Classification Of Land Cover Types Research Articles

Soil Organic Carbon Prediction Based on Vis-NIR Spectral Classification Data Using GWPCA-FCM Algorithm.

Soil visible and near-infrared reflectance spectroscopy is an effective tool for the rapid estimation of soil organic carbon (SOC). The development of spectroscopic technology has increased the application of spectral libraries for SOC research. However, the direct application of spectral libraries for SOC prediction remains challenging due to the high variability in soil types and soil-forming factors. This study aims to address this challenge by improving SOC prediction accuracy through spectral classification. We utilized the European Land Use and Cover Area frame Survey (LUCAS) large-scale spectral library and employed a geographically weighted principal component analysis (GWPCA) combined with a fuzzy c-means (FCM) clustering algorithm to classify the spectra. Subsequently, we used partial least squares regression (PLSR) and the Cubist model for SOC prediction. Additionally, we classified the soil data by land cover types and compared the classification prediction results with those obtained from spectral classification. The results showed that (1) the GWPCA-FCM-Cubist model yielded the best predictions, with an average accuracy of R2 = 0.83 and RPIQ = 2.95, representing improvements of 10.33% and 18.00% in R2 and RPIQ, respectively, compared to unclassified full sample modeling. (2) The accuracy of spectral classification modeling based on GWPCA-FCM was significantly superior to that of land cover type classification modeling. Specifically, there was a 7.64% and 14.22% improvement in R2 and RPIQ, respectively, under PLSR, and a 13.36% and 29.10% improvement in R2 and RPIQ, respectively, under Cubist. (3) Overall, the prediction accuracy of Cubist models was better than that of PLSR models. These findings indicate that the application of GWPCA and FCM clustering in conjunction with the Cubist modeling technique can significantly enhance the prediction accuracy of SOC from large-scale spectral libraries.

Land Cover Dynamics and Environmental Implications in Türkiye’s Riparian Zones: A Comprehensive Seven-Year Study

This study examines the Riparian Buffer Zones (RBZs) in Türkiye, analyzing land cover changes from 2017 to 2023 to assess their effectiveness in ecological conservation and water resource management. The research employs Esri-Sentinel land cover data, Shuttle Radar Topography Mission digital elevation models, and RiverATLAS data to identify and delineate RBZs, classify land cover types, and evaluate trends and changes over seven years. The findings indicate that 44.8% of RBZs are covered by ‘Rangelands’, 32.2% by ‘Crops’, and 16.1% by ‘Trees’ classes. The study notes a considerable increase in built areas and agricultural fields, which may indicate ongoing urban expansion and its potential impact on the effectiveness of RBZs. Despite regulatory efforts, the substantial presence of agriculture within RBZs indicates a need for more robust environmental protection policies. The results indicate a pressing need for a policy reassessment to enhance forest cover within RBZs and reduce agricultural encroachment. This would improve water quality and facilitate the achievement of sustainability goals numbered 2, 3, 6, 11, 14 and 15. The findings thus call for a strategic enhancement of RBZ management to better protect water resources against the adverse effects of agricultural and urban activities.

Performance evaluation of Machine Learning algorithms for LULC classification: A case study of Fez-Meknes region

Significant advancements have been made in remote sensing technologies, with increasingly refined applications in creating LULC maps. The integration of Machine Learning-based approaches has been explored to develop LULC maps with varying levels of precision, leveraging diverse satellite imagery. However, the task of producing LULC maps for extensive areas like the Fez-Meknes region, covering an area of approximately 40,075 km2, can be challenging using traditional methodologies. Thus, this study prioritized the major objective of establishing a reference for extracting LULC information. This endeavour involves the comparative assessment of the performance of different LULC classification approaches: Recursive Partitioning (Rpart), k-nearest neighbors (KNN), random forest (RF), Linear Discriminant Analysis (LDA), support vector machine (SVM), and extreme gradient boosting (XGBoost). For map production, remote sensing data and a supervised classification algorithm based on LANDSAT images of the Fez-Meknes region were employed. The accuracy of the generated maps was assessed using overall accuracy and Kappa coefficient. This methodology holds the potential to be replicated in other regions, utilizing a variety of available remote sensing satellite images to generate LULC maps. Essentially, the approach proposed in this study will be a valuable tool for planners, facilitating the acquisition of LULC maps at various time intervals. This will facilitate the classification of land cover types in a faster and more cost-effective manner.

A Method of Chestnut Forest Identification Based on Time Series and Key Phenology from Sentinel-2

Chestnut trees hold a prominent position in China as an economically significant forest species, offering both high economic value and ecological advantages. Identifying the distribution of chestnut forests is of paramount importance for enhancing efficient management practices. Presently, many studies are employing remote sensing imaging methods to monitor tree species. However, in comparison to the common classification of land cover types, the accuracy of tree species identification is relatively lower. This study focuses on accurately mapping the distribution of planted chestnut forests in China, particularly in the Huairou and Miyun regions, which are the main producing areas for Yanshan chestnuts in northeastern Beijing. We utilized the Google Earth Engine (GEE) cloud platform and Sentinel-2 satellite imagery to develop a method based on vegetation phenological features. This method involved identifying three distinct phenological periods of chestnut trees: flowering, fruiting, and dormancy, and extracting relevant spectral, vegetation, and terrain features. With these features, we further established and compared three machine learning algorithms for chestnut species identification: random forest (RF), decision tree (DT), and support vector machine (SVM). Our results indicated that the recognition accuracy of these algorithms ranked in descending order as RF > DT > SVM. We found that combining multiple phenological characteristics significantly improved the accuracy of chestnut forest distribution identification. Using the random forest algorithm and Sentinel-2 phenological features, we achieved an impressive overall accuracy (OA) of 98.78%, a Kappa coefficient of 0.9851, and a user’s accuracy (UA) and producer’s accuracy (PA) of 97.25% and 98.75%, respectively, for chestnut identification. When compared to field surveys and official area statistics, our method exhibited an accuracy rate of 89.59%. The implementation of this method not only offers crucial data support for soil erosion prevention and control studies in Beijing but also serves as a valuable reference for future research endeavors in this field.

The response of ecological security to land use change in east and west subtropical China.

Regional land use change and ecological security have received considerable attention in recent years. The rapid economic development of Kunming and Fuzhou has resulted in environmental damage such as water pollution and urban heat island effect. It is thus important to conduct a comparative analysis of the ecological security response to land use/land cover change (LUCC) in different natural zones. Using the Google Earth Engine (GEE) platform, random forest and support vector machine methods were used to classify land cover types in the study area, after which the ArcGIS platform was used to analyze LUCC. The driving force-pressure-state-impact-response (DPSIR) model and entropy weight method were used to construct an ecological security evaluation system, and gray correlation was used to compare the ecological security responses to LUCC in Kunming and Fuzhou. The findings revealed that: (1) The average dynamic degrees of comprehensive land use in Kunming and Fuzhou from 1995 to 2020 were 1.05% and 0.55%, respectively; (2) From 1995 to 2020, the ecological security index values for Kunming and Fuzhou increased from 0.42 to 0.52 and from 0.36 to 0.68, respectively, indicating that Fuzhou's index is rising more rapidly; and (3)There is a strong correlation between LUCC and ecological security, the correlation between the woodland and the ecological security index is very strong in both places. The expansion of construction land may be an important reason for the reduced ecological security level in Fuzhou City, while water resources have a significant impact on the ecological security level of Kunming City.

UNET NEURAL NETWORK IN AGRICULTURAL LAND COVER CLASSIFICATION USING SENTINEL-2

Abstract. The article discusses a method for classifying land cover types in rural areas using a trained neural network. The focus is on distinguishing agriculturally cultivated areas and differentiating bare soil from quarry areas. This distinction is not present in publicly available databases like CORINE, UrbanAtlas, EuroSAT, or BigEarthNet. The research involves training a neural network on multi-temporal patches to classify Sentinel-2 images rapidly. This approach allows automated monitoring of cultivated areas, determining periods of bare soil vulnerability to erosion, and identifying open-pit areas with similar spectral characteristics to bare soil. After training the U-Net network, it achieved an average classification accuracy of 90% (OA) in the test areas, highlighting the importance of using OA for multi-class classifications, instead of ACC. Analysis of our main classes revealed high accuracy, 99.01% for quarries, 92.3% for bare soil, and an average of 94.8% for annual crops, demonstrating the model's capability to differentiate between crops at various growth stages and assess land cover categories effectively.

Мониторинг земель лесного фонда провинции Кон Тум Вьетнама по данным дистанционного зондирования Земли

In recent years, cloud computing technology has become increasingly useful and practical in many fields, including forestry. Monitoring forest dynamics throughout a relatively large area requires collecting a large amount of input data, and processing it is very complex and time-consuming. In this study, we demonstrated the potential of applying cloud computing technology in the Google Earth Engine platform, in conjunction with remote sensing data to monitor forest land changes in Kon Tum province, Vietnam. The use of the javascript editor on Google Earth Engine (GEE) automated the process of collecting and processing remote sensing data to meet the specified criteria, while saving time, effort, and computer resources. Computing the normalized difference vegetation index and classifying land cover types using the Random Forest machine learning method on the GEE platform also showed accuracy in representing the distribution of vegetation cover and evaluating the status and changes in forest areas in Kon Tum province. The study showed that the policies of Kon Tum province administration in recent years have had a positive impact on restoring natural forest areas and reducing resource losses. So, the application of remote sensing data on the cloud computing platform of Google Earth Engine is a promising method for conserving and managing forest resources in Kon Tum province and throughout Vietnam

Spatio-Temporal Evolution of Forest Landscape in China’s Giant Panda National Park: A Case Study of Jiudingshan Nature Reserve

The continuous prohibition of commercial logging and intensifying conservation endeavors have encompassed the implementation of the Natural Forest Conservation Program (NFCP) and the Grain-to-Green Program (GTGP) by the Chinese government since 1999. Nevertheless, the efficacy of the commercial logging ban and its effectiveness in halting deforestation remain uncertain. Likewise, the destructive aftermath of the 7.9 magnitude Wenchuan earthquake in 2008 continues to be under scrutiny, necessitating ongoing study and analysis. Thus, there exists a pressing need to comprehensively monitor the spatio-temporal evolution of the forest habitat and assess the ecological status over the past two decades. The Jiudingshan Nature Reserve (JNR) is situated in the upper reaches of the Tuojiang River basin in Sichuan province, China, constituting an integral part of the Giant Panda National Park (GPNP). In this study, we classified land cover types and conducted a meticulous monitoring of forest habitat alterations within JNR, by a multilayer perceptron model (MLP) with a highly learning-sensitive algorithm. To quantify these changes, the Simple Ratio Index (SRI) and the Normalized Difference Vegetation Index (NDVI) were computed from Landsat TM/OLI images of four years (i.e., 1997, 2007, 2008, and 2018). Additionally, elevation, slope, aspect, and other topographic data were acquired from the Digital Elevation Model (DEM). The findings of our study unveil a notable expansion in both the scope and proportion of mixed conifer and broadleaf forest from 1997 to 2004. The growth of coniferous forest and the augmented areas of mixed conifer and broadleaf forest signify a substantial improvement in panda habitat. However, the seismic event of 2008 exhibited a pronounced adverse impact on vegetation, particularly within forested regions. Although there is evidence of forest recovery spanning 21 years, concerns regarding fragmentation linger. It is pivotal to acknowledge the potential long-term adverse implications arising from widespread socio-economic development and a multitude of geohazards. Hence, sustained long-term monitoring coupled with effective management strategies remain pivotal for the preservation and rehabilitation of the Giant Panda National Park (GPNP) and giant panda habitat in the future.

Supervised Classification of Tree Cover Classes in the Complex Mosaic Landscape of Eastern Rwanda

Eastern Rwanda consists of a mosaic of different land cover types, with agroforestry, forest patches, and shrubland all containing tree cover. Mapping and monitoring the landscape is costly and time-intensive, creating a need for automated methods using openly available satellite imagery. Google Earth Engine and the random forests algorithm offer the potential to use such imagery to map tree cover types in the study area. Sentinel-2 satellite imagery, along with vegetation indices, texture metrics, principal components, and non-spectral layers were combined over the dry and rainy seasons. Different combinations of input bands were used to classify land cover types in the study area. Recursive feature elimination was used to select the most important input features for accurate classification, with three final models selected for classification. The highest classification accuracies were obtained for the forest class (85–92%) followed by shrubland (77–81%) and agroforestry (68–77%). Agroforestry cover was predicted for 36% of the study area, forest cover was predicted for 14% of the study area, and shrubland cover was predicted for 18% of the study area. Non-spectral layers and texture metrics were among the most important features for accurate classification. Mixed pixels and fragmented tree patches presented challenges for the accurate delineation of some tree cover types, resulting in some discrepancies with other studies. Nonetheless, the methods used in this study were capable of delivering accurate results across the study area using freely available satellite imagery and methods that are not costly and are easy to apply in future studies.

Research on land cover type classification method based on improved MaskFormer for remote sensing images.

High-resolution remote sensing images have the characteristics of wide imaging coverage, rich spectral information and unobstructed by terrain and features. All of them provide convenient conditions for people to study land cover types. However, most existing remote sensing image land cover datasets are only labeled with some remote sensing images of low elevation plain areas, which is highly different from the topography and landscape of highland mountainous areas. In this study, we construct a Qilian County grassland ecological element dataset to provide data support for highland ecological protection. To highlight the characteristics of vegetation, our dataset only includes the RGB spectrum fused with the near-infrared spectrum. We then propose a segmentation network, namely, the Shunted-MaskFormer network, by using a mask-based classification method, a multi-scale, high-efficiency feature extraction module and a data-dependent upsampling method. The extraction of grassland land types from 2 m resolution remote sensing images in Qilian County was completed, and the generalization ability of the model on a small Gaofen Image Dataset (GID) verified. Results: (1) The MIoU of the optimised network model in the Qilian grassland dataset reached 80.75%, which is 2.37% higher compared to the suboptimal results; (2) the optimized network model achieves better segmentation results even for small sample classes in data sets with unbalanced sample distribution; (3) the highest MIOU of 72.3% is achieved in the GID dataset of open remote sensing images containing five categories; (4) the size of the optimized model is only one-third of the sub-optimal model.

Radiometric Correction Model and Land Cover Classification of Snow-Covered Terrains for ICESat-2 Photon-Counting Lidar

The signal strength is a fundamental parameter in radiometric applications for satellite lidars. Different from full-waveform lidars, photon-counting lidars cannot record the returned signal strength but only respond to the presence of the photon event and may miss some returned photons due to the dead time effect, i.e., introduce radiometric distortion. Based on the lidar equation and the response mechanism of photon-counting detectors, we propose a radiometric correction model to remove the impact of the nonlinear response and dead time of detectors for the photon-counting lidar borne on ICESat-2. The returned signal photon number is corrected by the proposed model with respect to the photon event number per shot (PNPS) and surface slope derived from ATL03/ATL08 products. Then, the optical throughput calibration factor of ICESat-2 is obtained from ATL06 products over high Antarctic plateau where has given reflectance and clear atmosphere, which is generally equal to 0.52. The atmospheric attenuation induced by the molecular, cloud, and aerosol is calculated from ATL09 products. In addition, the corrected radiometric parameters including the calculated surface reflectance and apparent surface reflectance (ASR) are applied to classify land cover types along laser tracks over snow-covered terrains. The results indicate that the signal strength and calibration constant are reliable after corrections, but the atmospheric attenuation is sometimes inaccurate, which further influences the derived surface reflectance. In classifications, the overall accuracy and Kappa coefficient based on the corrected ASR can achieve the best classification results with 88.80% and 0.69. The proposed radiometric correction model is very essential to radiometric applications for photon-counting lidars such as ICESat-2, especially for data captured on ice and bare land with relatively high reflectance.

An Analysis of Surface Temperature Changes for Urban Green Space using Unmanned Aerial Vehicles

Background and objective: The urban heat island (UHI) effect is recognized as a representative environmental problem that occurs in cities in summer. This study aimed to quantitatively determine the surface temperature (ST) of UGSs using high-resolution images taken by an unmanned aerial vehicle (UAV), and analyze time-series changes in ST according to spatial characteristics (conifers, deciduous trees, shrubs, grass, metal sculptures, pavements).Methods: In this study, ST data of UGSs were established and acquired through UAV flight and filming, and orthoimages of such data were produced using the Pix4D program. In addition, by comparing RGB orthoimages and green space status data (location of trees and facilities) obtained from field surveys, a green-space type map was prepared using ArcGIS (v10.3.1) software to classify land cover types in green spaces (GSs). ST distribution by GS type was analyzed and statistical significance was verified through one-way ANOVA.Results: As a result, the ST of conifers, deciduous trees, shrubs, and grass, which are vegetation, was found to be lower than that of paved roads: for conifers, 4.1-12.5℃ lower than paved roads; for deciduous trees, 3.0-10.8℃; for shrubs 3.4-11.2℃; and for grass 1.7-8.1℃. In addition, the variations in ST over time were greatest for metal sculptures (28.1℃), followed by pavement (20.4℃), grass (19.4℃), shrubs (14.0℃), conifers (13.3℃), and deciduous trees (13.0℃).Conclusion: Based on the results of this study, it is necessary to consider the components of GS for the efficient planning and management of UGS in terms of improving the urban thermal environment. Insufficient and unsystematic planning and management of UGSs may deteriorate the function of GSs. Therefore, it is necessary to determine and evaluate the ST characteristics of GSs in terms of improving the urban thermal environment.

Determination of the optimum number of sample points to classify land cover types and estimate the contribution of trees on ecosystem services using the I-Tree Canopy tool.

The process of producing information about dynamic land use and land cover and ecosystem health quickly with high accuracy and low cost is important. This information is one of the basic data used for sustainable land management. For this purpose, remote sensing technologies are generally used, and sampling points are mostly assigned. Determination of the optimum number of sampling points using the I-Tree Canopy tool was the main focus of this study. The I-Tree Canopy tool classifies land cover, revealing the effects of tree cover on ecosystem services, such as carbon (C) sequestration and storage, temperature regulation, air pollutant filtering, and air quality improvement, with numerical data. It is used because it is practical, open source, and user-friendly. This software works based on sampling point assignment, but it is unclear how many sampling points should be assigned. Therefore, determining the optimum number of sample points by statistical methods will increase the effectiveness of this tool and guide users. For this purpose, reference data were created for comparison. Then, 31 I-Tree Canopy reports were created with 100-point increments up to 3100. The data obtained from the reports were compared with the reference data, and statistical analysis based on Gaussian and a second-order polynomial fit was performed. At the end of the analysis, the following results were obtained; the results of this study demonstrated that the optimum number of sample points for a 1-ha area is 760 ± 32 from the comparison of the real area and I-Tree Canopy results. Similar results from the Gaussian fit of annually sequestered and stored C and carbon dioxide (CO2 ) amounts in trees and the reduction in air pollution in grams were obtained as 714 ± 16. Therefore, we may conclude that taking more than 800 sample points will not be statistically significant. Integr Environ Assess Manag 2023;19:726-734. © 2022 SETAC.

Comparing Machine and Deep Learning Methods for the Phenology-Based Classification of Land Cover Types in the Amazon Biome Using Sentinel-1 Time Series

The state of Amapá within the Amazon biome has a high complexity of ecosystems formed by forests, savannas, seasonally flooded vegetation, mangroves, and different land uses. The present research aimed to map the vegetation from the phenological behavior of the Sentinel-1 time series, which has the advantage of not having atmospheric interference and cloud cover. Furthermore, the study compared three different sets of images (vertical–vertical co-polarization (VV) only, vertical–horizontal cross-polarization (VH) only, and both VV and VH) and different classifiers based on deep learning (long short-term memory (LSTM), Bidirectional LSTM (Bi-LSTM), Gated Recurrent Units (GRU), Bidirectional GRU (Bi-GRU)) and machine learning (Random Forest, Extreme Gradient Boosting (XGBoost), k-Nearest Neighbors, Support Vector Machines (SVMs), and Multilayer Perceptron). The time series englobed four years (2017–2020) with a 12-day revisit, totaling 122 images for each VV and VH polarization. The methodology presented the following steps: image pre-processing, temporal filtering using the Savitsky–Golay smoothing method, collection of samples considering 17 classes, classification using different methods and polarization datasets, and accuracy analysis. The combinations of the VV and VH pooled dataset with the Bidirectional Recurrent Neuron Networks methods led to the greatest F1 scores, Bi-GRU (93.53) and Bi-LSTM (93.29), followed by the other deep learning methods, GRU (93.30) and LSTM (93.15). Among machine learning, the two methods with the highest F1-score values were SVM (92.18) and XGBoost (91.98). Therefore, phenological variations based on long Synthetic Aperture Radar (SAR) time series allow the detailed representation of land cover/land use and water dynamics.

Intelligent classification of land cover types in open-pit mine area using object-oriented method and multitask learning

Although the exploitation of mineral areas brings wealth to society, it inevitably leads to the degradation of the surrounding natural environment. To understand and assess the influences of mining activities on the geological and ecological environment, land cover classification in open-pit mine areas (LCCMA) is of great significance. This research proposes an intelligent classification framework for LCCMA based on an object-oriented method and multitask learning (MTL), named the MTL Classification Framework (MTLCF). With the help of MTL, each land cover type in open-pit mine areas obtains its exclusive and receivable object-oriented feature sets using the model-agnostic method. After that, the feature sets are fused with the original images. EfficientNet, a spatial pyramid pooling module, and a global attention upsample module are assembled as the segmentation models with the structure of the encoder and decoder to classify intelligently each land cover type in open-pit mine areas. Finally, the models were trained, and ablation experiments were performed. The experimental results show that our proposed framework -MTLCF was effective for classification in LCCMA, and the overall accuracy and the mean of F1 score for the MTLCF in LCCMA were 85.6% and 86.06%, respectively.

Land Cover Classification from Hyperspectral Images via Weighted Spatial-Spectral Kernel Collaborative Representation with Tikhonov Regularization

Precise and timely classification of land cover types plays an important role in land resources planning and management. In this paper, nine kinds of land cover types in the acquired hyperspectral scene are classified based on the kernel collaborative representation method. To reduce the spectral shift caused by adjacency effect when mining the spatial-spectral features, a correlation coefficient-weighted spatial filtering operation is proposed in this paper. Additionally, by introducing this operation into the kernel collaborative representation method with Tikhonov regularization (KCRT) and discriminative KCRT (DKCRT) method, respectively, the weighted spatial-spectral KCRT (WSSKCRT) and weighted spatial-spectral DKCRT (WSSDKCRT) methods are constructed for land cover classification. Furthermore, aiming at the problem of difficulty of pixel labeling in hyperspectral images, this paper attempts to establish an effective land cover classification model in the case of small-size labeled samples. The proposed WSSKCRT and WSSDKCRT methods are compared with four methods, i.e., KCRT, DKCRT, KCRT with composite kernel (KCRT-CK), and joint DKCRT (JDKCRT). The experimental results show that the proposed WSSKCRT method achieves the best classification performance, and WSSKCRT and WSSDKCRT outperform KCRT-CK and JDKCRT, respectively, obtaining the OA over 94% with only 540 labeled training samples, which indicates that the proposed weighted spatial filtering operation can effectively alleviate the spectral shift caused by adjacency effect, and it can effectively classify land cover types under the situation of small-size labeled samples.

Improving Land Cover Classification Using Genetic Programming for Feature Construction

Genetic programming (GP) is a powerful machine learning (ML) algorithm that can produce readable white-box models. Although successfully used for solving an array of problems in different scientific areas, GP is still not well known in the field of remote sensing. The M3GP algorithm, a variant of the standard GP algorithm, performs feature construction by evolving hyperfeatures from the original ones. In this work, we use the M3GP algorithm on several sets of satellite images over different countries to create hyperfeatures from satellite bands to improve the classification of land cover types. We add the evolved hyperfeatures to the reference datasets and observe a significant improvement of the performance of three state-of-the-art ML algorithms (decision trees, random forests, and XGBoost) on multiclass classifications and no significant effect on the binary classifications. We show that adding the M3GP hyperfeatures to the reference datasets brings better results than adding the well-known spectral indices NDVI, NDWI, and NBR. We also compare the performance of the M3GP hyperfeatures in the binary classification problems with those created by other feature construction methods such as FFX and EFS.

Analysis of Land use and Land cover Characteristics of Kanke, Pankshin, and Langtang Local Government Areas, Plateau State, Nigeria

This study assessed landuse and landcover changes and how they affect the agrarian production in Kanke, Pankshin and Langtang North Local Government Areas of plateau state, Nigeria. The study adopted the survey design. The primary data was obtained through field observations, interview of stakeholders, satellite (landsat8 and shuttle radar terrain mission) and questionnaire administration which asked questions on socioeconomic status of respondents, constraints to farmers and the respondents’ perception on the existing adaptation strategies in place. ArcGIS10.6version software was employed for the classification of land cover types while supervised classification method was adopted using maximum likelihood algorithm for the classification of feature types. Data generated by Landsat8 and ArcGIS10.6 version software were subjected to Pan-Sharpening processing for clarity of terrain features. The study findings revealed that 66.13% of the earth’s surface in the study area is covered by rock outcrops while water body; one of man’s most precious resources occupied less than 1% (0.15%). The distribution of farmers based on constraints to farming indicated that poor soils and small farm land sizes constituted the most severe challenges to farming activities in the study area. As an adaptation strategy to inadequate farm lands, terrace farming practice and dry season farming, large scale quarrying activities is recommended to serve as a source of employment and income to authorities and a means of surface leveling to convert the dominant rock outcrops to productive land and prioritizing farming in the limited plains.

A METHOD FOR QUALITATIVELY MAPPING THE FLOOD PHYSICAL VULNERABILITY OF RESIDENTIAL AREAS

Mapping flood physical vulnerability is spatially limited because it requires input data such as building structures and materials, which are unavailable on large spatial scales. In this study, we propose a new method for qualitatively evaluating the flood vulnerability of residential areas in the context of the exposure and resilience to flood hazard on large spatial scales. This method utilizes the possible correlations between the structural physical vulnerability and residential types obtained from the statistical classifications of multispectral satellite images. Because multispectral classification is well-established as an inexpensive technique for automatically classifying land cover types over wide areas, our method is feasible and efficient for mapping the physical vulnerability of residential areas. As a case study, we present an application of the proposed approach to the Thach Ha district, Ha Tinh province, Vietnam, using the Japanese type 2 Advanced Visible and Near Infrared Radiometer (AVNIR-2) images and Phased Array type L-band Synthetic Aperture Radar (PALSAR) images captured by the Advanced Earth Observing Satellite (ADEOS).

Assessing the use of cross-orbit Sentinel-1 images in land cover classification

ABSTRACT Land cover is the easiest detectable indicator of human intervention on land. Urban, peri-urban and agriculture areas present a complex combination of land cover, which makes classification challenging. Getting more detailed information is the aim of any classification method. In this study, improving land cover type classification using cross-orbit Sentinel-1 images is evaluated. To avoid uncertainties, three sites in different weather conditions and locations but approximately the same land cover is selected. For each study area, three datasets including polarimetric features extracted from (1) ascending orbit image, (2) descending orbit image, and (3) combined ascending and descending orbit images are produced and used for classification. Land cover classification is performed following a supervised Support Vector Machine (SVM) exploiting all three datasets. Consequently, the radar cross-orbit integrated dataset produced the most accurate land cover map. Classifications show overall accuracies (OA) of 84%, 85%, and 75%, and Kappa coefficients (K) of 0.67, 0.75, and 0.55 for Skane, Tehran, and Sherbrooke regions, respectively. Fortunately, the accuracy results are at least 4% better than single-orbit classification. In other words, the proposed mapping approach has proved that using information from both the ascending and descending dual-polarized images could achieve a more accurate classification map than using a single-orbit image individually.