Multispectral cameras are commonly found among Agriculture & Forestry applications of drones. The role of a multispectral sensor is to measure for a plant, the percentage of reflected light at different wavelengths, both in the visible and non-visible parts of the light spectrum that are of interest for vegetation. Typically plants reflect green light and absorb red and blue. But plants also reflect light in non-visible spectrum, in the Near-Infrared (NIR).
This remote sensoring method allows to access to valuable information about plant health, its evolution in relationship to its environment : soil, sunlight, humidity, reaction to agricultural inputs such as fertilizers, among other factors affecting plant life.
Let us have a tour on how a multispectral camera works.
Light absorption and reflection by plants actually differs according to the light wavelength. Each color in the visible spectrum, for instance, corresponds to a certain wavelength.
A multipectral camera uses as many single sensors as the numbers of wavelengths to monitor. For instance a Micasense RedEdge is made of 5 sensors each fitted with a specific optical filter to capture light in the targetted wavelength. Combining the inputs of each sensor allows to compute the vegetation indices that are key to evaluate plant and crops condition; collected images pixels values actually account for the percentage of reflectance at the specific wavelength.
The list of delair.ai currently supported multispectral cameras can be found here.