Skip to content

meet-sapu/Crop-Yield-Prediction-Using-Satellite-Imagery

Repository files navigation

Yield-Prediction-Using-Sentinel-Data

The motive here is to predict the yield of crops of a particular farm by the change in pixels of the image of farm yearly . The satellite data used is of sentinel, sentinel is a launched by European Space Agency and the data is open source.

  1. Sentinel Image Products Scraping - The script in this folder scraps a patch of land containing our farm of interst, this scripts automatically scraps all the images through out the year.

  2. Preprocessing and Cropping - The script in this folder will crop the farm which we need to analyze according to coordinates provided by the user and also does feature engineering on the cropped images.

  3. Tensorflow Model - The script in this folder has the algorithm , it is still under testing ang the algorithm is somewhat similar to algorithm from this paper: https://cs.stanford.edu/~ermon/group/website/papers/jiaxuan_AAAI17.pdf.

In the above mentioned paper the data used is of NASA which is of low resolution as compared to that of Sentinel.