Commercial potato farmers anywhere in the world are familiar with the fact that the potato is susceptible to wide range of diseases. Two of the most common and dangerous of these are late blight (Phytophthora infestans) and early blight (Alternaria solani).
An untold number of potato farmers around the world apply chemicals – often very frequently – to their potato crops during the growing season as part of their management strategies to combat the potential damage that can be caused by these two diseases.
Until recently, broad spectrum fungicides were used to protect potato crops against both early blight and late blight. With the increase of public concern about the negative impact of agrochemicals on the environment, the shift from broad to small spectrum fungicides allowed the advance of early blight in potato cultivation.
This increasing infection pressure has an additional economic cost in terms of excessive man-hours as well as expensive chemicals needed for regular applications. Any reduction in the usage of manpower and chemicals during potato production will yield major economic and environmental benefits for all players involved in the potato production chain.
By putting precision agriculture methodologies into practice, this increasing cost burden can be reduced, since less manpower and less chemicals are required when these chemicals are applied in a very localized and precise way.
The potential of hyperspectral imaging for disease detection
Plenty of research had in the past been conducted already in many countries around the world to evaluate the potential of hyperspectral imaging for disease detection.
Most of these studies have concluded that hyperspectral imaging (using +100 wavelengths or ‘colors’) is indeed a very practical and most certainly a useful method for characterizing plant health status, as well as indicating the presence and progression of disease infection in a crop.
Hyperspectral imaging can without a doubt be adapted and applied appropriately to enable early detection of disease presence in potato crops.
This then was the premise of a team of researchers at the Flanders Research Institute for Agriculture, Fisheries and Food (ILVO) in Belgium who designed a project and put it into practice to test the value of hyperspectral imaging as it applies to the detection of early blight in a cultivated potato crop.
Putting hyperspectral imaging into practice: Building a “hypercart”
In order to acquire high quality hyperspectral images with a resolution of about 0.3mm/pixel, a motor-driven, controlled and steered “hypercart” was built by the research team. This high-tech cart was then used in a potato field where the research team wanted to detect the presence of early blight in the crop.
The hypercart was constructed from an aluminum frame with a width of 2.25m, a height of 2.3m and a length of 3m. Different sensors were mounted on a moving beam in order to scan crops in a plot of 0.85m by 3m.
The hypercart was equipped with five sensors to gather RGB, multispectral, hyperspectral and height information about the canopy i.e.
(1) a multispectral camera (Sequoia, Parrot),
(2) a hyperspectral snapshot mosaic camera (made by 3D-one and based on an IMEC CMOS-chip, 41 bands in the wavelength range 470-975 nm) and
(3) a hyperspectral linescan sensor (Imspector V9, 430-900 nm)
During field measurements, the hypercart was covered with a black cloth to eliminate influence of external light. A combination of 18 halogen spot lights mounted on the moving beam produced artificial light conditions in order to minimize the impact of variable lighting conditions.
The hypercart can be operated completely autonomously thanks to the use of two lithium-ion accu packs (each 5kWh) and a highly efficient power management system. Propulsion and the steering are realized with two brushless DC motors. Communication between user, sensors on the moving beam, and the steering and propulsion system is made possible by an integrated shuttle pc.
The data collected with the hypercart was analyzed in python by extracting color information (spectra) for healthy leaves and infected tissue. This allowed the research team to identify the wavelengths (colors) which are highly altered by an Alternaria infection.
The main regions (colors) of interest were identified as the NIR-region (720 – 900nm), the red region (around 680nm) and the green region (around 550nm).
The alterations in the NIR-region is caused by the destruction of cell structure in the leaves by the pathogen. Changes in the red region can be explained by the destruction of chlorophyll, a complex pigment (responsible for the green color of plants) that allows plants the absorb the energy of the light, and this can hence be described as the “motor” driving plant growth.
The last change (in the green region) is related to pigments produced by the plants when stressed. These pigments cause the final yellowing of the leaves.
Research findings and future UAV applications
From these experiments the research team learned that both a high resolution of <0.3mm/pixel (!) and information in the NIR region are indeed highly valuable.
This year, close-range information gathered will be translated into data useful for UAV application. The researchers will use a high-resolution consumer camera as well as multispectral cameras for the purpose of this kind of application. This will allow farmers who apply UAV’s to obtain an accurate overview of the Alternaria infections in scouted fields, and it will undoubtedly assist in preventative management decisions by farmers to combat early blight infection in their potato crops.
This article was first published in the February issue of the Global Potato News magazine