Feasibility study on Lepton 3.5 in terms of accuracy for measuring leaf temperature of crops

Abstract

The precise monitoring of leaf temperature is becoming more important as crop leaf temperature is utilized more frequently for different uses such as irrigation, disease and pest detection. This study aims to explore the potential usage of a cost-effective Lepton 3.5 camera to measure the crop canopy temperature. The accuracy of the Lepton 3.5 will be compared to a FLIR E8-XT thermal camera and an MLX90614 infrared thermometer. With the usage of three devices: a custom Lepton 3.5 camera, an Implexxio LT-1T thermistor and an MLX90614, the temperature of the target leaf of a laboratory plant was automatically measured every five minutes. The data would then be recorded on a private cloud server and manually measured with a handheld FLIR E8-XT. The performance of these three devices was evaluated to the standard of a highly accurate Implexxio LT-1T thermistor using the mean absolute error and root mean squared error. Among the non-contact sensors- MLX90614, Lepton 3.5 and FLIR E8-XT- the MLX90614 sensor showed the highest accuracy. However, the Lepton 3.5 module had an accuracy of less than ±2°C, which was similar to FLIR E8-XT and much better than the error value specified for the Lepton. The low-cost Lepton 3.5 can be used to periodically measure leaf temperature with an accuracy comparable to that of an intermediate-level thermal imager.