ASSESSMENT OF CROP ROTATION USING MULTI POLARIZED SENTINEL-1 TEMPORAL SAR DATA OVER PARTS OF KUMAUN REGION OF UTTARAKHAND

5
1 rating - Please login to submit your rating.

Abstract 

Accurate information about crop rotation is essential for administrators, managers and various government departments for assessment, monitoring, and management of various resources for crop escalation. Radar remote sensing, because of its all-weather capability and assured uninterrupted data supply can show a substantial part in the evaluation of crop rotation. Twenty-Nine (29) Level-1 Ground Range Detected (GRD) Sentinel-1 C band scenes acquired in the Interferometric Wide (IW) swath mode during the period from 11th November 2017 to 30th November 2018 over Kumaun region, Uttarakhand were utilized to explain the feasibility of multi-temporal SAR data for crop rotation. Principal Component Analysis was adopted to demarcate different crop rotation trends throughout the study area. The results show that ‘wheat to paddy’ is the highest practiced crop rotation with 27.95% coverage, whereas ‘mustard to sugarcane’ is the lowest practiced crop rotation with 2.86% coverage of the entire area of study provided with 80% of an overall accuracy acquired from the ground-truth data from IIRS/ISRO. It is a benefit for the decision-makers, administrators and policymakers to take reliable decisions about a particular region regarding land management and natural resources.

Instructions: 

The B. Tech final project was of 4-months duration from December 18th, 2018 to April 25th, 2019. The present study aimed to investigate the potential of sentinel-1 SAR data for the estimation and monitoring of crop pattern of the monsoon season and winter season crops using large temporal data set over the Kumaun region, Uttarakhand. The crop rotation is an indicator of various socio-economic trends of a given area. It also shows the penetration of line departments in rural areas of these districts along with increasing or decreasing patterns/ trends of water resources and soil health.
The proposed research work dealt with the development of a methodology for the viability of Synthetic Aperture Radar (SAR) data used for crop assessment using freely available multi-polarized SAR data so that the line departments can utilize the proposed methodology without large budget requirement. Thus, it is essential for administrators, managers and various government departments for assessment, monitoring, and management of various resources for crop intensification.
Considering the sensitivity of Synthetic Aperture Radar towards frequency, polarization, and the surface roughness, surface moisture of various crops; the Principal Component Analysis (PCA) was adopted to demarcate different crop rotation trends in the study area. This successfully demonstrated the feasibility of the use of SAR data in crop rotation analysis.
Radar Remote Sensing is relatively a new branch of Remote Sensing and therefore use of SAR data for many vital agricultural applications is under-explored and under-reported. The use of radar remote sensing has been demonstrated over parts of the Kumaun region where no such detailed study has been carried so far.
My guide appreciated me for my holistic contribution because I was the candidate who has individually carried out the research work. My overall contribution to the project includes in the following steps:
1. 29 Level-1 GRD Sentinel-1 C band scenes were acquired in the IW swath mode from the European space agency's Copernicus programme which has freely available dataset during the period from 11th November 2017 to 30th November 2018.
2. Pre-processing was conducted using the Sentinel-1 toolbox (S1TBX) in the SNAP software provided by ESA. The steps include Radiometric calibration, Terrain correction, and Speckle filtering.
3. After the completion of the pre-processing stages, the temporal data were stacked for taking subsets of VV and VH individually for analysis. The images were stacked using the nearest neighborhood resampling method, minimum output extent was selected for co-registration.
4. PCA was used to demarcate different trends/ patterns of crop rotation in the rabi, Kharif and whole season images of SAR.
5. In the project, the unsupervised classification was carried out.
6. Finally, the crop rotation map had been generated to find out various crop rotation classes pertaining to the study area. The crop accuracy map had also been generated to access the accuracy of various crop rotation classes of the study area.