Drones equipped with cameras sensitive to near-infrared wavelengths are increasingly being used in environmental assessment studies and in agriculture. These cameras are sensitive to vegetation cover, extent of eutrophication in water bodies, and aspects of crops such as growth vigour, biomass and potential yield. Single-sensor (‘RGB’) cameras with modified spectral filters that allow for capturing near-infrared wavelengths offer a low-cost alternative to multi-sensor multispectral cameras or spectrometers. However, some studies point to lower measurement accuracies by such infrared converted sensors. So, to what extent can infrared converted cameras be used to quantify vegetation condition? This case study compared vegetation indices calculated from infrared converted camera imagery to those measured by a drone-borne multispectral camera and a handheld NDVI meter, as captured over soybean and potato fields. The study found that infrared converted camera derived NDVI was consistently lower over vegetation than NDVI measured from the multispectral and handheld sensors. Further, all considered indices of the infrared converted camera displayed relative underestimation at high index values, but over estimation at low index values. The study builds on previous case studies with similar results by further evaluating the reflectance patterns of the individual image bands to find possible reasons for the discrepancy in vegetation index measurements. There is good agreement between the near-infrared bands of the respective sensors (r=0.87), but the respective red bands have weak correlation (r=−0.03). We discuss possible reasons for the lower vegetation index measurements observed by the infrared converted camera, noting broad band sensitivities, and differing central wavelengths, which may have caused overestimated reflectance in the red band. All processing and analysis were executed with open-source software, and source code is made available to support reproducible research.
Read full abstract