Applied Geomatics最新文献

Mapping wetland ecosystems at the species level provides critical information for understanding the nutrient cycle, carbon sequestration, retention and purification of water, waste treatment and pollution control. However, wetland ecosystems are threatened by climate variability and change and anthropogenic activities; thus, their assessment and monitoring have become critical to inform proper management interventions. Contemporary studies show that satellite-based Earth observation (EO) has significant potential for achieving this task. While many multispectral EO data are freely and readily available, its broad spectral bands limit its utility in differentiating subtle differences among similar plant species. In contrast, hyperspectral data has a high spectral resolution, which is superior in discerning minute differences in similar plant species. However, this data is associated with high dimensionality and multicollinearity, which negatively affect the performance of traditional, parametric classification algorithms. To this end, machine algorithms are often preferred to classify hyperspectral data due to their robustness to various data distributions and noise. The current study compared the performance of three advanced machine learning classifiers, i.e., Support Vector Machine (SVM), Random Forest (RF), and Partial Least Squares Discriminant Analysis (PLS-DA), in discriminating four dominant wetland plant species, i.e., Crocosmia sp., Grasses, Agapanthus sp. and Cyperus sp. using simulated hyperspectral data from an upcoming sensor, i.e., nSight-2. The results revealed that SVM is superior, with an overall accuracy of 93.18% (and class-wise accuracies > 85%). In comparison, there were minor differences in the performances of RF and PLS-DA, i.e., 84.09% and 83.63%, respectively. Overall, the results demonstrated that all the evaluated classifiers could achieve acceptable mapping accuracies. However, SVM is more robust, providing exceptional accuracies, and should be considered for operational mapping once the sensor is in space.

Recent flood events (FE) in Germany have shown that the extent and impact of extreme flood events cannot be estimated solely based on numerical models. For analyzing the development of such an event and to develop and implement safety measures more efficiently, additional data must be collected during the event. Within the scope of this research, the possibilities of near real-time recording using an unmanned aerial vehicle (UAV) and data processing with the Structure from Motion (SfM) method were tested in a case study. Different recording parameter combinations were tested in the Laufer Muehle area on the Aisch river in Germany. The focus of the investigations was the identification of a parameter combination that allows a short recording interval for aerial imagery. Based on these findings, the identification of changes in the study area by comparing multitemporal photography (flood prevention), as well as the recording of flooded areas during a FE should be possible. The accuracy analysis of the different parameter combinations between two point clouds as well as the process of change detection was done by a Multiscale Model to Model Cloud Comparison (M3C2) and including ground control points. As a result, a parameter combination was identified which led to the desired results in the study area. The processes were transformed into fully automated and scripted workflows. The results serve as a basis for establishing a workflow for near real-time analyses in future studies.

With the increasing use of drones for capturing images in urban areas, correcting for distortion and sawtooth effects on orthophotos generated with these images has become a critical issue. This is particularly challenging due to the larger displacements generated by high objects and lower flight altitude of drones compared to crewed aircraft. In addition, image-based point cloud generation methods often fail to produce complete point clouds due to occluded areas and radiometric changes between overlapping images, especially near the borders of high objects. To address these issues, a novel method is proposed in this article for improving the generated point clouds with image-based methods using a deep learning network, called urban-SnowflakeNet, which comprises the following steps: 1) preparing and normalizing the roof's point cloud; 2) completing the point clouds of the building using the proposed deep learning network; 3) restoring the completed point clouds of the buildings to the real coordinates and combining them with the background point cloud; and, 4) correcting the DSM and generating the final true orthophotos. On two different image datasets, our method reduced distortions at the building's edges by 40% on average when compared to the most recent orthophoto enhancement method. However, by maintaining this success on more datasets, the approach has the potential to improve the accuracy and completeness of point clouds in urban regions, as well as other applications such as 3D model improvement, which require further testing in future works.

Climate variability is a highly debated and unavoidable global environmental challenge that has adverse effects on Ethiopia, a developing country. Hence, the objective of this research is to examine the changes in rainfall patterns in Addis Ababa City, Ethiopia, from 1981 to 2018, considering both spatial and temporal aspects. The study utilized a time-series dataset of climate information, which had a spatial resolution of 4 × 4 km, obtained from the National Meteorological Agency of Ethiopia. Supplementary data was also acquired from the Ethiopian Space Science and Geospatial Institute. To examine the rainfall variability, statistical measures such as the coefficient of variation (CV) and standardized anomaly index (SAI) were employed. Geospatial technologies and “R” programming were also used to perform a non-parametric Mann-Kendall (MK) test and Sen’s slope estimator for the investigation of both the trend and magnitude of changes. The annual, Kiremt (main rainy), and Belg (spring) seasons rainfall exhibited low to moderate variability with CV < 20% and CV < 30%, respectively, and very high variability for the Belg season (CV > 30%). The Bega season’s variability was extreme (CV > 70%). In contrast, decadal rainfall variability was generally very low (CV < 10%). The months from October to March showed higher inter-monthly variability, with CV exceeding 100%. In contrast, the Kiremt season, July, and August, experienced lower inter-monthly variability (CV < 30%). The western, north-east, and southern parts of Addis Ababa demonstrated relatively higher rainfall variability, and the trends decreased in all seasons and months, except the Kiremt season and the months of May, June, and September. However, none of these seasonal and monthly changes were statistically significant (P > 0.05). The study identified 6 years (1982, 1984, 1997, 1999, 2014, and 2015) with varying degrees of drought. Consequently, the spatio-temporal variability of precipitation should be considered in development plans, disaster risk reduction strategies, and policy measures such as flood management.

Within the context of agricultural planning, spatial data have played a crucial role, replacing conventional tabular-based data. Plantation, one of the key agricultural commodities, has been of interest since they occupy large coverage of landmass. Primary data supplies have been provided by space agencies, allowing detailed, updated satellite data to monitor this resource, with the aid of machine learning. This article discusses the opportunity of implementing support vector machines (SVM) and relevance vector machines (RVM) for estimating tree girth as a predictor of tree maturity and plantation productivity. The current research indicated that baseline SVR models were unable to yield a sufficient outcome. The complexity of the problem suggested that only the radial basis function (RBF) kernel was promising. Tuning SVM on linear and polynomial kernels did not enhance the quality of the models, although it appeared that the phenomenon of diminishing return existed. After parameter tuning, this research yielded a model with root mean squared error (RMSE) around 8.5 cm with R² around 0.69. Although it was recently introduced, RVM with the same RBF kernel did not yield a sufficient model with RMSE about 52 cm. This concludes that the optimal model should be sought through examining a wide range of machine learning approaches.

Wildfires have become a growing global concern due to the environmental and economic damage they cause. Climate change is a primary cause of wildfires as it increases the frequency, extent, and severity of wildfires. In addition to climate change, human activities have become a major cause of wildfires, particularly in the Mediterranean region. Since wildfire is a very complicated environmental problem, effectively responding to and minimising the danger of a wildfire necessitates the integration of all available information into decision-making systems. The complexity of wildfires can have a negative impact on decision-making, particularly when decisions are temporally made under dynamic, uncertain, and contradictory conditions. Since the early 1990s, there has been a rise in the occurrence of “mega-fires” throughout Europe, which are characterised by wildfires that surpass the present firefighting capabilities. Controlling mega-fires exceeds the response capacity of the individual institutions as effective wildfire management requires extensive coordination of the institutions and all available resources at a local, regional, and national level. This cooperation necessitates the integration of advanced technologies with scientific knowledge, as well as the combination of various heterogeneous spatial and non-spatial data. GIS technology provides an efficient, expedited, and economical process of data collection and analysis. In the last decades, GIS-based decision support systems have been used to improve the efficiency of firefighting processes like planning, management, and decision-making. In this study, a conceptual framework of a GIS-based decision support system for wildfire prevention and fighting in Turkey was proposed. The presented conceptual design aims to improve the firefighting capacity by providing decision-oriented spatial information on wildfire risks and dangers timely through integrated functional tools efficiently.

The examination of changes in land use and land cover (LULC) holds a pivotal role in advancing our comprehension of underlying processes and mechanisms. The advancement of sophisticated earth observation programs has opened unprecedented opportunities to meticulously observe geographical areas, courtesy of the vast array of satellite imagery available across time. However, effectively analyzing this wealth of data to process LULC information remains a significant challenge within remote sensing. Recent times have witnessed the introduction of diverse techniques for scrutinizing satellite images, encompassing remote sensing technologies and machine/deep learning (M/DL) methods. This research endeavors to explore the transformation of LULC within the N’fis watershed, situated in the Western High Atlas region of Morocco, covering the timeline from 1984 to 2022. By harnessing remote sensing technologies, we have traced alterations in dams, forests, agriculture, and soil over this duration. Moreover, we have conducted comparisons among multiple machine and deep learning (M/DL) models to simulate and forecast LULC changes specifically for the year 2030. Our study outcomes manifest remarkable accuracy in LULC classification, consistently ranging between 91% and 97% for most years, with the kappa coefficient maintaining a range between 89% and 95%. Regarding predictive analysis, the Random Forest (RF) model emerges as the most precise, displaying an accuracy rate of 91%.

Landslide susceptibility assessment and prediction are among the main processing for disaster management and land use planning activities. Therefore, the general purpose of this research was to evaluate GIS-based spatial modeling of landslides in the western Algiers Province using five models, namely, frequency ratio (FR), weights of evidence (WoE), evidential belief function (EBF), logistic regression (LR), and analytical hierarchy process (AHP), and then compare their performances. At first, a landslide inventory map was prepared according to Google Earth satellite images, historical records, and extensive field surveys. The recorded landslides were divided into two groups (70% and 30%) to establish the training and validation models. In the next step, GIS techniques and remote sensing data were used, to prepare a spatial database containing 13 landslide conditioning factors: lithology, distance to lithological boundaries, permeability, slope, exposure, altitude, profile curvature, plan curvature, precipitation, distance to rivers, topographic wetness index, normalized difference vegetation index, and distance to roads. Finally, the landslide susceptibility maps were produced using the five models and validated by the areas under the relative operative characteristic curve (AUC). The AUC results showed a significant improvement in susceptibility map accuracy; the FR model has the best performance in the training and prediction process (90%), followed by LR (88%, 89%), WoE (88%, 87%), EBF (86%,86%), and AHP (76%,75%), respectively. The produced maps in the current study could be useful for land use planning and hazard mitigation purposes in western Algiers Province.

For areas without perennial surface water sources, groundwater might be considered the second-largest source of drinking water after surface water. However, groundwater is highly prone to contamination as the groundwater reservoir is formed by the movement of surface water into the subsoil; in its due course of motion, it may dissolve any probable contaminants such as agrochemicals, landfill leachates, the oil spill from underground pipelines, and sewer waste and further convey the contaminated water to join some groundwater aquifers from where the water is again pumped out for human consumption. Therefore, prior to its potable applicability, the quality of groundwater should be evaluated for the presence of alkalinity, hardness, and undesirable and heavy minerals. The Central Ground Water Board (CGWB), Bhubaneswar, collects data on 61 stations in the Kalahandi District for 15 physiochemical parameters, including pH, bicarbonate, hardness, sulphate, Cl⁻, total dissolved solids, Mg⁺⁺, K⁺, Na⁺, total alkalinity, nitrate, fluoride, carbonate, electrical conductivity, and calcium, to assess the quality of the groundwater. The goals were to pinpoint the major elements influencing water quality and comprehend the groundwater quality measures’ regional distribution. Data from the Central Groundwater Board (CGWB) were collected as part of our research, and PCA was used to identify the major impacting elements. To further minimize the dataset’s multidimensionality, a principal component analysis is used. Together, the first three major components explain 76.64% of the overall variability. The first two principal factors themselves explain about 56.9% of the total variance. The three principal factors indicate salinity, hardness, and relative alkalinity and acidity, respectively, in the groundwater.

To manage the historic built heritage, it is of fundamental importance to fully understand the urban area under study, so that all its characteristics and critical issues related to historical conformation, stratification, and transformations can be better understood and described. Geometric surveying allows a deeper investigation of these characteristics through analytical investigation in support of urban planning theories as well. To date, geomatics provides various tools and techniques to meet the above-mentioned needs, and mobile mapping system (MMS) is a technology that can survey large areas in a short time, with good results in terms of density, accuracy, and coverage of the data. In this context, the article aims to verify whether this approach can also be useful in the complex and stratified reality of the historic urban context. The case analyzed—the historical center of Sabbioneta—presents some criticalities found in many urban centers of historical layout. Examples are narrow streets inserted in an urban context with multi-story buildings and consequent difficulty in receiving the GNSS signal and difficulty in following general MMS survey guidelines (trajectories with closed loops, wide radius curves). The analysis presented, relating to a survey carried out with Leica Pegasus:Two instrumentation, in addition to describing the strategies used to properly develop the survey, aims to analyze the resulting datum by discussing its possibilities for use in urban modeling, for cartographic or three-dimensional information modeling purposes. Particular attention is paid to assessing whether the quality of the data (accuracy, density) is suitable for the urban scale. Finally, an analysis of the data obtained from MMS was made with the geographic-topographic database (DBGT), in a GIS (Geographic Information System) environment, to check the possibilities of use and integration between the two models.