The increasing complexity of spatial data necessitates advanced research into spatial distributions and the interrelationships of geographical phenomena. Classical geographically weighted regression (GWR) methods struggle with abrupt changes in local ranges, such as sudden variations in property prices or pollution levels. This challenge primarily arises because classical GWR models do not account for relationships between the independent variables and their surrounding environmental attributes, limiting their ability to adapt to complex spatial variations. Herein, an entropy-based high-dimensional attribute similarity GWR method—i.e., Entropy and Dimensional Similarity-based Geographically Weighted Regression (EDSGWR)—is proposed to overcome this limitation. This approach dynamically adjusts regression neighborhoods by computing local entropy, refining spatial regression targets for the GWR model. Next, neighborhood contextual data is incorporated into a high-dimensional attribute matrix, where the Hilbert Schmidt independence criterion is used to measure nonlinear dependencies between a target point and its neighbors. These dependencies generate attribute similarity weights, which are then combined with Euclidean spatial weights to form a hybrid model. Finally, Bayesian optimization determines the optimal neighborhood scale and the number of observations in the attribute matrix. Experiments using four open-source spatial datasets demonstrate that EDSGWR outperforms current methods in accuracy, stability, and interpretability, particularly in cases involving abrupt changes and directional data.

The Airborne LiDAR scanning technology is widely used in various fields due to its ability to quickly acquire geospatial features. As a crucial step in data applications, the automatic acquisition of point-by-point labels has become a hot topic in current research on point cloud data processing. Recent deep-learning algorithms for feature extraction in the form of kernel point convolution are currently considered mainstream due to their outstanding performance. Among these, Kernel Point Convolution (KPConv) is the most representative. Although it has demonstrated good performance in various point-cloud classification tasks, the local window convolution approach overlooks global semantic information and does not focus on the relationship between point distribution and features. These shortcomings are particularly prominent in large-scale point-cloud classification. This paper proposes a novel network structure, LI-Net, that considers geographic location and feature fusion. First, location-enhanced kernel point convolution (LFKPConv) was designed. Subsequently, a new mask attention module was introduced to integrate the global features. Compared to traditional attention mechanisms, this module focuses only on a small number of prominent feature points, allowing for effective feature interactions while reducing memory consumption. Finally, to enable cross-layer feature fusion, a feature-weighting unit was designed in the encoding phase to enhance the significance of the semantic features. The proposed method achieved competitive results on the ISPRS, LASDU, and DFC2019 datasets, as well as a new state-of-the-art result on the GML dataset, with an average F1 score of 72.0% and an accuracy of 97.3%.

Fuzziness is an inherent property of geographical phenomena and the processes of data acquisition, processing, and analysis often introduce uncertainty. Existing methods predominantly use fuzzy set (FS) theory to capture the fuzziness of geographical phenomena as fuzzy spatial objects. However, this approach has a conceptual confusion regarding fuzziness, uncertainty, and vagueness, and the membership degree is expressed using accurate values that ignore uncertainty. Furthermore, FS-based methods lack a vague temporal descriptor. Herein, a vague-spatiotemporal-object model based on the interval type-2 FS theory is proposed to express the vagueness of spatiotemporal objects. To verify the feasibility and superiority of the proposed method, the fuzzy and vague clustering algorithm was used to classify the vegetation cover types on Poyang Lake Plain, China. Furthermore, the classification accuracy was validated via field investigations, and its ability to identify the wet season of the area was verified via the annual vague water area changes of Poyang Lake. The results indicate that, compared with the spatial object model based on FSs, the proposed method increases the ability to measure membership error and express spatiotemporal vagueness.

With the development of airborne light detection and ranging (LiDAR) technology, it has become a common and efficient way to collect large-scale 3D spatial information. However, efficient and automatic semantic segmentation of LiDAR data, in the form of 3D point clouds, remains a persistent challenge. To address this, a dual attention neural network (DA-Net) is proposed, consisting of two different blocks, namely augmented edge representation (AER) and elevation attentive pooling (EAP). First, the AER can adaptively represent local orientation and position, thereby effectively enhancing geometric information. Second, the captured local features of centroid points are utilized to further encode discriminative features using the EAP with the learned attention scores. Finally, a location homogeneity (LH) module is devised to explore the long-range relationship in an encoder-decoder network. Benefiting from the dual attention module, geometric information hidden in unorganized point clouds can be effectively propagated. Besides, the LH forces the network to pay attention to the semantic consistency of elevated objects, which facilitates both point- and object-level point cloud semantic segmentation for scene understanding. A benchmark dataset is used to assess the proposed method, which achieves an overall accuracy of 85.98% and an average F1 score of 72.31%. In addition, comparisons with other latest deep learning methods on the 2019 Data Fusion Contest dataset further demonstrate the robustness and generalization ability of the proposed method.

Map construction algorithms attempt to derive a spatial graph representing a road network from GPS-sampled movement trajectories. Existing methods commonly use trajectories without considering the specific sampling methodology. Hence, the movement information is not preserved in the map construction results. The proposed map-construction method considers the particularities of the sampling process and how they affect the trajectory data to improve the overall result quality. Specifically, our proposed algorithm constructs nodes by clustering turn points. We use an adaptive clustering approach that considers when a turn point was sampled in relation to the ‘true’ node location based on the trajectory geometry. As nodes are the aggregates of turn points, edges are constructed by conflating trajectories that either connect turn points or are in close proximity to inferred nodes. Experiments using trajectory datasets at different spatial scales, data complexities, and data sources in combination with several assessment methods show that the proposed movement-aware map construction method produces maps of greater accuracy than those from the existing approaches.

Recently, collaborative representation for hyperspectral image analysis has received great interest. Due to the effectiveness of local manifold in a tangent space, this letter extends the collaborative representation classification (CRC) mechanism into the tangent space. Specifically, this letter uses simplified tangent distance and a new regularization term and designs a modified classifier innovatively. Moreover, two variants with weighted diagonal matrices to adaptively adjust the regularization terms are developed to further improve the classification performance. In the experiments, two real hyperspectral images were adopted for performance evaluation, and the experimental results demonstrate that the proposed algorithms can significantly improve classification results compared with the original CRC algorithm and other related classifiers.

Band clustering is applied to dimensionality reduction of hyperspectral imagery. Different from unsupervised clustering using all the pixels or supervised clustering requiring labeled pixels, the proposed semisupervised band clustering needs class spectral signatures only. After clustering, a cluster selection step is applied to select clusters to be used in the following data analysis. Initial conditions and distance metrics are also investigated to improve the clustering performance. The experimental results show that the proposed algorithm can outperform other existing methods with lower computational cost.

Band selection is often applied to reduce the dimensionality of hyperspectral imagery. When the desired object information is known, it can be achieved by finding the bands that contain the most object information. It is expected that these bands can provide an overall satisfactory detection and classification performance. In this letter, we propose a new supervised band-selection algorithm that uses the known class signatures only without examining the original bands or the need of class training samples. Thus, it can complete the task much faster than traditional methods that test bands or band combinations. The experimental result shows that our approach can generally yield better results than other popular supervised band-selection methods in the literature.