Hyperspectral imaging is one of the most advanced remote sensing technologies, renowned for its ability to capture light reflected across hundreds of narrow spectral bands. A distinctive feature of hyperspectral systems is their capability to generate a complete reflectance spectrum for every pixel in an image. This unique characteristic makes these systems highly valuable for assessing the condition of plants and conducting detailed, multidimensional analyses of their physiological and biochemical properties - both at the whole-plant level and for specific organs or localised regions.

In recent years, the use of hyperspectral imaging has expanded significantly worldwide, particularly in agricultural and environmental research. It has become a powerful tool in agroecology and crop science, with applications ranging from detecting nutrient imbalances and water stress to diagnosing plant diseases, tracking the development of invasive species, and predicting crop yields.

Due to their high spatial and spectral resolution, hyperspectral cameras are capable of capturing detailed spectral signatures of plant materials (often referred to as “fingerprints”). These rich datasets are ideal for training machine learning algorithms aimed at assessing plant condition, and they can also be used to validate remote sensing results obtained through platforms like UAVs.

Successful application of machine learning for crop disease detection hinges on the quality of the input data. The accuracy of hyperspectral measurements depends heavily on correct sample handling, precise equipment calibration, and careful image acquisition. Factors such as lighting conditions, focus, and scan settings must be optimised to ensure high-quality, reliable results.

This training guide offers a hands-on introduction to the use of hyperspectral imaging in a controlled laboratory environment. It includes concise instructions for preparing plant samples and acquiring hyperspectral data using the FigSpec FS-13 camera. Additionally, it provides a step-by-step walkthrough for processing and analysing the collected data using Breeze software by Prediktera.

The methods presented for disease diagnosis are based on direct experience working with the FS-13 model of hyperspectral camera in the lab, as well as analysis of the resulting multispectral datasets using the Breeze platform.

This guide is funded by the Science Committee of the Ministry of Science and Higher Education of the Republic of Kazakhstan (Grant No. AP23485162) and was prepared within the framework of the project “Development of an innovative method for monitoring and early diagnosis of grain crop diseases using hyperspectral sensing technology”.