Browsing by Author "Vishakha Sood"

Now showing 1 - 4 of 4

A novel deep learning change detection approach for estimating spatiotemporal crop field variations from Sentinel-2 imagery
(Elsevier B.V., 2024) Neelam Dahiya; Gurwinder Singh; Dileep Kumar Gupta; Kleomenis Kalogeropoulos; Spyridon E. Detsikas; George P. Petropoulos; Sartajvir Singh; Vishakha Sood
The analysis of crop variation and the ability to quantify it is a critical and challenging task. Remote sensing (RS) has proven to be an effective tool for monitoring crops and detecting seasonal variations worldwide. This opens new opportunities for developing effective crop monitoring models, with deep learning models showing great promise. This study presents a deep learning-based U-Net v5 Change Detection (UCD) model capable of identifying and monitoring the spatio-temporal variations in crop fields. The application of the model is demonstrated using Sentinel-2 imagery over Patiala district in India to monitor the seasonal crop variation (rabi crop) during 2017–2018. The results have shown that the UCD model has achieved better results (95.6–98.4%) in accuracy for classified maps and more than (91.6%–96.6%) in accuracy for change maps. This study will be useful for crop monitoring, precision agriculture and crop yield prediction and can assist in decision and policy making towards a more sustainable environment. © 2024
Retrieval of optical vegetation indices from SCATSAT-1 Ku-band backscatter: A comparative analysis with MODIS and Proba-V sensors
(Elsevier, 2025) Dileep Kumar Gupta; Ayushi Gupta; Ayuchtesh Dixit; Dhiraj Kumar Singh; Singh Sartajvir; Vishakha Sood; Naveen Kumar
The extraction of vegetation indices from radar backscatter measurements offers a solution to optical remote sensing, especially in cloud-dominated areas where optical satellite observations are poor. This research explores the capability of Scatsat-1 Ku-band backscatter observations for normalised difference vegetation index (NDVI) estimation from two optical satellite instruments namely Moderate Resolution Imaging Spectroradiometer (MODIS) and Proba-V. Total of 60 different study sites are considered for taking the observations of Scatsat-1 backscattering coefficients and NDVI from two different sensors. A nonlinear statistical model has been developed for the retrieval of NDVI (MODIS and Proba-V) using Scatsat-1 backscattering coefficients at horizontal transmit, horizontal receive (HH) and vertical transmit, vertical receive (VV) polarisation by least square optimisation techniques. The validation results demonstrate that VV-polarised backscatter yields better NDVI retrieval accuracy than HH polarisation, with a higher correlation (R = 0.796) and lower root mean square error (RMSE) (0.058) for MODIS NDVI compared to R = 0.579 and RMSE = 0.057 for Proba-V NDVI. The bias values are near zero, showing no strong systematic overestimation or underestimation of the retrieval models. Yet, retrieval errors are more evident in low NDVI situations, where vegetation sparsity adds variability to backscatter response. The research validates that Scatsat-1 Ku-band backscatter data can be utilised to estimate NDVI effectively, offering an alternative for vegetation monitoring when optical sensors are obscured by cloud cover. Future studies need to investigate multi-temporal data fusion, machine learning methods, and other vegetation indices to further improve the accuracy and reliability of radar-based NDVI retrieval models. © 2026 Elsevier Ltd. All rights reserved..
Snow Cover Response to Climatological Factors at the Beas River Basin of W. Himalayas from MODIS and ERA5 Datasets
(2023) Sunita; Pardeep Kumar Gupta; George P. Petropoulos; Hemendra Singh Gusain; Vishakha Sood; Dileep Kumar Gupta; Sartajvir Singh; Abhay Kumar Singh
Glaciers and snow are critical components of the hydrological cycle in the Himalayan region, and they play a vital role in river runoff. Therefore, it is crucial to monitor the glaciers and snow cover on a spatiotemporal basis to better understand the changes in their dynamics and their impact on river runoff. A significant amount of data is necessary to comprehend the dynamics of snow. Yet, the absence of weather stations in inaccessible locations and high elevation present multiple challenges for researchers through field surveys. However, the advancements made in remote sensing have become an effective tool for studying snow. In this article, the snow cover area (SCA) was analysed over the Beas River basin, Western Himalayas for the period 2003 to 2018. Moreover, its sensitivity towards temperature and precipitation was also analysed. To perform the analysis, two datasets, i.e., MODIS-based MOYDGL06 products for SCA estimation and the European Centre for Medium-Range Weather Forecasts (ECMWF) Atmospheric Reanalysis of the Global Climate (ERA5) for climate data were utilized. Results showed an average SCA of ~56% of its total area, with the highest annual SCA recorded in 2014 at ~61.84%. Conversely, the lowest annual SCA occurred in 2016, reaching ~49.2%. Notably, fluctuations in SCA are highly influenced by temperature, as evidenced by the strong connection between annual and seasonal SCA and temperature. The present study findings can have significant applications in fields such as water resource management, climate studies, and disaster management.
Spatiotemporal Vegetation Variability and Linkage with Snow-Hydroclimatic Factors in Western Himalaya Using Remote Sensing and Google Earth Engine (GEE)
(Multidisciplinary Digital Publishing Institute (MDPI), 2023) Dhiraj Kumar Singh; Kamal Kant Singh; George P. Petropoulos; Priestly Shan Boaz; Prince Jain; Sartajvir Singh; Dileep Kumar Gupta; Vishakha Sood
The mountain systems of the Himalayan regions are changing rapidly due to climatic change at a local and global scale. The Indian Western Himalaya ecosystem (between the tree line and the snow line) is an underappreciated component. Yet, knowledge of vegetation distribution, rates of change, and vegetation interactions with snow-hydroclimatic elements is lacking. The purpose of this study is to investigate the linkage between the spatiotemporal variability of vegetation (i.e., greenness and forest) and related snow-hydroclimatic parameters (i.e., snow cover, land surface temperature, Tropical Rainfall Measuring Mission (TRMM), and Evapotranspiration (ET)) in Himachal Pradesh (HP) Basins (i.e., Beas, Chandra, and Bhaga). Spatiotemporal variability in forest and grassland has been estimated from MODIS land cover product (MCD12Q1) using Google Earth Engine (GEE) for the last 19 years (2001–2019). A significant inter- and intra-annual variation in the forest, grassland, and snow-hydroclimatic factors have been observed during the data period in HP basins (i.e., Beas, Chandra, and Bhaga basin). The analysis demonstrates a significant decrease in the forest cover (214 ha/yr.) at the Beas basin; however, a significant increase in grassland cover is noted at the Beas basin (459 ha/yr.), Chandra (176.9 ha/yr.), and Bhaga basin (9.1 ha/yr.) during the data period. Spatiotemporal forest cover loss and gain in the Beas basin have been observed at ~7504 ha (6.6%) and 1819 ha (1.6%), respectively, from 2001 to 2019. However, loss and gain in grassland cover were observed in 3297 ha (2.9%) and 10,688 ha (9.4%) in the Beas basin, 1453 ha (0.59%) and 3941 ha (1.6%) in the Chandra basin, and 1185 ha (0.92%) and 773 ha (0.60%) in the Bhaga basin, respectively. Further, a strong negative correlation (r = −0.65) has been observed between forest cover and evapotranspiration (ET). However, a strong positive correlation (r = 0.99) has been recorded between grassland cover and ET as compared to other factors. The main outcome of this study in terms of spatiotemporal loss and gain in forest and grassland shows that in the Bhaga basin, very little gain and loss have been observed as compared to the Chandra and Beas basins. The present study findings may provide important aid in the protection and advancement of the knowledge gap of the natural environment and the management of water resources in the HP Basin and other high-mountain regions of the Himalayas. For the first time, this study provides a thorough examination of the spatiotemporal variability of forest and grassland and their interactions with snow-hydroclimatic factors using GEE for Western Himalaya. © 2023 by the authors.