Papers by Keyword: Remote Sensing

Abstract: Deforestation is a significant threat to the sustainability of the ecosystem, leading to adverse effects such as climate change, biodiversity loss, and socio-economic consequences. Timely monitoring of forest destruction enables effective implementation of preventive mechanisms supported by law enforcement. Advancements in remote sensing, coupled with enhanced deep learning techniques, boost efficient deforestation monitoring as these technologies support real-time analysis of complex satellite images. Thus, this study aimed to develop a classification model to identify forest areas from non-forest areas using Landsat-8 data acquired for Wilpattu park, Sri Lanka, between 2015 to 2024. We explored model building using minimal input of two bands in satellite data, facilitating low resource needs. Seven deep learning models were explored, progressing from Convolution Neural Networks to Transformer-based models to build the classifier using a set of patches of size 100×100. The results were evaluated using standard metrics such as accuracy, precision, recall, F1 score, and Kappa index. We found that SegNet outperformed the remaining models with an overall accuracy of 96.36%, F1 score of 0.97, and Kappa index of 0.92, demonstrating excellent ability to distinguish the classes. However, the efficiency of the model needs further improvement. The proposed system will contribute to deforestation detection, offering a simpler model development approach with minimum input requirements. The proposed method can be adopted to other domains where the chosen band combination supports effective detection, such as water body identification.

Abstract: In this study, survey data from distinct sources includes Total Station measurements, ArcGIS, and Global Mapper at a planned dam location at Qila Abdullah district in Balochistan are compared. Because of the study area's complicated topography, a precise elevation assessment is essential to the feasibility of the project. Elevation information from open-source NASA SRTM models integrated using ArcGIS and Global Mapper was contrasted with high-precision Total Station data, which served as the reference benchmark. In the field, 146 survey points were gathered, and AutoCAD Civil 3D was used to process all of the datasets for a thorough analysis. Significant elevation differences between the datasets were observed in the results, with SRTM-based models demonstrating large deviations from ground-truth observations. The analysis between total station and ArcGIS Pro reports a mean error of 15.647917m and standard deviation of 3.894677m. The results between total station and Global Mapper give similar results reporting a mean error of 14.870448m and standard deviation of 3.960269m. These differences directly affect feasibility studies, especially when it comes to cost estimation, design precision, and possible overestimation of material requirements. Because relying on broad open-source data might result in significant errors in project planning and execution, the findings point out the significance of accurate survey methodologies for infrastructure projects in rough terrains.

Abstract: Poor waste handling causes problems in the form of increasing waste generation. The availability of waste treatment plants is needed to process waste before it is disposed of in landfills. In addition, rapid population growth leads to increased waste production. Therefore, the procurement of Reduce, Reuse, Recycle-based Waste Management Sites can be a solution. The procurement of Reduce, Reuse, Recyclebased Waste Management Sites must be designed by taking into account regulations and the physical appearance of the area. This needs to be done to protect the surrounding environment and not interfere with residents' activities. In planning the determination of 3R Waste Processing Sites locations, remote sensing and geographic information systems can be utilized, as well as the Analytical Hierarchy Process (AHP) approach to map the suitability of 3R Waste Processing Sites locations. In data processing, 5 parameters are used, land use, road class, distance to roads, distance to settlements, and distance to rivers. The results of the AHP analysis show that land use is the criteria with the highest weight at 53%, distance to the road is worth 22%, road class is worth 11%, distance to settlements is worth 8%, and distance to the river is 7%. The AHP model built has a Consistency Ratio (CR) value of 6.7% so it is considered valid. The final results show that the areas suitable for the provision of 3R Waste Processing Sites are bushes, vacant land, and plantations in the western part of Singosaren Village.

Abstract: An exploratory study was conducted using Landsat-8 (L8) and Sentinel-2 (S2) satellite images for the extraction of chlorophyll-a and SST, followed by determining their relationship with fish catch at Baron Beach due to the increasing fish catch in certain months. L8 and S2 can observe chlorophyll-a but not as optimum as low-resolution imagery such as MODIS due to the complex optical characteristics of seawater and their limited band types. Aside from observing the chlorophyll-a, L8 can observe SST value but S2 cannot because it currently has no thermal infrared band. Therefore, both images need to be compared to know their capability of extracting chlorophyll-a and SST. Data processing for chlorophyll-a and SST extraction used Google Earth Engine (GEE) and QGIS. Data extraction preparation involved cloud masking with four scenarios. Chlorophyll-a extraction used Ocean Color (OC) algorithm, while SST extraction at L8 used thermal-infrared band and optical band approach at S2. Differences in extraction results were analyzed using a non-parametric significance test with α = 0.05. The relationship between chlorophyll-a, SST and fish catch was assessed using Catch per Unit Effort (CPUE) values in the Spearman correlation test. The extraction results showed changes in chlorophyll-a and SST values each month in 2022 where both images show an increasing chlorophyll-a within June until October and decreasing within those months. However, the extraction results from both images are significantly different. Aside from the significantly different extraction results, there is a positive correlation between chlorophyll-a and fish catch, but the SST correlation varied between L8 and S2 images. This difference is thought to be caused by image characteristics, cloud masking, and extraction models that are not yet suitable for the Baron coastal area, which is characterized by high sedimentation coastal areas. In this context, correlation analysis showed a relationship between chlorophyll-a concentration and SST with fish production, but direct comparison data at Baron Beach is needed for further analysis.

Abstract: Monitoring of Land Use and Land Cover changes is very important in the planning and management programs required for development activities at the regional levels of any country. The main goal of this study is to observe the dynamics of the vegetation cover of Siyazan region for 7 years from 2010 to 2017 using Remote Sensing and Geographic Information System. Sentinel-2A MSI (Multi Spectral Imager), Landsat-5 TM (Thematic Mapper) satellite images are used to create vegetation maps. Vegetation change in the study area is calculated by the Normalized Vegetation Cover Index (NDVI), and the results show that the vegetation cover increased from 0.8% in 2010 to 22.5% in 2017. Supervised classification is performed using the Maximum Likelihood Classification (supervised classification). The 5 main classes considered for classification are: Watersheds, cropland/vegetation, gray land, settlements and productive land. The ArcGIS software package is used to carry out the proposed study and the accuracy assessment is carried out by taking the base values for appropriate classification through the Google Earth Pro software. The results show that the overall accuracy of the proposed system is 78.12%.

Abstract: On April 2-4, 2021, severe cyclone Seroja occurred in Kupang City, resulting in a landslide hazard. Landslide disaster management is essential for disaster mitigation. However, the susceptibility landslide map for the cyclone Seroja in Kupang City has not yet been done. In this paper, we map landslide locations using remote sensing. We used free remote sensing data. The data are Landsat 8 imagery, Digital Elevation Model (DEM), and rainfall. To create a landslide map in Kupang City, we overlaid three maps. Those are landcover, slope and rainfall maps. We created a landcover map using the Landsat 8 imagery. The landcover types were classified using the Supervised Classification method of Support Vector Machines (SVM) in the QGIS software. Each land cover type was determined based on training sites taken in the field using GPS. DEM was used to create a slope map. We used the accumulative rainfall data for three days to map the rainfall. For data validation, we took GPS points from the landslide locations. The study result shows a landslide map in Kupang City after cyclone Seroja. This work informs that remote sensing can be used to determine the location of landslides in inaccessible areas. Remote sensing can also be used to map landslide areas with a limited budget and limited data.

Abstract: With the increasing population, the demand for food products is increasing day by day. The agriculture sector is adapting to technological reforms of traditional processes to maintain a proper balance between the demand-supply relationship. The intervention of the technology is resulting in the enhanced productivity of the agricultural process, and at the same time, it is also helping in the workload management of the farmers. In the last two decades, unmanned aerial vehicles (UAVs) or drones have emerged as indispensable tools in modern agricultural processes. Drones and allied smart technologies are being used for a variety of applications in this sector. This work presents a comprehensive survey of drones in the agriculture sector. The latest trends in the usage of drones from agricultural viewpoints are discussed. The work emphasizes the drone’s architectures, sensor integration, and availability in the open market. Furthermore, the challenges associated with this technology are also outlined.

Abstract: Remote sensing technology is essential to various industries such as agriculture, meteorology, surveillance, defence, manufacturing and processing industries. Several sectors widely adopt this technology, so much research has been conducted in this domain. In satellite applications, research in remote sensing has been performed for seven decades. Images and videos captured by satellites have less resolution, which undoubtedly reduces object detection and data analysis accuracy. After analysis, the imprecise nature of captured data might cause difficulties in fields such as defence and agriculture. To combat this problem, in this research, we developed a hexacopter-based modern remote sensing device that can fly with manual intervention and also has an emergency autopilot function. The proposed system is equipped with a compact high-resolution camera which captures images with a higher frame rate. The developed system uses the YOLO v4 algorithm, which is fast and accurate to recognise and track an item or a particular individual in real time. Logged data is shared with the ground station to perform the desired task. The hexacopter-based system has more mobility than the satellite-based system, which overcomes the drawback of the limited range of the proposed system. In this proposed system, we have connected a precise flight controller and a Raspberry Pi 3 Model A+ microprocessor board with other electronic components to more accurately control hexacopter flying and real-time object identification and tracking.

Abstract: The central dry zone area of Myanmar is characterized as vulnerable area due to water stressed and one of the most food insecure regions in the country. In this region, the adverse effects of climate change are believed to be a major constraint to vulnerability. Theses extreme climatic events are likely increase in frequency and magnitude of serious drought periods and extreme floods. For vulnerability assessment we used remote sensing (RS) and geographical information system (GIS) technology and develop a numerical model, using spatial principle component analysis (SPCA) in ArcGIS software and evaluate two decade (1995, 2005 & 2016) vulnerability evaluation. The model contains following indicators: discharge change, climate moisture, drained area, flood risk, irrigation, evapotranspiration, precipitation, surface runoff, nitrogen load and population distribution. According to the numerical results, the vulnerability is classified into five levels: slight, light, medial, heavy and very heavy level by means of the cluster principle. The results show that vulnerability in the study area from 1995 to 2016 is at medial (25%) and heavy (25%) level and presents from south-west to north east direction. The vulnerability change trend show worst situation in 1995 (29.80) and best one in 2005 (17.45) but again vulnerability was increase in 2016 (21.58). In the study area the main driving forces for dynamic change in vulnerability is the intensive land use and high population density. This spatial approach allowed the analysis of different indicators, providing a platform for data integration as well as a visually powerful overview of the study area.

Abstract: Site surface characterization is an important factor to identify a suitable area for riverbank filtration (RBF) location. However, selecting the suitable area for RBF location using conventional methods is costly and time consuming, usually restricted to a small area. In this research, a site suitability for RBF location methodology was proposed using spatial data techniques to determine the site suitability of the potential RBF location in Kota Lama Kiri, Kuala Kangsar study area. A high resolution GeoEye-1 satellite imagery acquired in 2012 was classified using the supervised classification process for land cover. The classified image was further analyze using overlaying, buffering and Boolean analysis, to identify the suitable site for RBF based on location, distance from the river and distant from built-up area. In addition, the geology and hydrological data were extracted from published maps, which were then converted and integrated into GIS spatial database. The results show that the classified GeoEye-1 image produces the overall accuracies of 83.50% % with kappa statistic value of 0.806. The site suitability map for the potential RBF locations in the study area were produced confirms the location of an existing RBF well developed by Lembaga Air Perak (LAP). The methodology can be readily used to provide information of suitability area for RBF location in which can be used by water supply management to locate the RBF well for extraction purposes.