With the increasing complexity of mobile robot systems, multi-robot collaboration has become essential. However, such collaboration places heavy demands on both computational power and communication bandwidth. Cloud computing offers elastic and scalable infrastructure, allowing robotic functions to be encapsulated and delivered as on-demand services. Meanwhile, edge computing provides low-latency processing near the data source. The integration of these paradigms enables efficient coordination between cloud and edge infrastructures, forming a solution for real-time, scalable, and intelligent robotic control. This paper presents a cloud-edge collaborative system for mobile robots. Based on cloud-native technologies, the cloud control platform is implemented on Kubernetes (K8s) to orchestrate containerized control services, enabling dynamic resource allocation and automated deployment. An edge offloading mechanism is designed to support real-time data exchange and task coordination between robots and the cloud. Furthermore, a collaborative mapping and control framework is developed, where computationally intensive tasks such as map construction and trajectory generation are offloaded to the cloud, while real-time control and data pre-processing are executed at the edge. Experimental results in outdoor environments verify that the proposed framework significantly enhances system scalability and performance. These findings confirm the feasibility and generality of the proposed approach, highlighting its strong potential for large-scale deployment in cloud-based robotic systems.

While significant advances exist in pseudo-label generation for semi-supervised semantic segmentation, pseudo-label selection remains understudied. Existing methods typically use fixed confidence thresholds to retain high-confidence predictions as pseudo-labels. However, these methods cannot cope with network overconfidence tendency, where correct and incorrect predictions overlap significantly in high-confidence regions, making separation challenging and amplifying model cognitive bias. Meanwhile, the direct discarding of low-confidence predictions disrupts spatial-semantic continuity, causing critical context loss. We propose Confidence Separable Learning (CSL) to address these limitations. CSL formulates pseudo-label selection as a convex optimization problem within the confidence distribution feature space, establishing sample-specific decision boundaries to distinguish reliable from unreliable predictions. Additionally, CSL introduces random masking of reliable pixels to guide the network in learning contextual relationships from low-reliability regions, thereby mitigating the adverse effects of discarding uncertain predictions. Extensive experimental results on the Pascal, Cityscapes, and COCO benchmarks show that CSL performs favorably against state-of-the-art methods. Code and model weights are available at https://github.com/PanLiuCSU/CSL.

The occlusion problem has consistently posed a significant challenge in the field of segmentation. Most existing segmentation methods require additional annotations and fail to capture the contour information of occluded regions, thus not truly addressing the occlusion issue. Although segmentation tasks involving particle objects also suffer from occlusion problems, the homogeneity of particle objects offers new possibilities for overcoming this challenge. In this paper, we propose an occlusion segmentation framework for particle objects that does not require additional annotations. This framework only necessitates instance-level segmentation labels to obtain complete contour information of particle objects, including occluded regions. First, we decompose the occlusion segmentation task into a generic instance segmentation task and an occlusion repair task for occluded objects. Then, in order to train the occlusion repair model with only instance segmentation-level labels, we quantitatively analyze the occlusion phenomenon, including the mathematical descriptions of occlusion relationships, degrees, and distributions. Next, we geometrically transform and layer overlay the unobscured samples to construct occlusion samples containing labeling information of the occluded regions. These sample sets are used to train a generative model that predicts the contour information of occluded regions. Finally, we fine-tune or post-process the pre-segmentation model with the particle objects containing restored complete contour information to achieve the final occlusion segmentation. We conducted extensive ablation experiments on both the ore-particle dataset and publicly available cell-particle datasets. The experimental results validate the effectiveness, accuracy, and generalizability of our method. Note to Practitioners—Particle segmentation has been faced with the occlusion problem. In this paper, inspired by the similarity between particle objects, we propose a self-supervised occlusion segmentation framework that does not require additional annotation of occlusion layers. Our approach requires only instance segmentation level annotation without more complex additional manual annotation, which is crucial for practical applications. In addition, we decouple the complex occlusion relation modeling into a binary classification problem without knowing precisely the occlusion hierarchy between particles, which further reduces the difficulty of practical applications. Then, we also propose shading transformations to characterize the inter-particle shading distribution to construct shading sample sets from existing samples. Finally, we use these learned and constructed occlusion sample sets to pre-train the generative model for regenerating the occluded objects to complete the final occlusion segmentation. Although in a generic segmentation task, our approach may have some limitations because the segmented objects may not have an apparent similarity. However, our approach using self-supervision and the objects’ properties provides valuable ideas for solving the occlusion problem. In the future, we will solve the occlusion problem regarding the properties of each class of objects rather than just considering the similarity among particle objects.

The absorption and scattering of light in different turbid media cause images to suffer from poor visibility and contrast, which severely affects the performance of many computer vision tasks. To address this issue, we propose a fast scene recovery method based on the Ambient light similarity prior (ALSP). In this method, the ambient light similarity metric is designed from both magnitude and orientation, which is embedded into the optical imaging model, and the estimation of scene transmission is derived by simplification and approximation. The estimation of the transmission map is very simple, and its time complexity is O(N), where N is the size of the input image. Moreover, we propose a progressive manner to determine the ambient light for both the near and far regions separately, which can effectively improve the brightness and color saturation of the restored image. Experiments performed in different scenes demonstrate that our method outperforms several state-of-the-art competitors in terms of efficiency and scene recovery performance.

Unlike general image segmentation, highly complex particle images have significant challenges in labeling and segmentation due to the information occlusion and texture disturbance. Aiming at the highly under-segmentation problem caused by complex particle image segmentation, this paper proposes a Semi-supervised Hybrid-training Particle Segmentation framework (SHPS) based on skeleton-guided pseudo-labels. First, a pre-trained model is obtained by training a popular segmentation algorithm on partially labeled data. Then, a Background Skeleton-guided Pseudo-label generation algorithm (BSP) is proposed to generate pseudo-labels closer to the ground truth in terms of structural integrity based on coarse segmentation. The final segmentation model is obtained by training a mixed dataset consisting of labeled data and pseudo-labels from another partition on the pre-trained model. The skeleton differences of pseudo-labels and coarse segmentation are added to the loss function. Experimental results show that our method achieves 84.4% accuracy on mIoU with uniform label data distribution, which is 2.1% higher than the accuracy of UNet and reduces the degree of under-segmentation.

The last two decades have seen a clear trend toward crafting intelligent vehicles based on the significant advances in communication and computing paradigms, which provide a safer, stress-free, and more enjoyable driving experience. Moreover, emerging applications and services necessitate massive volumes of data, real-time data processing, and ultrareliable and low-latency communication (URLLC). However, the computing capability of current intelligent vehicles is minimal, making it challenging to meet the delay-sensitive and computation-intensive demand of such applications. In this situation, vehicular task/computation offloading toward the edge cloud (EC) and vehicular cloudlet (VC) seems to be a promising solution to improve the network’s performance and applications’ Quality of Service (QoS). At the same time, artificial intelligence (AI) has dramatically changed people’s lives. Especially for vehicular task offloading applications, AI achieves state-of-the-art performance in various vehicular environments. Motivated by the outstanding performance of integrating reinforcement learning (RL)/deep RL (DRL) to the vehicular task offloading systems, we present a survey on various RL/DRL techniques applied to vehicular task offloading. Precisely, we classify the vehicular task offloading works into two main categories: 1) RL/ DRL solutions leveraging EC and 2) RL/DRL solutions using VC computing. Moreover, the EC section-based RL/DRL solutions are further subcategorized into multiaccess edge computing (MEC) server, nearby vehicles, and hybrid MEC (HMEC). To the best of our knowledge, we are the first to cover RL/DRL-based vehicular task offloading. Also, we provide lessons learned and open research challenges in this field and discuss the possible trend for future research.