What we do
Derive satellite-based information for crop-specific monitoring
The process
GeoWatch Labs convolutes multiple layers of information, comprising of satellite imageries, remote sensing climatic variables, crop-specific characteristics and local farming practices to take into account all parameters influencing yields.
Collection of data
Satellite imageries and remotely-sensed climatic data are collected for the studied area, ensuring the scalability of the monitoring system as well as spatial and temporal homogeneity regarding input data.
Collection of satellite images
Before (left) and after (right) the extraction of agricultural areas in Mauritania in 2019.
Crop type classification
A first step is to identify agricultural fields in order to filter satellite signals only to cultivated lands. Crop classification comes next in order to adapt growth modelling to the plants' characteristics.
Crop growth modelling
Crop growth modelling requires various layers of information to be taken into account to reproduce the local environment of the plant. Climatic data and satellite imageries are combined based on agronomic considerations to monitor the plant growth on a daily basis.
Yield forecast and final estimation
Monitoring the growth of the plants enables GeoWatch Labs to forecast the expected yield learning from the local historical agricultural seasons. Final estimations of yields are rendered once the harvest has been detected.
Case studies
Burkina Faso – Hauts-Bassins Region: rice modelling
1. Extraction of agricultural areas and crop-type classification
2. Crop-specific modelling and final yield estimation
Yield formation evolution between July and October 2019 in the area considered – snapshot every two weeks.
Extraction of agricultural areas (bottom, rice fields in blue, cowpeas fields in yellow).
3. Final estimations of rice yield
(mean average in the area considered in 2019):
5.5 t/ha