Nitrogen and chlorophyll status in romaine lettuce using spectral indices from RGB digital images

Destructive methods for crop nutrition estimation are accurate and standardized but costly and limited by spatial scale. Non-destructive techniques such as the use of digital imaging provide fast and reliable results in situ; however, there is limited information on these non-destructive techniques in leafy vegetable crops. During the study it was estimated the concentration of nitrogen (N) and chlorophyll (Chl) in romaine lettuce using spectral indices derived from the RGB (red, green and blue) digital images. The lettuce crop was grown in plastic tunnels and irrigated with five N levels: 0, 4, 8, 12, and 16 mEq L-1 NO3-N (corresponding to 0, 248, 496 and 744 mg L-1, respectively) based on a modified Steiner solution. The treatment of 16 mEq L-1 NO3-N showed the highest growth at 42 days after transplanting (DAT). Digital images of the plants were acquired weekly with a RGB camera and processed to obtain scale mosaics and twelve spectral vegetation indices. Correlation analysis of the spectral indices indicated that normalized green-red difference index (NGRDI), excess green (ExG) and green channel (g) indices showed a positive linear correlation with the concentration of N (r > 0.93) and Chl (r > 0.93). Besides, an exponential correlation with leaf area (r > 0.86) was founded, which was stronger in the last 21 DAT due to the acceleration in leaf growth during the vegetative stage of the crop. These results show that RGB digital images are a low cost, non-destructive, reliable and accurate method to estimate N and Chl concentration and leaf area in romaine lettuce during production. Therefore, this technique could be an affordable alternative that combined with portable meters (i.e. SPAD) provides real-time monitoring of the nutritional status of the lettuce crop, especially in crop factories.