SAR signature investigation of rice crop using RADARSAT data

Abstract

The current study investigates the potential of multi‐temporal RADARSAT ScanSAR Narrow Beam B (SCNB) data to monitor rice crop growth and condition where special emphasis was given to the signature analysis of the crop. The study area is located in the Baleshwar and Bhadrak districts of Orissa. The temporal variations of radar backscatter of all land‐cover classes were analysed as a function of time. The analysis of the Synthetic Aperture Radar (SAR) backscatter coefficient (σ⁰) of rice crop shows significant temporal behaviour and a large dynamic range during its growth period, which is due to the interaction of microwave radiation with the crop canopy, increasing from the transplanting stage to the reproductive stage. This temporal variation of SAR backscatter clearly differentiates rice fields from other land‐cover classes.

Separability studies among different class pairs carried out using t‐test and Bhattacharya distance show that all the rice classes are separable from each other except the early rice, which was mixed with far sides of hills and shadow. Knowledge‐based decision rule classifier based on the temporal evaluation of SAR backscatter was attempted to classify rice and non‐rice areas, and achieved more than 98% accuracy in the case of rice class. The results are promising and confirm the possibility of operational use of RADARSAT data for rice crop growth monitoring.

Acknowledgements

We would like to place on record a deep sense of gratitude to Shri J. S. Parihar, Group Director, ARG/RESIPA for his keen interest and constant encouragement during the completion of this work. The authors are grateful to the scientists of ORSAC for extending their support for the collection of ground data. The authors are also grateful to all the colleagues and friends for the moral boost and support while carrying out this work successfully.