Smart One-Chip NDVI Camera for Plant Detection

Volker Dworak, Joern Selbeck, and Karl-Heinz Dammer


smart camera, NDVI, image processing, plant sensor


The use of problem adapted (smart) cameras for automation and mapping in agricultural applications can be beneficial because their use could conserve fertilizer, pesticides, and machine time. Protecting the environment is one of the major scientific challenges in agriculture research. Smart cameras for agricultural applications are typically plant cameras, which use the NDVI (normalized differential vegetation index) to discriminate plant information from the soil background. The highest difference in the reflectance is found between the NIR (near infrared) and the red channel optical frequency band. Traditional NDVI cameras with two CCD (charge coupled device) chips for the red and NIR channel are expensive and require a precise alignment of the two sensors. The advantage of the presented smart camera approach is the use of a one-chip, low-cost sensor to design a NDVI camera. In this study, a reassembly of the optical filter design was performed to enable the NIR-sensitivity and disable the usual blue and green optical sensitivity. The new color pixel response was then used to calculate the NDVI with respect to the Bayer pattern and the different color intensity content in the pixels. “Blue” and “green” corresponded to the pure NIR intensity, and the “red” represented the sum of red and NIR intensities.

Important Links:

Go Back