Estimating complex 3D topographic surface changes including rigid spatial movement and non-rigid morphological deformation is an essential task to investigate Earth surface dynamics. However, for current 3D point comparison approaches, it is challenging to separate rigid and non-rigid topographic surface changes from multitemporal 3D point clouds. Additionally, these methods are affected by challenges including topographic surface roughness and point cloud heterogeneities (i.e., discrete and irregular point distributions). To address these challenges, in this paper, we consider the dynamic evolution of topographic surfaces as the geometric changes of Riemann manifold surfaces. By building Euclidean (straight) and non-Euclidean (curved) coordinate systems on Riemann manifold surfaces that are represented from point clouds, the rigid transformation and non-rigid deformation of the Riemann manifold surfaces are solved to conceptualize rigid and non-rigid change tensors, respectively. On this basis, we design rigid (i.e., translation and rotation) and non-rigid (i.e., stretch and distortion) change features to describe various topographic surface changes and quantify the associated uncertainties to capture significant changes. The proposed method is tested on pairwise point clouds with simulated and real topographic surface changes in mountain regions. Simulation experiments demonstrate that the proposed method performed better than the baseline (i.e., M3C2) and state-of-the-art methods (i.e., LOG), with a higher translation accuracy (more than 50% improvement), a lower translation uncertainty (more than 61% reduction), and strong robustness to varying point densities. These results also show that the proposed method accurately quantifies three additional types of change features (i.e., the mean accuracies of rotation, stretch, and distortion are 1.5 degrees, 0.5%, and 3.5 degrees, respectively). Moreover, the real-scene experimental results demonstrate the effectiveness and superiority of the proposed method in estimating various topographic changes in real environments, the applicability in analyzing geomorphological processes, and the potential contribution for understanding spatiotemporal patterns of Earth surface dynamics.

3D object detection, a pivotal task in autonomous driving systems, confronts the challenge of performance degradation under occlusion and adverse weather conditions. Fusing multi-modal point clouds from LiDAR and 4D radar is an effective way to solve this problem. However, current LiDAR and 4D radar fusion methods fuse the features of heterogeneous point clouds at the bird’s-eye-view (BEV) level and ignore geometric inconsistencies, which makes them sensitive to adverse conditions. To address the issues above, we propose a LiDAR and 4D radar fusion model (FusionBev) for accurate and robust 3D object detection. In this study, we focus on how to make the network fully fuse LiDAR and 4D radar data at the voxel level and ensure geometric consistency. Further, we propose a cross-fusion module (CF) to aggregate the features of LiDAR and 4D radar voxels. After voxel encodings by CF module, we design a redundant down-sampling strategy (RD) to learn the multi-scale features. Finally, a geometry-consistent module (GC) is designed to solve the problem of geometric offset between sensors. We conduct extensive experiments across multiple public datasets to evaluate the effectiveness and robustness of our model. Notably, FusionBev achieves 89.2 % mAP on the VoD dataset and 64.9 % mAP on the K-Radar dataset. Compared to the recent LiDAR-4D radar fusion method (L4DR), we achieve more than twice the inference speed (27.6 FPS> 13.1 FPS) with less than half the GPU memory (2.81 GB < 6.31 GB).

For the development of accurate shallow landslide (translational debris and earth slides with a depth < 2 m) susceptibility assessments and further hazard or risk analyses, it is essential that complete and accurate landslide inventory data is available. Various methods are applied for the construction of shallow landslide inventories. However, it is known that the most used methods underreport landslides in forests, e.g. with visual interpretation of satellite/aerial imagery and manual mapping of landslides during field visits. To address this issue, several studies have instead used topographic Light Detection and Ranging (LiDAR) data to create their landslide inventories. These studies showed that landslides under forest cover can be mapped using topographic LiDAR, as LiDAR can penetrate the vegetation cover. The methods used in these studies can be divided into (1) methods using raster data derived from filtered LiDAR point-cloud data and (2) methods working directly on point-cloud datasets. The benefit of the raster-based methods is their computational speed and scalability, while point-cloud based methods are difficult to apply to larger areas, due to their high computational requirements, but have a greater measurement accuracy (e.g., landslide depth). This difference in accuracy is especially important for the mapping of shallow landslides, which often leave only limited traces in the landscape. This study investigates how both methods can be combined to derive a semi-automatic workflow for mapping shallow landslides using LiDAR data that is accurate and scalable. The investigation focusses on mapping shallow landslides under forest, and on how the derived workflow for mapping landslides needs to be adapted to forested and non-forested areas. In a first step, potential landslide-prone areas are identified using the difference of pre- and post-event digital terrain models, an after-event digital terrain model and their related topographic derivatives such as the roughness coefficient and slope. In the next step, the identified areas are segmented and man-made topographic changes are removed, before they are further analyzed with a more accurate mapping technique using point-cloud data from the multiscale model-to-model cloud comparison (M3C2) algorithm. In addition to the M3C2 distances, the point-cloud based mapping will also make use of 3D shape features describing point location and orientation to increase the accuracy and robustness of the topographic change detection and estimation. The scalability of the workflow is tested by applying the workflow to several areas in the Tyrolean Alps (Austria). First results, derived with a logistic regression model using the raster-based derivatives, show a distinct difference in the feature importance of the topographic derivatives when forested and non-forested areas are compared. In addition, the performance of the model also greatly benefits from a separate training in forested and non-forested areas, with an increase in the Area Under the Curve (AUC) value from 0.84 to 0.89 for, respectively, unseparated and separated training.

The availabilities of 3D multisource sensors and neural networks make it possible to improve spatial observation and data processing capacities for urban environment understanding, via multisource point cloud fusing and multiple-task associated learning. However, it is still unexplored that how to effectively fuse the multisource point clouds and extract multiple-level scene clues. This paper proposes a solution to multiple-level urban environment understanding, with the improvements of point cloud observation and processing capacities. Specifically, to increase the geometric completeness and physical variety of point clouds, we fuse respectively the geometric coordinates and physical properties of multisource point clouds, and then represent high-dimensional geometric and physical features from the fused point clouds. To increase the processing capacity especially for multisource aggregated point clouds, a multiple-task (semantic classification, instance segmentation, and relationship reasoning) interactive learning network is designed to comprehensively extract semantic, instance, and relationship information. The proposed method is tested on the multisource point clouds (i.e., laser-scanning, RGB-color, and multispectral) collected from UAV in an urban community scene. The experimental results show that the aggregation of multisource point clouds achieved significant accuracy improvements (+10 % IoU in semantic classification, +5 % AP_25 % in instance segmentation, and +51 % OA in relationship reasoning), and the interaction of semantic classification, instance segmentation, and relationship reasoning outperformed the traditional methods without interactive learning in both accuracies and levels of urban environment understanding. Moreover, we found that high-dimensional geometric and physical features are consistently effective for urban environment understanding, and their combination can accumulate the effectiveness.

3D topographic point cloud change estimation produces fundamental inputs for understanding Earth surface process dynamics. In general, change estimation aims at detecting the largest possible number of points with significance ( i.e., difference > uncertainty) and quantifying multiple types of topographic changes. However, several complex factors, including the inhomogeneous nature of point cloud data, the high uncertainty in positional changes, and the different types of quantifying difference, pose challenges for the reliable detection and quantification of 3D topographic changes. To address these limitations, the paper proposes a graph comparison- based method to estimate 3D topographic change from point clouds. First, a graph with both location and orientation representation is designed to aggregate local neighbors of topographic point clouds against the disordered and unstructured data nature. Second, the corresponding graphs between two topographic point clouds are identified and compared to quantify the differences and associated uncertainties in both location and orientation features. Particularly, the proposed method unites the significant changes derived from both features ( i.e., location and orientation) and captures the location difference ( i.e., distance) and the orientation difference ( i.e., rotation) for each point with significant change. We tested the proposed method in a mountain region (Sellrain, Tyrol, Austria) covered by three airborne laser scanning point cloud pairs with different point densities and complex topographic changes at intervals of four, six, and ten years. Our method detected significant changes in 91.39 %- 93.03 % of the study area, while a state-of-the-art method ( i.e., Multiscale Model-to-Model Cloud Comparison, M3C2) identified 36.81 %- 47.41 % significant changes for the same area. Especially for unchanged building roofs, our method measured lower change magnitudes than M3C2. Looking at the case of shallow landslides, our method identified 84 out of a total of 88 reference landslides by analysing change in distance or rotation. Therefore, our method not only detects a large number of significant changes but also quantifies two types of topographic changes ( i.e., distance and rotation), and is more robust against registration errors. It shows large potential for estimation and interpretation of topographic changes in natural environments.

Real 3D building models have become indispensable data sources for building spatial information bases for smart cities by leveraging structural correlations and rich semantic expressions of real-world scene entities. The essential prerequisite for real 3D reconstruction is real-time and dynamic detailed-level observations. Lowaltitude multicopter UAV platforms are optimal for automatic and periodic building scene observations. However, there are still several challenges in UAV-based path planning for real 3D data capture while maintaining the overall fidelity of architectural details due to observational scale variations, surrounding uncertainties, structural complexity, and topological delicacy. We propose a scene information guided aerial photogrammetric mission recomposition method in response to this challenge. Depending on the architectural complexity, the two proposed observation patterns, parallel inspection and surface enveloping, can be recomposed to achieve UAV obstacle avoidance and complete coverage of individual buildings in a restricted space, capturing global surface detail with millimeter resolution and low texture distortion. The virtual simulation environment, which is constructed based on the semantics and elevation values of the surroundings, provides a basis for selecting the observation pattern and optimal flight parameters based on the reconstruction requirements of the building. In order to achieve quality control of 3D reconstruction models, this paper introduces a reconstruction quality assessment scheme consisting of four quantitative evaluation metrics, namely coverage, resolution distribution, texture distortion score, and geometric accuracy, which effectively establishes a close relationship between mission planning and 3D reconstruction. The observation capability of the proposed method is better than other typical observation patterns, obtaining a model of globally homogeneous resolution distribution over the main body of the building, reaching an average level of 7.01 mm and the highest level of 2.12 mm (fa & ccedil;ade region), which can provide high-quality data for the semantic extraction and instantiation of multiple surface elements of buildings.

The correspondence between BIM and construction instances is crucial to construction management. However, spatial deviation and geometric heterogeneity between BIM and construction point clouds pose great challenges. This paper establishes high-dimensional point cloud feature tensor to devise a point cloud semantic segmentation network and an incremental point-to-point correspondence estimation strategy against spatial deviation and geometric heterogeneity. In the construction of the stadium for the 31st Summer World University Games, the method achieved semantic segmentation accuracy (OA) of 93.8% - 99.9% and reduced BIM-and-construction correspondence error from 16 cm to 3 cm, and automatically documented four-phase construction progresses of 38 317 instances with progress monitoring error of 1%. These results demonstrate the effectiveness and applicability of the proposed method in BIM-to-construction semantic transfer and BIM-and-construction instance correspondence for large complex buildings. Future research will design transfer learning networks to achieve fully BIM-driven semantic understanding and knowledge mining for intelligent construction.

The increasing availability of point cloud acquisition techniques makes it possible to significantly increase 3D observation capacity by the registration of multi-sensor, multi-platform, and multi-temporal point clouds. However, there are geometric heterogeneities (point density variations and point distribution differences), small overlaps (30 % similar to 50 %), and large data amounts (a few millions) among these large-scale heterogeneous point clouds, which pose great challenges for effective and efficient registration. In this paper, considering the structural representation capacity of graph model, we propose an incremental registration method for large-scale heterogeneous point clouds by hierarchical graph matching. More specifically, we first construct a novel graph model to discriminatively and robustly represent heterogeneous point clouds. In addition to conventional nodes and edges, our graph model particularly designs discriminative and robust feature descriptors for local node description and captures spatial relationships from both locations and orientations for global edge description. We further devise a matching strategy to accurately estimate node matches for our graph models with partial even small overlaps. This effectiveness benefits from the comprehensiveness of node and edge dissimilarities and the constraint of geometric consistency in the optimization objective. On this basis, we design a coarse-to-fine registration framework for effective and efficient point cloud registration. In this incremental framework, graph matching is hierarchically utilized to achieve sparse-to-dense point matching by global extraction and local propagation, which provides dense correspondences for robust coarse registration and predicts overlap ratio for accurate fine registration, and also avoids huge computation costs for large-scale point clouds. Extensive experiments on one benchmark and three changing self-built datasets with large scales, outliers, changing densities, and small overlaps show the excellent transformation and correspondence accuracies of our registration method for large-scale heterogeneous point clouds. Compared to the state-of-the-art methods (i.e., TrimICP, CoBigICP, GROR, VPFBR, DPCR, and PRR), our registration method performs approximate even higher efficiency while achieves an improvement of 33 % - 88 % regarding registration accuracy (OE).

Accurate indoor 3D models are essential for building administration and applications in digital city construction and operation. Developing an automatic and accurate method to reconstruct an indoor model with semantics is a challenge in complex indoor environments. Our method focuses on the permanent structure based on a weak Manhattan world assumption, and we propose a pipeline to reconstruct indoor models. First, the proposed method extracts boundary primitives from semantic point clouds, such as floors, walls, ceilings, windows, and doors. The primitives of the building boundary are aligned to generate the boundaries of the indoor scene, which contains the structure of the horizontal plane and height change in the vertical direction. Then, an optimization algorithm is applied to optimize the geometric relationships among all features based on their categories after the classification process. The heights of feature points are captured and optimized according to their neighborhoods. Finally, a 3D wireframe model of the indoor scene is reconstructed based on the 3D feature information. Experiments on three different datasets demonstrate that the proposed method can be used to effectively reconstruct 3D wireframe models of indoor scenes with high accuracy.

The forest canopy height is a key indicator for measuring global forest carbon stocks. Spaceborne LiDAR, a satellite remote sensing technology, plays an essential role in large-scale canopy height estimations. However, there are still some problems with existing methods of the spaceborne LiDAR canopy height estimates: the retrieval accuracy is degraded by the topographic relief and vegetation cover, as well as uneven spatial distribution of mapping height uncertainties. In this paper, we investigated the possibility of fusing multimodal spaceborne LiDAR and optical images to improve these above problems. We proposed a hybrid model fusing spaceborne full-waveform and photon-counting LiDAR data with optical imagery. Specifically, our approach divided the regional extent into multiple fusion patterns based on the spatial distribution of the LiDAR footprints in an object-oriented method. We then constructed canopy height models corresponding to each pattern and finally integrated the model results using a weighting scheme considering geospatial distances. We used GEDI (full-waveform LiDAR), ICESat-2 (photon-counting LiDAR) and Sentinel-2 (optical imagery) products as the input data and validated the model accuracy in four representative biomes of global forest ecosystems (i.e., evergreen broadleaf forests, deciduous broadleaf forests, savannas and coniferous forests). The experimental results demonstrated that fusing multisource spaceborne LiDAR data and optical images can not only enhance the canopy height estimation accuracy (R2 0.65 ∼ 0.90 and RMSE 0.57 ∼ 4.15 m in four biomes) but also maintain stable accuracy under undulating slope and large vegetation cover. Moreover, the uncertainty of canopy height estimation was low (meanerror −0.20 ∼ 0.03 m) and uniformly distributed in space (stdev 0.71 ∼ 4.45 m). We also compared the performances with two other advanced canopy height models, as well as two global canopy height products, and our model showed significant advantages in each test region. Our study demonstrates the effectiveness of fusing multimodal spaceborne LiDAR data and optical imagery for canopy height estimation accuracy improvement.

With the rapid development of sensor technology and observation platform, point cloud data that is viewed as primary data of remote sensing, has gradually become an important information carrier. Moreover, it plays an increasingly significant role in the national major strategic needs such as geological disaster situation awareness, natural resources quantitative investigation and road traffic safety services. At the same time, driven by point cloud observation equipment and national major strategic needs, spatial scenes have changed from perception to cognition, and new requirements for cognitive processing algorithms and computing power have also been put forward. Therefore, based on the basic framework of point cloud scene cognition, this paper analyzes the research status of multi-source point cloud coupled observation, summarizes the key progress of point cloud scene cognition and typical applications in major national strategic needs, and summarizes the main problems facing point cloud scene cognition at present. On this basis, this paper focuses on the cutting-edge challenges of cloud scene cognition, avoids the traditional Euclidean space and turns to the high-dimensional tensor manifold space for point cloud data processing, proposes the scientific concept and technical framework of generalized point cloud, and provides a new research idea for the algorithm and computing power of cognitive processing of point cloud scene.

Building Information Modeling (BIM) has increasingly been adopted as an as-planned construction state to provide essential support for fine construction. However, some deviations are inevitably induced between the BIM model and actual construction states in practical construction processes, and these deviations obstruct the unidirectional interaction between the BIM model and actual construction states, thus seriously degrading the fine construction quality. This paper proposes a bidirectional interaction mechanism between BIM and construction processes using a multisource geospatial data enabled point cloud model. In the forward interaction, parametric BIM model are obtained to provide essential information for fine construction. Particularly, a multisource geospatial data enabled point cloud modeling strategy is the core of the proposed method, which overcomes complex buildings characterized by severe occlusions, specular surfaces and similar components to capture accurate and complete 3D point cloud model that reflects the actual construction state. In the backward interaction, the point cloud model provides feedback to adjust the BIM model for ensuring a fit between the BIM model and actual construction state. Moreover, the resulting up-to-date BIM model can further instruct the subsequent construction processes; thus, BIM and construction become a closed task-oriented loop. The proposed method was applied to the curtain wall construction of the main stadium for the Chengdu 2022 31st Summer World University Games and compared with four other state-of-the-art point cloud registration methods. The results demonstrated the superior accuracy of our method over other methods. The results also showed that the proposed method could effectively maintain the interaction between the BIM model and actual construction states throughout the construction lifecycle and ensure the quality of fine construction.

Point cloud semantic segmentation in urban scenes plays a vital role in intelligent city modeling, autonomous driving, and urban planning. Point cloud semantic segmentation based on deep learning methods has achieved significant improvement. However, it is also challenging for accurate semantic segmentation in large scenes due to complex elements, variety of scene classes, occlusions, and noise. Besides, most methods need to split the original point cloud into multiple blocks before processing and cannot directly deal with the point clouds on a large scale. We propose a novel context-aware network (CAN) that can directly deal with large-scale point clouds. In the proposed network, a local feature aggregation module (LFAM) is designed to preserve rich geometric details in the raw point cloud and reduce the information loss during feature extraction. Then, in combination with a global context aggregation module (GCAM), capture long-range dependencies to enhance the network feature representation and suppress the noise. Finally, a context-aware upsampling module (CAUM) is embedded into the proposed network to capture the global perception from a broad perspective. The ensemble of low-level and high-level features facilitates the effectiveness and efficiency of 3-D point cloud feature refinement. Comprehensive experiments were carried out on three large-scale point cloud datasets in both outdoor and indoor environments to evaluate the performance of the proposed network. The results show that the proposed method outperformed the state-of-the-art representative semantic segmentation networks, and the overall accuracy (OA) of Tongji-3D, Semantic3D, and Stanford large-scale 3-D indoor spaces (S3DIS) is 96.01%, 95.0%, and 88.55%, respectively.

Visual localization and mapping have received considerable attention in the fields of computer vision, photogrammetry, and remote sensing. Image matching is key to visual localization and mapping. However, light is complex and uneven in real scenarios, creating difficulties in feature extraction and matching. Hence, feature mismatch and loss reduce the efficiency, accuracy, and robustness of visual localization and mapping. We developed a visual localization and mapping method for complex light scenarios based on image enhancement. Starting with initial images, the irradiance and reflectance components were separated based on logarithmic transformation. Our method strengthened high-frequency components and restrained low-frequency components with improved homomorphic filtering, restraining the light component and enhancing the important reflection component. The SIFT algorithm was used for feature detection and matching. The proposed method was tested on images with uneven light captured using a stereo vision camera in an indoor environment, focusing on visual localization and mapping. The experimental results emphasized that the method improved the rate of image localization and number of point clouds, as well as the reprojection error, which ranged from 0.85 to 0.82 on average. Thus, the proposed method is robust rather than probabilistic for improving visual localization and mapping under complex light conditions.

To meet the requirement of high-resolution and high-efficiency unmanned aerial vehicle (UAV)-borne multi-spectral remote sensing, using the miniaturized large-array commodity complementary metal-oxide semi-conductor (CMOS) camera is an effective solution. Given the characteristics of the new sensor and platform, almost no systematic and feasible radiometric calibration method has been specifically developed. In this paper, we proposed an indoor and outdoor integrated radiometric calibration method. To develop a systematic indoor calibration method, we explored the optimal methods for dark current offset, vignetting effect correction, and quantum efficiency calibration. According to the comparison results of three different methods, the lookup table (LUT) method was chosen to correct vignetting effect rather than nonlinear regression. Further, we proposed an exponential nonlinear model to replace the traditional linear model for quantum efficiency calibration, which improved the R-squares from around 0.92 to around 0.99. The outdoor calibration included atmospheric path radiance and reflectance correction. We proposed an empirical line method based on the dark target method to correct the atmospheric path radiance before the reflectance correction. Based on our method, the mean absolute percentage errors (MAPE) between the observed reflectance and the true reflectance were around 10%. More-over, the method can greatly improve the calculated reflectance accuracy of low reflectance targets. Our method can serve as a useful reference for the radiometric calibration of large-array commodity CMOS multispectral cameras. It can also contribute to the application of UAV-borne multispectral remote sensing.

Highly accurate 2D maps can supply basic geospatial information for efficient and accurate indoor building modeling. However, problematic scenarios, which are characterized by few features, similar components and large scales, seriously influence data association and cumulative error elimination, and thus degrade simultaneous localization and mapping (SLAM)-based mapping quality. In this paper, a cross-correction LiDAR SLAM method is proposed for constructing high-accuracy 2D maps of problematic scenarios. The method comprises two models. The first model, namely, pose correction for rough mapping (PCRM), increases the data association capacity and generates a rough map with cumulative errors. In the PCRM model, a rough mapping module is developed against the scenario with few features for accurate data association. This module improves the robustness of the data association by using the initial poses from the local pose correction module, especially in similar-component scenarios. The other is a map correction for pose optimization (MCPO) model, which enhances cumulative error elimination capacity. Here, a block-based local map correction module is proposed that takes both map and pose into consideration to construct accurate constraints. The constraints are then added to the global pose optimization module to significantly reduce the cumulative error of the rough map and thus construct a high-accuracy 2D map. The results demonstrate the superiority of our method over 5 other state-of-the-art methods in problematic scenarios. The overall performance of our method in these two scenarios is approximately 1 cm and 0.2% in terms of the absolute and relative map errors, respectively. Moreover, the modeling results demonstrate that our method can be applied to the efficient and accurate indoor modeling.

Urban land cover classification for high-resolution images is a fundamental yet challenging task in remote sensing image analysis. Recently, deep learning techniques have achieved outstanding performance in high-resolution image classification, especially the methods based on deep convolutional neural networks (DCNNs). However, the traditional CNNs using convolution operations with local receptive fields are not sufficient to model global contextual relations between objects. In addition, multiscale objects and the relatively small sample size in remote sensing have also limited classification accuracy. In this paper, a relation-enhanced multiscale convolutional network (REMSNet) method is proposed to overcome these weaknesses. A dense connectivity pattern and parallel multi-kernel convolution are combined to build a lightweight and varied receptive field sizes model. Then, the spatial relation-enhanced block and the channel relation-enhanced block are introduced into the network. They can adaptively learn global contextual relations between any two positions or feature maps to enhance feature representations. Moreover, we design a parallel multi-kernel deconvolution module and spatial path to further aggregate different scales information. The proposed network is used for urban land cover classification against two datasets: the ISPRS 2D semantic labelling contest of Vaihingen and an area of Shanghai of about 143 km2. The results demonstrate that the proposed method can effectively capture long-range dependencies and improve the accuracy of land cover classification. Our model obtains an overall accuracy (OA) of 90.46% and a mean intersection-over-union (mIoU) of 0.8073 for Vaihingen and an OA of 88.55% and a mIoU of 0.7394 for Shanghai.

Rapid urbanization has become a major urban sustainability concern due to environmental impacts, such as the development of urban heat island (UHI) and the reduction of urban security states. To date, most research on urban sustainability development has focused on dynamic change monitoring or UHI state characterization, while there is little literature on UHI change analysis. In addition, there has been little research on the impact of land use and land cover changes (LULCCs) on UHI, especially simulates future trends of LULCCs, UHI change, and dynamic relationship of LULCCs and UHI. The purpose of this research is to design a remote sensing-based framework that investigates and analyzes how the LULCCs in the process of urbanization affected thermal environment. In order to assess and predict the impact of LULCCs on urban heat environment, multitemporal remotely sensed data from 1986 to 2016 were selected as source data, and Geographic Information System (GIS) methods such as the CA-Markov model were employed to construct the proposed framework. The results showed that (1) there has been a substantial strength of urban expansion during the 40-year study period, (2) the farthest distance urban center of gravity moves from north-northeast (NEE) to west-southwest (WSW) direction, (3) the dominate temperature was middle level, sub-high level, and high level in the research area, (4) there was a higher changing frequency and range from east to west, and (5) there was a significant negative correlation between land surface temperature and vegetation and significant positive correlation between temperature and human settlement.

Image matching forms an essential means of data association for computer vision, photogrammetry and remote sensing. The quality of image matching is heavily dependent on image details and naturalness. However, complex illuminations, denoting extreme and changing illuminations, are inevitable in real scenarios, and seriously deteriorate image matching performance due to their significant influence on the image naturalness and details. In this paper, a spatial-frequency domain associated image-optimization method, comprising two main models, is specially designed for improving image matching with complex illuminations. First, an adaptive luminance equalization is implemented in the spatial domain to reduce radiometric variations, instead of removing all illumination components. Second, a frequency domain analysis-based feature-enhancement model is proposed to enhance image features while preserving image naturalness and restraining over-enhancement. The proposed method associates the advantages of the spatial and frequency domain analyses to complete illumination equalization, feature enhancement and naturalness preservation, and thus acquiring the optimized images that are robust to the complex illuminations. More importantly, our method is generic and can be embedded in most image-matching schemes to improve image matching. The proposed method was evaluated on two different datasets and compared with four other state-of-the-art methods. The experimental results indicate that the proposed method outperforms other methods under complex illuminations, in both matching performances and practical applications such as structure from motion and multi-view stereo.