Tuesday, October 29, 2019

Can we predict the yield before harvest?? Yes! We CAN, technology is here: Case study of Soybean in Dewas 2019.


Through Fasal Salah App, approximately 150,000 Lakh farmers are linked to us on one to one basis. These farmers are cultivating different crops. The sowing date, variety and agriculture management practices follows by them are received at the back end for analytical purposes. Based on these input from farmer’s yield determining factors like temperature, humidity, sunshine is crunched into a dynamic crop model specially developed for this purpose . We selected soybean as crop and Dewas of Madhya Pradesh as area of study. Dewas is situated between North latitude 22o 17’00” & 23o 20’00” and east longitude 22.962267, 76.050797 and covers 7,020km sq area .
Study Area

This study covered about 1200 farmers cultivated soybean who were connected on Fasal Salah App. Through this app farmers received weather basis advisory based on crop physiology of soybean being cultivated in real time. Farmers in farm were providing feedback of management practices being followed by them for application of farm input such as fertilizer, application of pesticides etc.

Besides, farmers also forward their crop pictures on weekly basis which helps us to monitor their crop more precisely. We ran artificial intelligence on the pictures received which fine tunes the crop yield determination by the dynamic model.  It is note worthy that the farmers field are
geo-located as also all the pictures received from them are geo located. This help us in correlating the ground recorded vegetation index with high resolution satellite data (In this case Sentinel-2A and Its NDVI are used). In this process we are able to expand the area of interest beyond the limits of farmers on Fasal Salah App.
Normalized Difference Vegetation Index (NDVI) of Dewas district were obtained. Collected GPS coordinates of soybean field were overlaid over calculated NDVI and NDVI values of each field were recorded as training sets. The refined training set data statistics were used for performing maximum likelihood supervised classification based on recorded vegetation data.


With this methodology we have estimated area under soybean in Dewas district for year 2019, to be 3.8 Lakh hectare. The average yield determined by the hybrid of crop model and crop pictures for Dewas district based on input from farmers using our Fasal Salah App came to 813 kg  per hectare thus total yield of soybean in Dewas district expected to be 3,078,000 Quintal in Kharif 2019.

While actual field estimates are yet to come this figure will stand the test of time when official results are posted. 

I would like to remind that this year Dewas experienced some excessive rains. Our calculations take into account such damages because of excessive rains. However what is not accounted for is the damage because of floods.

The innovation  which made our team win The Agriculture Grand Challenge 2019 and was awarded best App for Agriculture by Indian Mobile Congress 2019 ( all Mobile operators of India  form this congress)  is ready to be used  by the  Industry for actionable market intelligence for commodity trading. 

Come join our innovation

bk@weahersysbkc.com

1 comment: