Mesh is one of the most commonly utilized data formats for digital three-dimensional models in most existing 3-D applications. Recently, online mesh reconstruction from light detection and ranging (LiDAR) measurements has garnered significant interest because of its high efficiency. However, due to the lack of adaptability in adjusting vertex density, existing methods tend to generate either over-represented planar mesh or under-represented non-planar mesh. To address this issue, we propose a novel online mesh reconstruction method with a self-adaptive strategy which, respectively, processes planar and non-planar regions according to local geometric features. For planar regions, we propose a two-step points decimation and mesh reconstruction algorithm to reduce data redundancy based on the observation that the geometric structure of these regions is simple and can be represented by a few key vertices and triangles. For non-planar regions, we design a parallel direct meshing (PDM) algorithm with hole filling mechanism to model objects with complex geometric structure. Moreover, we propose a zipper-based connection strategy to handle the boundaries between planar and non-planar mesh regions. Experimental results demonstrate that our approach outperforms several state-of-the-art algorithms in terms of mesh quality and memory consumption. Remarkably, the entire process is capable of running in real-time on a standard desktop CPU. Code is available at https://github.com/Neo-cyber-hubb/LGFaware-Meshing.

Semantic segmentation of urban meshes plays an increasingly crucial role in the analysis and understanding of 3D environments. Most existing large-scale urban mesh semantic segmentation methods focus on integrating multi-scale local features but struggle to model long-range dependencies across facets effectively. Furthermore, owing to high computational complexity or excessive pre-processing operations, these methods lack the capability for the efficient semantic segmentation of large-scale urban meshes. Inspired by Mamba, we propose MambaMeshSeg-Net, a novel 3D urban mesh semantic segmentation method based on the State Space Model (SSM). The proposed method incorporates a hybrid scanning strategy that adaptively scans 3D urban meshes to extract long-range dependencies across facets, enhancing semantic segmentation performance. Moreover, our model exhibits faster performance in both inference and pre-processing compared to other mainstream models. In comparison with existing state-of-the-art (SOTA) methods, our model demonstrates superior performance on two widely utilized open urban mesh datasets.

Cross-view geo-localization aims to determine the geographic origin of street-view images by matching them with a repository of remote sensing (RS) images equipped with GPS tags. Due to the substantial dissimilarities in viewpoint and visual appearance between street-view and RS images, this task is highly challenging. Recently, many deep-learning-based methods have been proposed. The CNN-based cross-view geo-localization techniques often employ polar transform and fail to capture extensive spatial correlations. And traditional transformer-based methods are prone to losing fine-grained details as a result of downsampling. Addressing these challenges, we propose a novel multi-scale transformer architecture for cross-view geo-localization in this study. Our model leverages multi-scale feature extraction to bolster the accuracy of image matching. We have conducted experiments on an open benchmark, demonstrating our approach's superiority over the state-of-the-art methods.

3-D building models are one of the most intuitive and widely used forms for understanding urban buildings. Generating 3-D building models based on a single off-nadir satellite image is an economical and rapid method, particularly valuable in large-scale 3-D reconstruction scenarios with limited time. In this article, we propose a novel pipeline for automatically reconstructing level of detail 1 (LoD1) 3-D building models based on a single off-nadir satellite remote sensing image. Our pipeline is built upon a multitask neural network called off-nadir building reconstruction network (ONBuildingNet), which extracts building roof polygons and offsets from the image. Using this information, the pipeline computes the building footprint polygons and heights, constructs LoD1 building models, and then extract textures from the off-nadir image. ONBuildingNet introduces our proposed cross-field auxiliary task and multiscale mask head to extract building roof polygons with accurate shapes. We have demonstrated through extensive experiments that our pipeline can automatically and rapidly construct LoD1 3-D urban building models. In addition, our proposed ONBuildingNet outperforms current state-of-the-art methods in extracting more shape accurate building roof polygons, thereby enhancing the accuracy of the final 3-D models produced by our pipeline. Experimental results demonstrate that our method for reconstructing 3-D models of urban building scenes has strong visualization effects, with an average height error of 3.3 m.

Online maps are of great importance in modern life, especially in commuting, traveling and urban planning. The accessibility of remote sensing (RS) images has contributed to the widespread practice of generating online maps based on RS images. The previous works leverage an idea of domain mapping to achieve end-to-end remote sensing image-to-map translation (RSMT). Although existing methods are effective and efficient for online map generation, generated online maps still suffer from ground features distortion and boundary inaccuracy to a certain extent. Recently, the emergence of diffusion models has signaled a significant advance in high-fidelity image synthesis. Based on rigorous mathematical theories, denoising diffusion models can offer controllable generation in sampling process, which are very suitable for end-to-end RSMT. Therefore, we design a novel end-to-end diffusion model to generate online maps directly from remote sensing images, called MapGen-Diff. We leverage a strategy inspired by Brownian motion to make a trade-off between the diversity and the accuracy of generation process. Meanwhile, an image compression module is proposed to map the raw images into the latent space for capturing more perception features. In order to enhance the geometric accuracy of ground features, a consistency regularization is designed, which allows the model to generate maps with clearer boundaries and colorization. Compared to several state-of-the-art methods, the proposed MapGen-Diff achieves outstanding performance, especially a 5% RMSE and 7% SSIM improvement on Los Angeles and Toronto datasets. The visualization results also demonstrate more accurate local details and higher quality.

Road extraction from remote sensing imagery is a popular and frontier research focus, since road information plays an essential role in application fields, such as urban management, map updating and traffic planning. Deep learning based methods have shown their dominance on road extraction from remote sensing imagery. However, the performance of the existing road extraction methods relies heavily on a large amount of high -quality annotated training data, which is usually hard to obtain in practice. Current semi-supervised road extraction models can effectively reduce the dependency on the labeled data, nevertheless they cannot fully utilize the latent information of low-confidence pixels in pseudo-labels effectively. These pixels are usually a border between road and non-road area and of importance for the prediction accuracy of road extraction models. In order to address these issues, we proposed a novel semi-supervised road extraction network, named SemiRoadExNet. SemiRoadExNet is based on a Generative Adversarial Network (GAN), containing one generator with two discriminators. Firstly, both labeled and unlabeled images are put into the generator network for road extraction, and the outputs of the generator not only include road segmentation results but also the corresponding entropy maps. The entropy maps represent the confidence of prediction (road or non-road) for each pixel. Then, the two discriminators enforce the feature distributions keeping the consistency of road prediction maps and entropy maps between the labeled and unlabeled data. During the adversarial training, the generator is continuously regularized by exploiting the potential information from unlabeled data, thus the generalization capacity of the proposed model can be improved effectively. Compared to several state-of-the-art semi-supervised semantic segmentation methods, the proposed SemiRoadExNet achieves 0.96-5.38% IoU improvements on DeepGlobe Road Extraction, Massachusetts Roads and CHN6-CUG datasets respectively. The source code of d SemiRoadExNet is freely available at https://github.com/hchen118/SemiRoadExNet.

Remote sensing (RS) image semantic segmentation has attracted much attention due to its wide applications. However, deep learning-based RS image semantic segmentation methods usually require substantial manual pixelwise annotations, which are expensive and hard to obtain in practice. Although the existing semisupervised RS semantic segmentation methods effectively reduce dependence on labeled data, they generally focus on information consistency between labeled and unlabeled images, but ignore the potential context information between different areas of the RS image. In fact, the objects contained in an RS image usually have some long-range dependence between each other, since trees are usually on both sides of a road, and the middle of two rows of houses is commonly a road. Therefore, we believe that the potential dependencies between different areas of the RS image should be beneficial to reduce the label dependence of RS semantic segmentation. Based on this point, we propose a novel semisupervised RS image semantic segmentation network named SegMind, which is based on mean-teacher (MT) architecture and adopts masked image modeling (MIM) to enhance information interactions of different areas. Moreover, contrastive learning (CL) and entropy loss are introduced to SegMind framework to further improve the linear separability and prediction confidence of the proposed model. Experiments on three datasets have demonstrated the superiority of the proposed method over the state-of-the-art methods. The code is available at https://github.com/lzh-ggs-ddu/SegMind .

As an important branch of multi-objective optimization,preference-based multi-objective evolutionary algorithms have been widely used in scientific researches and engineering practices,which have important research significance.In order to obtain the extreme solutions and the knee solution with the most compromised performance over each optimization objective in multi-objective optimization problems,a definition of knee solution based on Chebyshev distance and its geometric interpretation is presented.Based on the definition,a multi-objective evolutionary algorithm HP-NSGA-II aiming to search for the extreme solutions and the knee solution is proposed.The regional dynamic updating strategy of the proposed algorithm updates the target regions dynamically in each iteration,and finally converges to the target regions.The strategy of maintaining the balance between regions ensures the balance of the number of individuals in each region,so that the individuals could be distributed evenly in each region.Based on widely used test functions,sufficient experimental verification is carried out,and the experimental results indicate that HP-NSGA-II algorithm can achieve better performance in terms of convergence,regional balance and regional controllability in two-dimensional and three-dimensional test problems,and can accurately obtain the extreme solutions and knee solution.

Change detection (CD) is one of the important applications of remote sensing and plays an important role in disaster assessment, land use detection, and urban sprawl tracking. High-accuracy fully supervised methods are the main methods for CD tasks at present. However, these methods require a large amount of labeled data consisting of bi-temporal images and their change maps. Moreover, creating change maps takes a lot of labor and time. To address this limitation, a simple semi-supervised change detection method based on consistency regularization and strong augmentation is proposed in this paper. First, we construct a Siamese nested UNet with graph attention mechanism (SANet) and pre-train it with a small amount of labeled data. Then, we feed the unlabeled data into the pre-trained SANet and confidence threshold filter to obtain pseudo-labels with high confidence. At the same time, we produce distorted images by performing strong augmentation on unlabeled data. The model is trained to make the CD results of the distorted images consistent with the corresponding pseudo-label. Extensive experiments are conducted on two high-resolution remote sensing datasets. The results demonstrate that our method can effectively improve the performance of change detection under insufficient labels. Our methods can increase the IoU by more than 25% compared to the state-of-the-art methods.

Maps can help governments in infrastructure development and emergency rescue operations around the world. Using adversarial learning to generate maps from remote sensing images is an emerging field. As we now know, the urban construction styles of different cities are diverse. The current translation methods for remote sensing image-to-map tasks only work on the specific regions with similar styles and structures to the training set and perform poorly on previously unseen areas. We argue that this greatly limits their use. In this work, we intend to seek a remote sensing image-to-map translation model that approaches the challenge of generating maps for the remote sensing images of unseen areas. Our remote sensing image-to-map translation model (RSMT) achieves universal and general applicability to generate maps over multiple regions by combining adversarial deep transfer training schemes with novel attention-based network designs. Extracting the content and style latent features from remote sensing images and a series of maps, respectively, RSMT generalizes a pattern applied to the remote sensing images of new areas. Meanwhile, we introduce feature map loss and map consistency loss to reinforce generated maps’ precision and geometry similarity. We critically analyze qualitative and quantitative results using widely adopted evaluation metrics through extensive validation and comparisons with previous remote sensing image-to-map approaches. The results of experiment indicate that RSMT can translate remote sensing images to maps better than several state-of-the-art methods.

Effective and fast Earth Observation Satellite (EOS) scheduling plays an essential role in new generation space-based information services, especially in emergency scenarios. Most existing research regards the EOS scheduling as a combinatorial optimization problem and adopts metaheuristic local search-based methods to solve it in a batch-wise fashion. To the best of our knowledge, no practical scheduling approach has been able to generate an optimized multi-EOS observation plan in an immediate response style. This paper proposes a novel real-time multi-satellite scheduling method consisting of a machine learning-based hierarchical prediction model and a heuristic local search algorithm. Firstly, the hierarchical prediction model based on a stacked multi-channel transformer network can learn from existing historical multi-EOS observation plans and generate a high-quality initial solution for the current scheduling scenario. Then, the local search algorithm based on Random Hill Climbing does further improvement and heuristic constraint handling on the initial solution and generates the final solution. To verify the effectiveness of the proposed method, simulation experiments are carried out. The results show that the method proposed in this paper consumes a short calculation time and has high planning profits compared with the state-of-the-art approaches.

近年来,随着电子技术的发展,地面站测控设备与数传设备逐渐趋同,呈现出功能一体化的特性,充分利用该特性可有效提高地面站设备资源的利用率,缓解星地通信中地面站资源相对匮乏的现实难题.针对问题特征和实际需求,建立了以最小化任务冲突时长、最大化天线负载均衡度以及最大化任务集聚度为优化目标的约束满足模型,提出了面向测控数传资源一体化场景的卫星地面站资源规划多目标优化算法KG-NSGA-Ⅱ-TTC&DT.该算法针对优化目标设计了负载均衡算子、任务集聚算子以及迭代修复冲突消解算子,并以膝点引导算法进程,有效提升了问题求解的优化性和针对性.实验结果表明,与NSGA-Ⅱ-TTC&DT算法相比,KG-NSGA-Ⅱ-TTC&DT在世代距离(GD)指标上实现了16.75％的平均性能提升,在最小化任务冲突时长、最大化天线负载均衡度以及最大化任务集聚度3个优化目标上分别实现了 6.67％、9.28％以及1.87％的平均优化性能提升,其中负载均衡算子、任务集聚算子以及迭代修复冲突消解算子的优化性能贡献率分别为31.50％、15.60％、70.57％.

Satellite onboard autonomous task planning is one of the key technologies for the operation of earth observation satellites,which has received great attention from researchers in recent years.Considering the limited computing resources,as well as the dynamic changes of observation tasks and resource onboard,the heuristic search algorithms are mainly used to solve the satellite onboard task planning problem,and the optimization of solution needs to be improved.Firstly,a new sequential decision-ma-king framework for observation tasks is constructed in this paper.Based on this framework,an earth observation satellite can decide the observation task to be performed in real-time,without generating a plan in advance.Then,an observation task decision model based on attention mechanism,and the corresponding input feature representation method and model training method are designed.An observation task sequence algorithm based on attention neural network is proposed.Finally,based on a set of random data,the performance of the proposed algorithm,two deep learning algorithms and two heuristic online search algorithms are compared.Experimental results show that the response time of the proposed method is less than one-fifth of the existing deep learning algorithm,and the profit gap is much smaller than that of the heuristic search algorithms,which confirm the feasibility and effectiveness of our method.

Change detection is one of the main applications of remote sensing images. Pixel-to-pixel change detection using deep learning has been a hot research spot. However, the current approach are not effective enough to fuse deep semantic features and raw spatial information, and the network does not have the ability to perform long-distance information aggregation due to the limitation of the convolutional kernel size. In this manuscript, we propose a Siamese UNet with a dense attention mechanism, named SUDANet to do change detection for remote sensing images. SUDANet add a channel attention mechanism and a self-attention mechanism to the dense skip connection between encoder and decoder which enable the model to fuse feature information in channel dimensions and spatial dimensions. Graph attention module is also added at the end of the encoder, enabling the model to perform correlation analysis and long-distance aggregation of deep semantic features. The experimental results on LEVIR dataset show that our method outperforms the state-of-the-art change detection methods.

如何根据应急航天任务,在满足各种约束的前提下,从型号众多、能力各异的卫星平台及有效载荷中快速决策效费比最佳的卫星构造方案引起了越来越多的关注,这就是适应任务的模块化卫星快速构建优化决策问题.在深入分析该问题特点的基础上,建立优化决策数学模型,提出基于遗传算法的优化决策方法,为处理遗传算法迭代过程中产生的不可行解引入基于罚函数法的约束处理方法.针对罚函数法中惩罚系数难于确定的特点,设计惩罚系数自适应调整的动态罚函数机制.定量化实验及分析结果表明:该方法能有效解决适应任务的模块化卫星快速构建优化决策问题.

Multi-step station-level crowd flow prediction (Ms-SLCFP) is to predict the count of people that would depart from or arrive at subway/bus/bike stations in multiple future consecutive time periods. By providing a long term view, it benefits the decision making in related applications, such as public safety, traffic management, etc. However, performing Ms-SLCFP is challenging as complicated spatiotemporal correlations are formed among stations due to the flowing crowd. Besides, the crowd flow at a single station fluctuates a lot though the regularity is obvious at the regional level. To tackle such issues, we propose a deep neural networks-based model with spatiotemporal hierarchical attention mechanisms, called ST-HAttn for short, for Ms-SLCFP. The notable contributions are that ST-HAttn performs attention mechanisms (AM) in two ways: 1) implementing AM at both station level and regional level; 2) implementing AM to explicitly model the pairwise correlations of station-region instead of station-station. The intuition is to alleviate the negative impact on Ms-SLCFP due to the fluctuation of the crowd flow at the station level. Verified on three real-world datasets, ST-HAttn outperforms the state-of-the-art methods in terms of Ms-SLCFP.

Onboard task planning can enhance the responsiveness of satellite to dynamic changes, which has attracted widespread attention. In this paper, the Satellite Onboard Observation Task Planning (SOOTP) problem is studied, and a data-driven onboard planning approach is proposed to decide the observation task to execute in real-time using machine learning techniques. In the approach, the satellite can learn how to make optimal decisions from the historical planning results. What is more, we design five types of features and employ three ensemble learning algorithms to solve the SOOTP. A comparison of the proposed method against two online searching algorithms indicates that the former has smaller profit gap and shorter response time, which verify the feasibility of our method.

Map is an essential medium for people to understand our changing planet. Recently, research on generating and updating maps through remote sensing images has been an important and challenging task in geographic information. Traditional methods for map generation are time-consuming and labor-intensive. Besides, most supervised learning methods for map generation lack labeled training samples. It is challenging to generate maps quickly and efficiently for emergency rescue operations such as earthquakes, fire disasters, or tsunami. In this article, we propose an unsupervised domain mapping model based on adversarial learning called MapGen-GAN. MapGen-GAN is a generative adversarial network (GAN) that can do end-to-end translation from remote sensing images to general map quickly, and trained with no human annotation data. In order to improve the fidelity and the geometry precision of generated maps, we employ circularity-consistency and geometrical-consistency constraints as a part of the loss function of the proposed model. And then, an improved residual block Unet is designed and adopted as the generator of MapGen-GAN to capture the geographic structure information of buildings, roads, and topography outlines under different resolutions in the map generation. By applying the proposed model to two distinct datasets, experiments demonstrate that our model can generate maps efficiently and quickly and outperform the state-of-the-art approaches.

In taxi dispatch systems, predicting citywide passenger pickup/dropoff demand is indispensable for developing effective taxi distribution and scheduling strategies to resolve the demand-service mismatch. Compared with predicting next-step only, predicting multiple steps is preferable since it can provide a long term view, thus preventing short-sighted strategies. However, multi-step citywide passenger demand prediction (MsCPDP) is challenging due to the complicated spatiotemporal correlations in the distribution of passenger demand and the lack of ground truth from pre-steps for the prediction of subsequent steps. In this paper, a deep-learning-based prediction model with spatiotemporal attention mechanism is proposed for MsCPDP. The model, called ST-Attn, follows the general encoder-decoder framework for modelling sequential data but adopts a multiple-output strategy without recurrent neural network units. The spatiotemporal attention mechanism learns to determine the focus on those parts of the city at certain periods that are more relevant to the passenger demand in the predicted region and time period. In addition, a pre-predicted result calculated by spatiotemporal kernel density estimation is fed to ST-Attn, which provides a reference for further accurate prediction. Experiments on three real-world datasets are carried out to verify ST-Attn's performance, and the results show that ST-Attn outperforms the baselines in terms of MsCPDP. (C) 2019 Published by Elsevier Inc.

The scheduling of Earth Observation Satellite (EOS) data transmission is a complex combinatorial optimization problem. With the development of remote sensing applications, a new special requirement named data transmission oriented to topics has appeared. It supposes that the data obtained from each observation activity by satellites belong to certain observation data topics, and every observation data topic has completeness and timeliness requirements. Unless all of the observation data belonging to one topic has been transmitted to the ground before the expected time, the value of the observation data will be decayed sharply and only a part of the rewards (or even no reward) for the data transmission will be obtained. Current researches do not meet the new data topic transmission requirements well. Based on the characteristics of the problem, a mathematic scheduling model is established, and a novel hybrid scheduling algorithm based on evolutionary computation is proposed. In order to further enhance the performance and speed up the convergence process of our algorithm, a domain-knowledge-based mutation operator is designed. Quantitative experimental results show that the proposed algorithm is more effective to solve the satellite observation data topic transmission scheduling problem than that of the state-of-the-art approaches.