Radiometric Performance Of Multispectral Camera Applied To Operational Precision Agriculture

The use of multispectral sensors onboard of unmanned vehicles is nowadays a common practice in precision agriculture. Their radiometric behavior is an important issue to ensure comparable information along space and time. A calibration of a Sequoia (R) camera is conducted in terms of reflectance and NDVI, taking benefit from a wheat cover field work. The output reflectance values in the green and red channel show linear behavior in the range of 0.00 to 0.45, with a saturation over this value. The red edge and NIR channels show linear behavior in the whole reflectance range, with better performance in the NIR than in the red-edge channel. The NDVI has a good agreement with field measurements, near the 1:1 line, showing the highest variability for the lower values due to heterogeneity of soil brightness linked to variations in soil moisture.