Smartphone-Based SPAD Value Estimation for Jujube Leaves Using Machine Learning: A Study on RGB Feature Extraction...

Area of Science:

• Agricultural Science
• Computer Science
• Plant Physiology

Background:

• Chlorophyll content in date palm leaves is vital for fruit yield and quality.
• Conventional chlorophyll detection methods are often complex and costly, hindering efficient agricultural management.

Purpose of the Study:

• To develop a rapid, cost-effective method for assessing chlorophyll content in date palm leaves.
• To evaluate the performance of machine learning and deep learning models in predicting chlorophyll levels from smartphone images.

Main Methods:

• Collected SPAD values and RGB images of Xinjiang date palm leaves.
• Extracted and selected 21 color features from preprocessed images using Python and OpenCV.
• Applied Principal Component Analysis (PCA) for feature downscaling.
• Trained and validated various models including Support Vector Regression (SVR), Relevance Vector Machine (RVM), Convolutional Neural Network (CNN), CNN-SVR, and CNN-RVM.

Main Results:

• The CNN-SVR model demonstrated superior performance, achieving R-squared values of 72.21% (training) and 77.44% (validation).
• Selected color features highly correlated with chlorophyll content were effectively utilized by the models.
• The proposed approach significantly outperformed other tested prediction models.

Conclusions:

• Smartphone-based image analysis combined with CNN-SVR offers a simple, accurate, and economical solution for chlorophyll detection in date palms.
• This novel method provides a valuable tool for precise crop management and health monitoring in the date industry.
• The technique shows potential for broad applicability in agricultural monitoring and precision farming.