Inferring the actual road segment purely based on one positioning point, known as single-point map matching (SMM), is vital for many urban applications, e.g., ride-hailing and geo-tagging. However, it is challenging due to inherent positioning errors and extrinsic heterogeneous environments. Existing methods either overlook the heterogeneity of different regions, or do not exploit the commonality of different matching tasks. In this paper, we treat each region as an individual SMM task to tackle the heterogeneity, and propose Spatial Hierarchical Meta-Learning for SMM (SHSMM) to learn the shared knowledge across tasks. SHSMM is equipped with a Dual-view Map Matcher to perform the matching, which can perceive the knowledge of road segments globally. To learn the task-specific model parameters, SHSMM modulates initial parameters and scales the local update learning rate based on hierarchical geographical and semantic knowledge about spatial tasks. A local update learning rate scheduling strategy is further proposed to facilitate the meta-training. Extensive experiments as well as case studies based on two real-world datasets demonstrate the effectiveness of the proposed method.

The Lifelong Planning A* (LPA*) algorithm demonstrates unique advantages in dynamic pathfinding by employing incremental update techniques that reuse previously computed search results, enabling rapid path replanning in dynamic road network environments. However, when changes occur near the starting point or specific positions, such as key intersections, bottlenecks, or locations with high connectivity within the path network, even minor changes can trigger significant adjustments, which significantly increase the re-routing search space and computational costs of LPA*. To address this limitation, we propose a Hierarchical Multi-target LPA* (HMLPA*) algorithm that partitions the indoor path network into multiple subgraphs using the METIS graph partitioning algorithm and constructs an abstract trunk graph based on the key nodes of these subgraphs, thereby forming a hierarchical structure for the indoor path network. By leveraging this hierarchical structure, The HMLPA*’s subalgorithm, Multi-target LPA* (MLPA*), initiates pathfinding and confines re-routing to affected subgraphs and the abstract trunk graph. This localized re-routing approach effectively limits the search scope and significantly reduces computational overhead. Experimental results demonstrate that HMLPA* substantially outperforms LPA* in rerouting efficiency, effectively mitigating the high computational costs associated with the dynamic computational environment of the indoor path network.

Rotation object detection is of great importance in remote sensing imagery where the orientation is arbitrary and objects are densely distributed. However, there are several challenges that need to be overcome, such as the angular boundary problem in the regression-based methods and the square-like problem in the classification-based methods. For square-like object rotation detection, classification-based methods [e.g., circular smooth label (CSL)] suffer from inconsistencies between angular coding and evaluation mechanisms due to the variation of aspect ratio. To address the angular inconsistencies of square-like object existing in current classification methods, we design a novel angular encoding mechanism based on aspect ratio. We make optimizations and improvements in the following two aspects: 1) proposing an aspect ratio-based bidirectional coded label (AR-BCL) to replace CSL for angle coding of square-like object, which significantly improves detection accuracy for square-like object and ii) designing a cross-fusion decoupled head (CF-DH) based on angular classification to replace the existing coupled head (CH), which can help extract features that suitable for angular classification. Extensive experiments on DOTA, a large-scale public dataset for aerial images, demonstrate the effectiveness of our method for square-like object detection.

In order to meet the demand of urban road information acquisition in the field of urban road rapid update and intelligent transportation, this paper proposes an urban road extraction method based on floating car trajectory clustering. The proposed method considers the influence of the track point position and direction Angle, identifies the road contour by path clustering based on position and direction Angle, and then extracts the road skeleton line based on Delaunay triangulation network. Experimental results show that the proposed method has a high coverage rate and a low rate of missed detection, and can achieve accurate extraction of urban roads, which is helpful for the maintenance of urban road information and rapid update of road network based on floating car data.

Anomalous trajectory detection is an important research content in the field of trajectory data mining. The anomaly detection method based on Top-k evolving trajectory outlier detection (TOP-EYE) is an effective anomaly detection method for moving objects that adopts evolutionary computation to detect anomalies from two perspectives of density and direction. However, the TOP-EYE method only considers two factors of density and direction, ignoring the multi-dimensional semantic features of trajectory data. In order to improve the accuracy of anomaly detection of trajectory data, this paper proposes an improved TOP-EYE method integrating semantic features, which adds trajectory semantic features for anomaly trajectory detection based on TOP-EYE. Experiments show that the proposed improved method integrating semantic features can more accurately identify anomalous trajectories than TOP-EYE, which is helpful for mining potential trajectory outliers with anomalous semantic features.

Abstract Acquired above variable clouds, aerial images contain the components of ground reflection and cloud effect. Due to the non‐uniformity, clouds in aerial images are even harder to remove than haze in terrestrial images. This paper proposes a divide‐and‐conquer scheme to remove the thin translucent clouds in a single RGB aerial image. Based on colour attenuation prior, we design a kind of veiling metric that indicates the local concentration of clouds effectively. By this metric, an aerial image containing thickness‐varied clouds is segmented into multiple regions. Each region is veiled by clouds of nearly‐equal concentration, and hence subject to common assumptions, such as boundary constraint on transmission. The atmospheric light in each region is estimated by the modified local colour‐line model and composed into a spatially‐varying airlight map for the entire image. Then scene transmission is estimated and further refined by a weighted ‐norm based contextual regularization. Finally, we recover ground reflection via the atmospheric scattering model. We verify our cloud removal method on a number of aerial images containing thin clouds and compare our results with classical single‐image dehazing methods and the state‐of‐the‐art learning‐based declouding method, respectively.

Aiming at satisfying the users' navigation requirements in complex indoor environments, an indoor route planning method with environment awareness is proposed in this paper. Considering the accessibility, simplicity and comfortable experience of user's navigation, the paper first describes the modeling method of indoor navigation network and defines the expressions of indoor environment semantics. And then a navigation cost function is presented, which includes environment semantics such as path complexity, crowded degree and blocking events into indoor navigation network model. Finally, this paper proposes an optimal indoor route planning method by introducing the navigation cost function and environment semantics into traditional Dijkstra algorithm. The experimental results show that the proposed method can effectively enhance the user's comfort experience in indoor navigation.

Summary: Analysis pipelines are an essential part of bioinformatics research, and ad hoc pipelines are frequently created by researchers for prototyping and proof-of-concept purposes. However, most existing pipeline management system or workflow engines are too complex for rapid prototyping or learning the pipeline concept. A lightweight, user-friendly and flexible solution is thus desirable. In this study, we developed a new pipeline construction and maintenance tool, PipelineDog. This is a web-based integrated development environment with a modern web graphical user interface. It offers cross-platform compatibility, project management capabilities, code formatting and error checking functions and an online repository. It uses an easy-to-read/write script system that encourages code reuse. With the online repository, it also encourages sharing of pipelines, which enhances analysis reproducibility and accountability. For most users, PipelineDog requires no software installation. Overall, this web application provides a way to rapidly create and easily manage pipelines. Availability and implementation: PipelineDog web app is freely available at http://web.pipeline.dog. The command line version is available at http://www.npmjs.com/package/pipelinedog and online repository at http://repo.pipeline.dog. Contact: [email protected] or [email protected] or [email protected]. Supplementary information: Supplementary data are available at Bioinformatics online.

MOTIVATION: A proper target or marker is essential in any diagnosis (e.g. an infection or cancer). An ideal diagnostic target should be both conserved in and unique to the pathogen. Currently, these targets can only be identified manually, which is time-consuming and usually error-prone. Because of the increasingly frequent occurrences of emerging epidemics and multidrug-resistant 'superbugs', a rapid diagnostic target identification process is needed. RESULTS: A new method that can identify uniquely conserved regions (UCRs) as candidate diagnostic targets for a selected group of organisms solely from their genomic sequences has been developed and successfully tested. Using a sequence-indexing algorithm to identify UCRs and a k-mer integer-mapping model for computational efficiency, this method has successfully identified UCRs within the bacteria domain for 15 test groups, including pathogenic, probiotic, commensal and extremophilic bacterial species or strains. Based on the identified UCRs, new diagnostic primer sets were designed, and their specificity and efficiency were tested by polymerase chain reaction amplifications from both pure isolates and samples containing mixed cultures. AVAILABILITY AND IMPLEMENTATION: The UCRs identified for the 15 bacterial species are now freely available at http://ucr.synblex.com. The source code of the programs used in this study is accessible at http://ucr.synblex.com/bacterialIdSourceCode.d.zip CONTACT: [email protected] SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.