NDVI prediction using Machine learning after Geofencing on satellite data of Sugarcane crop

Abstract

The manual fencing of the crop is too tedious and time-consuming for the farmers. The farmers have to physically see that they are not crossing the boundaries and usually, farmers fight related to the plots. In this paper focus is on the satellite data imagery of Satara district and if geofencing is done for sugarcane crops then monitoring the crop is comparatively easy and analysis can be further done related to sugarcane plots. Geofencing helps then to have the precise data of the farmer and boundaries can be detected to avoid hassles. The Cane plots that have been geofenced can then be labeled according to the plots of the farmers, thus streamlining th classification process. The digitization of the farmer plots is used in this paper and converted into shape files by using GIS which helps to do the further analysis of the crop. Further after doing the pre-processing by using GIS, the Normalized Difference Vegetation Index (NDVI) is predicted prior using Machine learning technique in python. The NDVI actual and predicted for one life cycle of sugarcane is shown in the paper. The NDVI predicted values can be helpful for sustainable agriculture of sugarcane crop in terms of disease detection, cane classification and prediction also. The Machine learning   algorithms can be applied further and the geofencing can be done district wise in future scope along with the vegetation   indices.