Vine Vigour Characterization using a Multi-temporal X-band SAR Very High Resolution dataset

In the vine-growing industry, the regular acquisition of image data throughout the vine-growing season is of great interest for accurate plot monitoring and grape yield prediction. In this context SAR backscattering data, which can be acquired at any time without being affected by clouds or night-time, seems particularly suitable, especially as their spatial resolution has significantly increased over the past few years. In this paper, we investigate the ability of X-band very high resolution (1 meter) SAR backscattering signal to characterize vine-plots. For that purpose, we used a multi-temporal dataset in order to study three vine-related factors : vine variety, ranks orientation and vine vigour. In our case, we want to monitor a known area for which vine variety and ranks orientation are well defined. As a consequence, we decided to focus on the most useful factor for vine-growers, i.e. vine vigour. The methodology used is a whole sensitivity analysis over a vine vigour classification and regularization process by using only the amplitude of the SAR backscattering.