Improving wheat yield predictions with crop image technology
Novel applications developed by researchers at BioSense Institute in Serbia are dedicated to make deep learning technology a widely accepted practice in agriculture, providing small and big farm holders to benefit from precision farming technology.
BioSense, the Serbian Research and Development Institute for Information Technologies in Biosystems, is a multidisciplinary research institute for agriculture of the future. The wheat yield prediction research conducted in Serbia aims to increase the collection of farm management data, help farmers understand more about their farm business by using sensor technology and IoT applications, and reduce farm labour.
Wheat yield experiments
Wheat is one of the most important crop types in food production worldwide. Due to increasing food demand and rising population, it is necessary to boost production and supplies of wheat and other cereals.
In 2019, BioSense Institute, observed wheat in different experimental field stages and did this for three consecutive seasons. Cameras used during the experiment were the FLIR SC620 in season one and two, and a thermal camera in the third season. By taking pictures of the wheat growing in their field (four weeks before harvest), and uploading it through a mobile application, farmers were able to gain information about the wheat yield estimate based on the current state of growth.
The objective of this research is to enable the farmer to use imagery to detect at an earlier stage when estimated yields are below average and timely apply agronomic treatments to improve yield.
Farm efficiency with data management
The automatization of ear density calculation (number of ears per unit ground area, usually 1m2), which is one of the main agronomic yield components in determining grain yield in wheat, can provide fast evaluation of this attribute and potentially save 200 hours of manual work, ease monitoring, and increase crop management practice efficiency. This will save money from potential yield reduction, which can cause big losses in the farmers’ investments.
The currently used process of yield prediction includes manual and tedious work. The farmer takes samples from the area of 1m2 from the field (if the field is larger, then from a few locations within a field), and measures the biomass. The next step is to separate and count the ears of wheat manually. Since the counting of one sample requires up to 1 hour, while the number of samples can easily exceed 200, this can result in more than 200 working hours, or two to three weeks of manual labour that could be avoided.
The collected dataset comprises RGB and thermal images. Thermal images give us information about the difference in temperature between the ears and their background through their colouring and ease ear detection. Images were taken in four dates on two locations in two stages of wheat growth.
Power of deep learning
Since we have witnessed a huge breakthrough of neural networks, especially in image processing, deep learning has greatly outperformed classical models and algorithms. The nature of deep learning is that the addition of more data improves the quality of results, so by uploading images from farmers (crowd sourcing), the initial database will be expanded, so the algorithm will achieve better and more accurate results.
For more information about the methodologies used in this research by BioSense Institute, visit the DRAGON website.
Data-driven precision agriculture by DRAGON
Agri-EPI Centre is a core partner within the data-driven agriculture services and skill acquisition project DRAGON. The aim of the project is to enable communication skill transfer and knowledge exchange between research organisations and end users through big data and effective data analytics.
This article is an extract from an article of Željana Grbović – Junior Researcher, BioSense Institute – published on www.datadragon.eu.