place recognition (VPR) is crucial for tasks such as simultaneous localization and mapping (SLAM), geo-localization, and the navigation of mobile robots. Inspired by the successful migration of animals by combining magnetic field information, this paper compactly encodes the categories and three-dimensional positions of visual semantic objects by combining the simple and reliable geomagnetic direction provided by geomagnetic sensors, and generates lightweight single-frame image descriptors based on the geomagnetic direction and distance between semantic objects, so as to realize the opposite-viewpoint VPR task with only single-frame image information in the scene with changes in appearance. The performance of the proposed method is assessed in comparison to fourteen cutting-edge VPR methods, employing three Precision-Recall metrics: Area Under Curve (AUC), Recall at 100% Precision, and Precision at 100% Recall. The results indicate that while many other methods decline in performance, our approach consistently exhibits superior results in scenarios with significant viewpoint and appearance variations. In the test datasets, our method obtains the highest AUC value on average, which balances the precision and recall well. Additionally, the robustness and real-time performance of our method are evaluated, demonstrating consistently high levels in both aspects.

In recent years, semantic segmentation has made significant progress in visual place recognition (VPR) by using semantic information that is relatively invariant to appearance and viewpoint, demonstrating great potential. However, in some extreme scenarios, there may be semantic occlusion and semantic sparsity, which can lead to confusion when relying solely on semantic information for localization. Therefore, this paper proposes a novel VPR framework that employs a coarse-to-fine image matching strategy, combining semantic and appearance information to improve algorithm performance. First, we construct SemLook global descriptors using semantic contours, which can preliminarily screen images to enhance the accuracy and real-time performance of the algorithm. Based on this, we introduce SemLook local descriptors for fine screening, combining robust appearance information extracted by deep learning with semantic information. These local descriptors can address issues such as semantic overlap and sparsity in urban environments, further improving the accuracy of the algorithm. Through this refined screening process, we can effectively handle the challenges of complex image matching in urban environments and obtain more accurate results. The performance of SemLook descriptors is evaluated on three public datasets (Extended-CMU Season, Robot-Car Seasons v2, and SYNTHIA) and compared with six state-of-the-art VPR algorithms (HOG, CoHOG, AlexNet_VPR, Region VLAD, Patch-NetVLAD, Forest). In the experimental comparison, considering both real-time performance and evaluation metrics, the SemLook descriptors are found to outperform the other six algorithms. Evaluation metrics include the area under the curve (AUC) based on the precision–recall curve, Recall@100%Precision, and Precision@100%Recall. On the Extended-CMU Season dataset, SemLook descriptors achieve a 100% AUC value, and on the SYNTHIA dataset, they achieve a 99% AUC value, demonstrating outstanding performance. The experimental results indicate that introducing global descriptors for initial screening and utilizing local descriptors combining both semantic and appearance information for precise matching can effectively address the issue of location recognition in scenarios with semantic ambiguity or sparsity. This algorithm enhances descriptor performance, making it more accurate and robust in scenes with variations in appearance and viewpoint.