Analysis of the Thermal Temporal Variation in the Citrus Canopy

This paper presents a study of thermal imaging as a potential approach for improving citrus fruit detection. Citrus tree canopy was monitored on 24 hour cycles using a thermal infrared camera. Four different trees (region of interest) were observed on four separate days. Surface temperature of the fruit, ambient temperature and relative humidity were measured using a portable dew point meter. The acquired thermal images were corrected for fruit emissivity, which was determined to be 0.90, using the ambient temperature, the relative humidity and the reflected temperature. Fruit and canopy temperature profile demonstrated a relatively large temperature gradient, which occurred from afternoon until midnight, which was evident from the acquired thermal images of the canopy. The fruits were successfully segmented in the thermal images using image processing techniques during the time of the largest temperature difference, which suggests potential application of thermal imaging for citrus fruit detection.