Sources of Variation in Assessing Canopy Reflectance of Processing Tomato by Means of Multispectral Radiometry

• Main Authors: Giorgio Gianquinto, Francesco Orsini, Giuseppina Pennisi, Stefano Bona
• Format: Article
• Language: English
• Published: MDPI AG 2019-10-01
• Series: Sensors
• Subjects: passive sensor; sampling methodology; bootstrap analysis; vegetation indices; green ratio of vegetation index; nir/560 or gvi (green vegetation index); green normalized difference vegetation index; normalized difference vegetation index on greenness (g-ndvi); soil mulching; solanum lycopersicum l
• Online Access: https://www.mdpi.com/1424-8220/19/21/4730

Canopy reflectance sensors are a viable technology to optimize the fertilization management of crops. In this research, canopy reflectance was measured through a passive sensor to evaluate the effects of either crop features (N fertilization, soil mulching, appearance of red fruits, and cultivars) or sampling methods (sampling size, sensor position, and hour of sampling) on the reliability of vegetation indices (VIs). Sixteen VIs were derived, including seven simple wavelength reflectance ratios (NIR/R460, NIR/R510, NIR/R560, NIR/R610, NIR/R660, NIR/R710, NIR/R760), seven normalized indices (NDVI, G-NDVI, MCARISAVI, OSAVI, TSAVI, TCARI), and two combined indices (TCARI/OSAVI; MCARI/OSAVI). NIR/560 and G-NDVI (Normalized Difference Vegetation Index on Greenness) were the most reliable in discriminating among fertilization rates, with results unaffected by the appearance of maturing fruits, and the most stable in response to different cultivars. Black mulching film did not affect NIR/560 and G-NDVI behavior at the beginning of the growing season, when the crop is more responsive to N management. Due to a moderate variability of NIR/560 and G-NDVI, a small sample size (5−10 observations) is sufficient to obtain reliable measurements. Performing the measurements at 11:00 and 14:00 and maintaining a greater distance (1.8 m) between plants and instrument enhanced measurement consistency. Accordingly, NIR/560 and G-NDVI resulted in the most reliable VIs.