LAUSR

Abstract Apple sunburn is a major challenge for the apple growers and so is its prediction for taking timely remedial measures. In this study, a thermal-RGB imager was developed and evaluated to estimate the real-time fruit surface temperature (FST) of an apple. The thermal-RGB images were captured for two apple varieties namely Golden Delicious and Red Delicious. Custom image processing algorithms were developed to estimate the FST from captured data. Two sets of experiments (1 and 2) were conducted on day-1 and -2, respectively. In experiment-1, the thermal-RGB imager estimated a maximum FST of 42.24 °C for Golden Delicious and 43.85 °C for the Red Delicious cultivar. Similarly, in experiment-2, the thermal-RGB imager estimated a maximum FST of 40.51 °C for Golden Delicious and 43.81 °C for the Red Delicious cultivar. The overall standard error for FST estimation from thermal-RGB imager ranged within 0.12–0.23 °C. The FST estimates from thermal-RGB imager were compared with the FST estimates from high resolution (pixel size = 0.17 μm) thermal imager (coefficient of determination, R2 = 0.80–0.90, root mean square error, RMSE = 0.50–1.08 °C and Pearson correlation coefficient, r = 0.90–0.95) and micro-climate sensor data (R2 = 0.87–0.90, RMSE = 0.98–1.97 °C and r = 0.92–0.99). Overall, the thermal-RGB imager was found to be an accurate estimator of apple FST (standard error = 0.12–0.23 °C). The thermal-RGB imager was advantageous in apple segmentation based on RGB imaging integrated with thermal imaging. The thermal imaging principle of capturing infrared radiations, irrespective of the unsteady weather parameters accounted for all types of heat flows occurring on the apple surface. Imager estimated the critical conditions for apple sunburn with numerous advantages over other techniques and can further ease out the formation of decision support system for apple sunburn management.