elocation-id: elocation-id: e4097
Monitoring and evaluating crop water status is critical to optimizing water use and promoting sustainable agricultural production. The measurement of canopy temperature using thermal images has established itself as a reliable technique for estimating plant water status; however, its adoption is limited by the high cost of thermal cameras and unmanned aerial systems. Due to the above, a low-cost thermal camera was developed based on the 32 × 24 pixel MLX90640 infrared sensor and the Teensy 3.6 microcontroller, with storage capacity on a microSD card. The device was integrated into an unmanned aerial system and evaluated in a corn crop, with its measurements compared with those of a LI-COR LI-1600 porometer (accuracy ±0.5 °C). The thermal images were enhanced by bicubic interpolation and fused with RGB images to obtain images with a resolution of 640 × 480 pixels, which were then processed and segmented using binary images to isolate the pixels corresponding to the crop canopy temperature. Subsequently, the temperature values from the thermal images and the LI-COR LI-1600 porometer at each sampling site were compared, yielding a root mean square error of 0.74 °C. The results show that the developed thermal camera offers adequate accuracy, low cost (135 USD), and high spatial representativeness, positioning itself as a promising tool for thermal canopy monitoring in precision agriculture and efficient water resource management applications.
canopy temperature, infrared sensors, RPAS, thermal camera, water stress.
Monitoring crop water status is essential to maintain productivity and quality; as reliable indicators of water stress, they not only describe the availability of water in plants but also allow predicting the effects of deficit irrigation on yield ( Noguera et al ., 2020 ; Abioye et al ., 2022 ; Jiménez et al ., 2022 ).
It is necessary to optimize and improve techniques for recognizing water stress that support precision irrigation management, thereby reducing production damage and yield loss ( Bijanzadeh et al ., 2022 ; Bwambale et al ., 2022 ; Melo et al ., 2022 ).
In this context, studies have shown that plant canopy temperature can be used to detect moderate to severe levels of water stress in crops ( García-Tejero et al ., 2018 ; Gheysari et al ., 2021 ; Bo et al ., 2023 ).
The relationship between leaf temperature and stomatal regulation establishes that, when a crop has an adequate water supply, stomatal conductance increases and leaf transpiration improves, thereby generating a decrease in temperature due to cooling by water evaporation, since the temperature decreases as water is lost through the stomata ( Gheysari et al ., 2021 ; Ouma et al ., 2024 ).
On the other hand, plants with a water deficit reduce the stomatal conductance of the leaf and the transpiration rate; consequently, the canopy temperature increases, since there is no heat dissipation by evaporation, thereby altering the plant’s energy balance ( Sagan et al ., 2019 ; Noguera et al ., 2020 ).
Therefore, the temperature of transpiring leaves should be close to the air temperature, whereas in the case of stressed leaves, it should be higher. This is the assumption that allows estimating water stress from canopy temperature ( López-López et al ., 2009 ; Bijanzadeh et al ., 2022 ).
Obtaining infrared radiometric images is a valuable tool in remote sensing applications for precision agriculture, as it enables irrigation monitoring, assessment of crop physiological status, and yield estimation. In this context, non-radiometric, uncooled, and low-cost IR cameras have emerged as viable alternatives for new implementations in agricultural management ( Gomes et al ., 2017 ; Dong et al ., 2023 ; Paciolla et al ., 2025 ).
Infrared cameras allow subtle changes in canopy temperature to be measured using the longwave radiation emitted ( Sagan et al ., 2019 ; Yun et al ., 2024 ). Their implementation on unmanned aerial vehicles (RPAS) makes it possible to obtain high-resolution spatial and temporal maps ( García-Tejero et al ., 2018 ; Wu et al ., 2023 ). This allows us to more precisely understand the variability in water and physiological conditions of crops, as well as in soil characteristics ( Wu et al ., 2023 ; Oz et al ., 2025 ).
In this context, the present work aims to develop a low-cost thermal camera based on a Teensy 3.6 development board and a 768-pixel MLX90640 infrared sensor to monitor canopy temperature, which is used to determine water stress in agricultural crops, with an accuracy comparable to commercial reference instruments.
The system was integrated into an RPAS platform to obtain georeferenced information on canopy temperature and assess crop water status, thereby providing an accessible tool for precision irrigation management.
A low-cost thermal camera device was developed to measure canopy temperature in agricultural crops and to monitor water status ( Figure 1 ).
The system is based on an MLX90640 infrared sensor (Melexis, Ieper, Belgium), integrated into a SparkFun development board (SparkFun Electronics, Niwot, Colorado, USA); the sensor consists of a 32 × 24 pixel thermal matrix that allows surface temperature information to be obtained from the thermal radiation emitted by the objects as a function of their emissivity.
The MLX90640 infrared sensor is an electronic transducer-type thermopile that converts thermal energy into an electrical signal, operating on the principle that all bodies emit thermal radiation in the far-infrared region ( Elsayed et al ., 2017 ; Paciolla et al ., 2025 ). It is capable of taking non-contact remote temperature measurements with 16-bit resolution and an operating range of -40-85 °C, considered suitable for agricultural applications.
For data processing, the Teensy 3.6 development board was used, which is equipped with a 32-bit 180 MHz processor, digital pins with interrupt capability and 3.3 V tolerance, and USB connectivity. The microcontroller was programmed in the Arduino IDE environment using the ‘Teensyduino’ plugin, which enabled us to calculate the values corresponding to the 768 pixels of the MLX90640 sensor and store the temperature information on a microSD card, taking advantage of the board’s built-in socket.
The electronic circuit was designed to ensure a proper power supply and operation of the thermal camera, complying with the voltage and current consumption requirements of the devices used: Teensy 3.6 development board (100 mA), MLX90640 sensor (23 mA) and light-emitting diodes (LEDs, 10 mA).
The electronic diagram ( Figure 2 ) shows the power supply and operation of the system, which has a four cell LiPo battery with a nominal voltage of 14.8 V and 4 000 mAh, responsible for supplying power to the flight controller. The battery voltage is regulated by a universal battery eliminator circuit that provides a 5 V output with a capacity of up to 4 A.
The thermal camera system integrates a Mini560 regulator, which reduces the input voltage from 5 V to 3.7 V with a maximum current of 1 A, to power the Teensy 3.6. Additionally, 1N4007 diodes were implemented at the output in order to protect the circuit against current spikes and reverse voltages and to reduce the electrical noise present in the main signal.
The program used on the device’s controller board was developed in the Arduino IDE environment (version 1.8.12). Figure 3 illustrates the complete routine, detailing the functions performed by the Teensy 3.6 microcontroller and the MLX90640 sensor.
The temperature data were corrected to estimate the actual values of the canopy temperature. The sensor output data were processed together with the stored calibration coefficients in order to normalize the sensor response and obtain
Elsayed, S.; Elhoweity, M.; Ibrahim, H. H.; Dewir, Y. H.; Migdadi, H. M. and Schmidhalter, U. 2017. Thermal imaging and passive reflectance sensing to estimate the water status and grain yield of wheat under different irrigation regimes. Agricultural Water Management. 189(1):98-110. https://doi.org/10.1016/j.agwat.2017.05.001.
García-Tejero, I. F.; Gutiérrez-Gordillo, S.; Ortega-Arévalo, C.; Iglesias-Contreras, M.; Moreno, J. M.; Souza-Ferreira, L. and Durán-Zuazo, V. H. 2018. Thermal imaging to monitor the crop-water status in almonds by using the non-water stress baselines. Scientia Horticulturae. 238(1):91-97. https://doi.org/10.1016/j.scienta.2018.04.045.
Gheysari, M.; Pirnajmedin, F.; Movahedrad, H.; Majidi, M. M. and Zareian, M. J. 2021. Crop yield and irrigation water productivity of silage maize under two water stress strategies in semi-arid environment: two different pot and field experiments. Agricultural Water Management, 255(1):1-9. https://doi.org/10.1016/j.agwat.2021.106999.
Melo, L. L. de; Melo, V. G. M. L.; Marques, P. A. A.; Frizzone, J. A.; Coelho, R. D.; Romero, R. A. F. and Barros, T. H. da S. 2022. Deep learning for identification of water deficits in sugarcane based on thermal images. Agricultural Water Management. 272(1):1-13. https://doi.org/10.1016/j.agwat.2022.107820.
Oz, N.; Sochen, N.; Mendlovic, D. and Klapp, I. 2025. End-to-end pipeline for simultaneous temperature estimation and super resolution of low-cost uncooled infrared camera frames for precision agriculture applications. Electrical Engineering and Systems Science. 250(1):1-24. http://arxiv.org/abs/2502.13985.
Sagan, V.; Maimaitijiang, M.; Sidike, P.; Eblimit, K.; Peterson, K. T.; Hartling, S.; Esposito, F.; Khanal, K.; Newcomb, M.; Pauli, D.; Ward, R.; Fritschi, F.; Shakoor, N. and Mockler, T. 2019. UAV-based high resolution thermal imaging for vegetation monitoring and plant phenotyping using ICI 8640 P, FLIR Vue Pro R 640 and thermomap cameras. Remote Sensing. 11(1)1-29. https://doi.org/10.3390/rs11030330b.
