Applied Geomatics最新文献

Reliable estimation of Tree Canopy Height (TCH) is pivotal for monitoring forest stand dynamics and carbon flux in forest ecosystems. Conventional methods like field surveys, Terrestrial LiDAR, and Airborne LiDAR provide precise measurements but are costly and impractical for large-scale applications. This study integrates Global Ecosystem Dynamics Investigation (GEDI) LiDAR footprints with the Shuttle Radar Topography Mission (SRTM) Digital Surface Model (DSM) to examine changes in TCH within planted mangroves along Bangladesh’s southwestern coast. The SRTM DSM was calibrated to a TCH model for 2000 using a published SRTM-based canopy height model (CHM). Interpolation techniques such as Natural Neighbour (NN), Ordinary Kriging (OK), and Inverse Distance Weighting (IDW) were employed to construct a comprehensive TCH model (TCHM) for 2020 from GEDI LiDAR footprints. TCHM was derived by subtracting the Digital Terrain Model (DTM) from the DSM, excluding urban vertical structures using the ESA World Cover V200 dataset. Results indicate a 19% overall increase in TCH between 2000 and 2020, with increments ranging from 2.5 to 20 m, indicating significant accumulation in aboveground biomass (AGB) and suggesting the area likely acted as a carbon sink. Validation revealed strong correlations between observed and GEDI-derived TCH values, with an R² of 0.81 and an RMSE of 2.38 m (p < 0.001), confirming GEDI’s effectiveness for TCH estimation. However, challenges such as GEDI data coverage gaps in the study region and SRTM’s limited canopy penetration persist. The GEDI coverage gaps and interpolation uncertainties will incur canopy height accuracy. Findings of this study provide insights into forest ecosystem dynamics and carbon cycling, highlighting the role of GEDI in monitoring terrestrial ecosystems.

Personalized recommendation in location-based services (LBS) is a challenging task due to the diversity of user preferences and the dynamic nature of location-based interactions. Existing methods often struggle to capture the complex interplay of user interests, leading to suboptimal recommendations. Addressing this problem is crucial for enhancing user experience and driving engagement in LBSs. Accurate and personalized point of interest (POI) recommendations can help users discover relevant places, optimize their time, and make informed decisions, because these points can play the role of the road landmarks. By improving POI recommendations, we can foster a more enjoyable and valuable experience for LBSN users. This paper presents a novel approach to enhancing LBS by leveraging deep learning (DL) techniques and data from location-based social networks (LBSNs) like Foursquare and Yelp. Our method effectively extracts user interests from LBSN data, incorporating individual preferences, social influence, and temporal dynamics. By combining these factors, we provide personalized POI recommendations that significantly improve user experience. We introduce an improved long short-term memory (LSTM) model to capture complex user behavior patterns and extract relevant POIs. Additionally, we address the challenge of data inconsistency across different social networks by harmonizing place categories. In the proposed model, the improved artificial bee colony (IABC) optimally updates the weights, biases, learning rate, and the number of units of the LSTM. This paper also introduces a new three-dimensional migration model to enhance the exploitation capabilities of the ABC algorithm. The performance of the proposed IABC-LSTM model is compared to several advanced machine learning (ML) algorithms, including recurrent neural networks (RNN), K-nearest neighbors (KNN), support vector machines (SVM), biogeography-based optimization (BBO), and orchard algorithm (OA). To evaluate our model, we utilized a rigorous evaluation framework, incorporating metrics such as root mean square error (RMSE), coefficient of determination (R²), convergence curve, precision, recall, F-score, and receiver operating characteristic (ROC) curve. Our research demonstrates the effectiveness of our approach by achieving the highest (:{R}^{2}) value of 96.41% on the validation set, and the lowest RMSE value of 0.84, whereas SVM model achieved the highest RMSE and the lowest(::{R}^{2}). Also, the combination of all three factors achieved the highest F-score of 0.893. Our findings highlight the importance of considering multiple factors for personalized recommendations and the potential of DL techniques to revolutionize LBS.

Soil salinity and sodicity are significant environmental concerns contributing to land degradation, hampering plant growth, disrupting microbial functions, reducing crop yields, and limiting arable land for agriculture. Understanding these soil properties from spatiotemporal perspectives is essential for decision-makers to devise effective crop management and agricultural planning strategies. This study aims to develop an approach integrating multi-source remote sensing data with an emsenble random forest (ERF) for soil salinity and sodicity assessment in Taiwan. The data spanning from 2012 to 2020 were processed, and predicted electrical conducitivty (EC) and sodium adsorption ratio (SAR) results were validated against field observations, demonstrating a robust agreement with correlation coefficient (r) values of 0.78 and 0.79, respectively. Although these models effectively captured the spatial patterns of soil salinity and sodicity, the root mean square error (RMSE) values of 1.04 and 1.4 for EC and SAR models indicated a room for further improving predictive accuracy and consistency. Salinity and sodicity classes derived from predicted results showed that slight salinity was widely distributed across the study region, whereas moderate to strong salinity and slight to moderate sodicity were more concentrated in low-lying coastal areas, highlighting the significant impact of saltwater intrusion and poor drainage. Finally, these salinity and sodicity classes were overlaid with township-level cultivated areas for targeted management interventions and improved agricultural practices.

The dynamic development of railway infrastructure requires a proactive approach of surveying work and Structural Health Monitoring (SHM). A proactive approach makes it possible not only to anticipate and identify future technology needs, but also to evolve and select the most optimal solutions, relevant to infrastructure maintenance. The main objective of the work is to analyze and evaluate the staking accuracy of special multi-track constructions (SMtC) represented in this work by six single turnouts. Modern Robotic Total Stations (RTS) were used, which work with mobile devices (MD). Three models of RTS were verified: the Trimble S7, Trimble S9 HP and Geomax Zoom90. Attention was paid to the location of the RTS station in use with the polar method. It has been shown that, the highest staking accuracies are presented by the Trimble S9 HP total station, due to its high accuracy and optical plummet. The most optimal location of the RTS station for staking out turned out to be a staking stations near the central construction zone for each turnout – near the mathematical point. Such a location of the RTS station allowed for an even distribution of errors on the staked points. The accuracy parameters of RTS imply their use in SHM of railway infrastructure. The results of the implemented work have confirmed that modern RTS working with MD, provide high accuracy in surveying work.

Forest fires are the main cause of forest degradation in Algeria, with serious consequences for the environment and the local economy. The variety of traditional models and methods used for fire prevention and suppression poses challenges in terms of specialization, precision, updating, and validation. These techniques and procedures are often slow and less reliable due to the complexity and diversity of forest ecosystems. Research in this field has demonstrated the effectiveness and speed of using remote sensing and Geographic Information Systems (GIS) to map the vulnerability of forest fires in the eastern Aurès region in Algeria, where the Turkish model (Erten) was applied in this study. This model integrates five key factors essential for assessing forest fire vulnerabilities, which include slope, aspect, the vegetation, proximity to roads, and proximity to human settlements. The results revealed that the region faces vulnerabilities ranging from low to very high. Where 691.71 km² (41.39%) of the surface area is in a high to very high vulnerability zone. The validity of these results was verified in two different ways. Initially, field surveys were carried out using reference data to pinpoint the regions of fire outbreaks in 2022 and compare the findings with the results obtained. In this study, it was discovered that 33 of the 47 fire points, roughly 70.2%, are situated in regions that are categorized as high and very high vulnerability for forest fires, while the outcomes were good. The second method was statistical validation based on the ROC analysis, which yielded an AUC value of 0.747, indicating good predictive performance. Additional accuracy metrics (overall accuracy = 64%, precision = 0.61, recall = 0.70) confirmed the model’s reliability in distinguishing between fire and non-fire areas. The results obtained are promising and contribute to improved forest fire prevention and informed decision-making.

The rapid increase in domestic and agricultural water demand in the Municipality of Gjilan has intensified the need for new, sustainable water resources. This study aims to identify the most suitable location for the construction of an accumulation dam in the “C.Z. Pograxhë” catchment by applying a GIS-based multi-criteria decision analysis framework grounded in the Analytic Hierarchy Process (AHP). Eight spatial and non-spatial criteria were considered: terrain slope, geology, catchment size, precipitation, distance from roads, distance from settlements and protected areas, distance from rivers, and land ownership. The criteria were weighted through pairwise comparisons supported by an expert panel, and integrated in a weighted overlay analysis to produce a suitability map for dam site selection. The results indicate that a substantial proportion of the study area exhibits moderate to high suitability for dam construction, with the most suitable zone located along the Llapushnik River corridor. Two design scenarios were evaluated, corresponding to planned dam heights of 32 m and 42 m. The estimated total storage capacities are 1.69 × 10⁶m³ and 3.29 × 10⁶m³, respectively, which would satisfy projected municipal water supply requirements for approximately 10 and 20 years. A validation and sensitivity analysis demonstrates that the spatial pattern of highly suitable areas is robust to reasonable variations (± 20%) in the factor weights, and that locations identified as highly suitable are consistent with expert judgement and existing planning documents. The study confirms the effectiveness of integrating AHP with GIS-based cartographic modelling for dam site selection in data-limited environments. Beyond providing a concrete candidate site for an accumulation dam in the “C.Z. Pograxhë” catchment, the proposed framework can support planners, water utilities, and policymakers in screening and prioritizing alternative locations for new water resources at municipal and regional scales.

This paper presents a system for sidewalk pothole measurement that applies structure-from-motion (SfM) to citizen-reported images and using raw time-difference, carrier-phase GNSS observations from smartphones to resolve the scale ambiguity in SfM reconstructions. The methodology transforms arbitrary SfM camera locations into actual locations using a closed-form absolute orientation technique. A weighting system is proposed to improve scale estimation by comparing SfM and smartphone data vector alignments. After that, pothole quantities, including perimeter and maximum depth, are extracted. RTK is used to collect reference data to validate camera locations. Estimated pothole quantities deviate from the reference by 35 cm for perimeter and 1 cm for maximum depth, demonstrating the method’s effectiveness. This approach enhances the accuracy of citizen-reported pothole information, supporting local authorities in infrastructure management and public safety improvements.

Glacial Lake Outburst Floods (GLOFs) pose a growing threat to Himalayan communities. Yet, integrated frameworks that link systematic regional-scale susceptibility assessments with quantitative, uncertainty-aware impact modelling are critically lacking. This study presents an integrated approach that combines multi-decadal remote sensing, a Fuzzy-Analytic Hierarchy Process (AHP), and ensemble hydrodynamic modelling to assess GLOF risk in the Garhwal region, India. Our inventory (1990–2023) reveals a 126% increase in total glacial lake area, expanding from 147.3 ± 0.32 ha to 332.9 ± 0.11 ha, with 21 lakes exhibiting expansion > 100%. The Fuzzy-AHP susceptibility model, which uniquely incorporates decadal lake expansion as a dynamic criterion and propagates measurement uncertainties, identified 11 out of 87 lakes as highly susceptible, characterized by large volumes (e.g., 4.96 ± 0.89 × 10⁶ m³), rapid growth, and unstable moraines in four different river basins, except for the Yamuna. Hydrodynamic (HEC-RAS) modelling of worst-case breach scenarios for the highest-priority lakes projected rapid breach development (< 0.6 h), peak discharges ranging from 377 to 1,149 m³/s, and flood depths up to 14.3 m in downstream settlements, with significant spatial variability across basins. The synthesis of dynamic susceptibility ranking and probabilistic flood impact ranges provides a transformative, actionable basis for risk reduction. We recommend immediate early warning systems at the four highest-risk lakes (e.g., GL-AL47) and dynamic hazard zoning for settlements like Badrinath and Ghansali. This framework advances GLOF risk assessment beyond static indices toward proactive, evidence-based decision-making.

Some of the sustainable development goals (SDG’s) include attainment of good health and well-being (SDG 3) and the access to safe water and sanitation (SDG 6) by all global human communities by 2030. Although some progress has been made towards realizing these goals, degrading water quality and humans continue to be a problem in low to middle-income countries. This is especially true in informal settlements where provision of both health and water services are limited due to highly contaminated surface water bodies. The surface water bodies in these settlements are often very small thus posing a challenge to the analysis through medium to coarse resolution satellite data. Thus, this study aimed at exploring the capability of high-resolution satellite data for mapping biophysical and chemical parameters for water quality assessment in the informal settlements of South Africa. The four-band PlanetScope (PS) data with 3 m spatial resolution and field data were used to map the potential water quality in the selected informal settlements. The correlation analysis (CA), and stepwise multiple linear (including logistic) regression analyses were used to model concentrations of pH, turbidity, total suspended solids (TSS) Escherichia coli (E. coli), total dissolved solids (TDS), total coliforms, salmonella, heavy metals, and chlorophyll-a (chl-a) amongst other indicators. The results have shown that turbidity was the most spectrally correlated variable (r = 0.61) to blue wavelength (0.485 𝜇m) of PS data, while the lowest spectral correlation (r = 0.01) was found between salmonella presence/absence with both blue and red (0.630 𝜇m) bands. The higher concentrations of E. coli were spatially models for Alexandra informal settlement compared to other settlements. The model developed for predicting salmonella yielded the highest correlation (R² = 0.86) while pH model (R² = 0.43) did poorly in predicting the distribution of pH in the informal settlements’ water bodies. The models developed for predicting heavy metals (Fe, Zn, Cu, and Cd) exhibited higher correlation (R² > 0.70) except for the Pb model (R² = 0.41).These results offer a capability promise of high spatial resolution PS data for mapping the biophysical and biochemical parameters of surface water bodies, although in general the correlations obtained were not very strong due to the optically inactive nature of most of the water quality parameters considered in this study.

The rapid and accurate three-dimensional (3D) data acquisition of urban fabric, including building information, can be achieved using airborne laser scanning (ALS), a cutting-edge 3D mapping technique. This paper presents a Deep Residual U-Net framework (DRU-Net3D) for urban building identification using 3D ALS point cloud data with three well-designed steps, namely point-based feature descriptor generation, point-based feature descriptor to image conversion, and DRU-Net3D method training and testing considering different levels of scene complexity and heterogeneity in the ALS datasets. The proposed DRU-Net3D model is straightforward to implement, where accurate building segmentation is effectively achieved in several complex cases, such as connected and overlapped objects with one of the objects as building; partially acquired building geometries and data gaps; different shapes, sizes and orientation of the buildings present in the ALS datasets. The proposed framework achieved an average Intersection-over-Union (IoU) of 81.56%, an overall accuracy of 94.27%, and an average F1-score of 82.16% across the evaluated airborne LiDAR datasets, demonstrating consistent segmentation performance under varying urban scene complexities. The proposed method was compared with various state-of-the-art techniques, namely U-Net, SegNet and DeepLabv3+, and demonstrated improved performance on the evaluated datasets.