Assessing accuracy over warm-up time of Lepton 3.5 thermal imaging for measuring leaf temperature of crops

Abstract

Accurately monitoring leaf temperatures is becoming more and more critical as more studies use a crop's leaf temperature for irrigation, disease, and pest detection. In the present study, the accuracy of a module for camera warm-up times of 30 sec, 1 min, 2 min, 3 min, 3.5 min, and 4 min was compared when leaf temperature was periodically measured using a Lepton 3.5 module, a low-cost thermal imaging camera. The experiment was conducted for a plant in a laboratory, and a high-accuracy LT-1T thermistor sensor was used together to compare module accuracy. The power consumption of both sensors decreased by using the sleep mode of the module, and all measurements were conducted in intervals of five minutes. The accuracy was compared using the R-Squared, MAE, and RMSE of the two values measured by LT-1T and Lepton 3.5. As a result of the experiment, the accuracy was the highest when the warm-up time was 3.5 minutes showing the mean absolute error (MAE) and the root means squared error (RMSE) values of 0.56 ? and 0.59 ?, respectively. The accuracy was 2.5%, which was substantially higher than the 5% accuracy of the device specification.