Please use this identifier to cite or link to this item: http://dspace.iitrpr.ac.in:8080/xmlui/handle/123456789/3416
Full metadata record
DC FieldValueLanguage
dc.contributor.authorTripathi, A.-
dc.date.accessioned2022-05-24T10:03:32Z-
dc.date.available2022-05-24T10:03:32Z-
dc.date.issued2022-05-24-
dc.identifier.urihttp://localhost:8080/xmlui/handle/123456789/3416-
dc.description.abstractThe inception of remote sensing on the global technological horizon has eased out the conventional field and laboratory-based time-taking soil health assessment techniques. The utilisation of Machine Learning (ML) in remote sensing research has truly revolutionised the modelling and estimation of crop yield and soil health parameters. Over the past many decades, most of the remote sensing research directed at soil health was done alongside crop monitoring or precision agriculture-related studies. Therefore, there is still a paucity of dedicated research using multi-sensor remotely sensed data for soil health studies. Optical-multispectral remotely sensed data had been widely used for the different soil type mapping and generation of various soil health-related indices worldwide. Still, this data is highly weather dependent and not readily available for days of having a cloud over the area under study. Remotely sensed SAR data, on the other hand being an active sensor, has penetration ability and all-weather, all-time data acquisition capability, but its soil healthrelated indices are highly data and target-dependent and are still under research. The field gathered soil health-related parameters like moisture, salinity and Soil Organic Carbon have been much relied upon for ages for soil-related research. However, this is a highly labour-intensive exercise and requires costly apparatus and skilled manpower to be deployed. This study aims to utilise both remotely sensed Microwave satellite data from Sentinel-1 and Optical data from Sentinel-2, and field data to estimate three important soil health parameters- Soil Moisture, Soil Salinity, and Soil Organic Carbon (SOC). After that, the soil health parameters, SAR backscatter and field data were used to estimate wheat crop yield. The study uses SAR backscatter of VV and VH polarisations from Sentinel-1, Normalised Differential Moisture Index (NDMI) and Normalised Differential Salinity Index (NDSI) from Sentinel-2, Volumetric Moisture Content (VMC), Electrical Conductivity (EC), and Temperature from the field, with SOC and pH from a laboratory test of soil samples. The data were used for the estimation of soil health parameters using various machine learning techniques. The neural network model performed best in crop yield estimation and gave R2 values of 0.712 and 0.635 in training and testing phases, respectively and Mean Absolute Error (MAE) of 0.32 and Root Mean Square Error (RMSE) value of 0.54. The novel feature of this study is that since Sentinel-1 SAR data operates in C-band with less penetration in vegetation, the study estimates soil health parameters for early stages of wheat crop growth when soil lies mostly exposed and utilises the estimates for crop yield prediction. This study has importance for early mitigation and monitoring of soil health when there is no availability of longer wavelength SAR data.en_US
dc.language.isoen_USen_US
dc.subjectRemote Sensingen_US
dc.subjectMachine Learningen_US
dc.subjectSoil health parametersen_US
dc.subjectCrop Yield Estimationen_US
dc.titleSynergistic application of synthetic aperture Radar (SAR) and optical multispectral satellite data for agroinformatics of Rupnagar, Punjab, Indiaen_US
dc.typeThesisen_US
Appears in Collections:Year-2021

Files in This Item:
File Description SizeFormat 
Full Text.pdf8.36 MBAdobe PDFView/Open    Request a copy


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.