Agricultural Crop Yield Prediction Using Artificial Intelligence and Satellite Imagery

Abstract

The influence of climate change and its unpredictability, has caused majority of
the agricultural crops to be affected in terms of their production and maintenance.
Forecasting or predicting the crop yield well ahead of its harvest time would assist the
strategists and farmers for taking suitable measures for selling and storage. Accurate
prediction of crop development stages plays an important role in crop production
management. Such predictions will also support the allied industries for strategizing the
logistics of their business. Several means and approaches of predicting and demonstrating
crop yields have been developed earlier with changing rate of success, as these don’t take
into considerations the weather and its characteristics and are mostly empirical. For this a
combined constructional and methodological approach is proposed like variety inception,
pesticide & fertilizer management, integrated cropping, rainwater harvesting, efficient
irrigation techniques etc. would also be needed. The neural network algorithm is less prone
to error than other machine learning and data mining techniques, making it an effective
machine learning tool for predicting crop yields. The ANN back propagation algorithm is
used to determine the appropriate weight value to calculate the error derivative. The
accuracy of the crop yield estimation for the diverse crops involved in strategizing and
planning is deliberated to be one of the utmost significant issues for agronomic production
purposes. The yield prediction is still considered to be a major issue that remains to be
explained based on available data for some agricultural areas. Crop monitoring and
forecasting of crop yields for the proposed system will be carried out via satellite images
with low resolution. The combination of extensive and extended topographical coverage
and its high temporal frequency make these images an appropriate choice for the
prediction of crop yields. To ease the training, the dimensionality of the data is reduced by supposing that the position of pixels doesn’t influence the typical crop yield. The prototype
distinguishes between crops, the infrared and temperature bands of images taken during
apex growing season contribute the most to the crop prediction.
Using Satellite Imagery and CNN algorithm to forecast crops in all states produces better
efficiency than only using ANN algorithm. The main aim is to compare the output of ANN
and CNN to verify whether the results are accurate for crop yield forecasting.