A low-cost RGB-D/thermal platform for monitoring fruit temperature with spatial resolution

Fruit sunburn damage in orchards is a growing concern exacerbated by climate change and more frequent heatwaves. As temperatures increase, the risk of sunburn intensifies due to excessive solar radiation and heat stress. This compromises the fruit marketability, reducing growers’ income. Understanding the impact of heatwaves on fruit sunburn occurrence and severity is crucial for safeguarding crop yields and ensuring growers’ profitability. This study is part of a European project aiming to create an alert system for fruit sunburn damage based on weather data. One aspect of the project involved developing a low-cost platform able to create a 3D thermal distribution of fruit temperature at both the plant and orchard levels to better understand fruit temperature dynamics in relation to sunburn damage occurrence. The system comprises consumer-grade depth and thermal cameras powered by Python and ROS2. The software aligns thermal, colour, and depth images of the scene. Using these data, an artificial intelligence algorithm automates the detection of well-exposed fruits. For each identified fruit, the system extracts its temperature, corrects it for camera distance, determines the fruit’s position as X, Y and Z coordinates relative to the tree trunk, and provides a graphical representation. Additionally, if GPS information is available, the system can geolocate the collected data. Preliminary results indicated an image alignment error of ±0-6 pixels, a fruit temperature estimation error of ±1.2°C (mainly influenced by camera-object distance), and a fruit positioning error of ±3.5 cm. The system is currently undergoing further development to improve its performances.