Estimation of crop water stress index for real-time durian orchard monitoring system

This thesis presents a comprehensive framework for estimating the crop water stress index (CWSI) in durian farming. By integrating the Kalman filter with a neural network, the framework improves the accuracy of canopy temperature estimation, identify the non-water-stressed baseline for durian tree, and reduces noise in CWSI calculations. Traditionally, stem psychrometers have been used to identify water stress in plants. However, they are not designed for long-term use and are not suitable for real-time monitoring. CWSI provides a suitable alternative, allowing for continuous monitoring and accurate estimation of water stress levels in durian trees. The main objective of this research is to estimate CWSI to enhance water management practices and optimize durian productivity. The framework’s evaluation was conducted on a durian farm in Chanthaburi, Thailand. where we collected data on the durian tree from December 2020 and November to December 2021. By utilizing IoT thermal cameras for data collection, the framework successfully estimated canopy temperature with a notable 27.82% improvement in correlation. One of the key contributions of this work is the proposed method to identify the non-water-stressed baseline for CWSI calculation, which plays a vital role in accurate estimation. Additionally, the framework reduces noise in CWSI by approximately 9.47%. Overall, the results demonstrate significant improvements in accuracy and correlation with stem water potential, with the framework achieving an impressive average correlation of -0.7790. This represents an enhancement of up to 40% in overall accuracy compared to not using the framework. However, it is essential to note that the evaluation is based on limited stem water potential data. The advantages of this framework extend beyond improved accuracy and correlation. It promotes water conservation and enables more ecient water management, which is crucial for sustainable durian production. By accurately estimating the CWSI, farmers can make informed decisions about irrigation scheduling, leading to better water management practices and increased durian productivity. This framework presents a valuable tool for the agricultural sector, aligning with the principles of precision farming and smart farming to contribute to the advancement of sustainable durian cultivation and support ecient water management practices.