Rapid online plant leaf area change detection with high-throughput plant image data

Area of Science:

• Agricultural Science
• Computer Vision
• Plant Biology

Background:

• High-throughput plant phenotyping (HTPP) generates massive image data, posing challenges for efficient plant trait prediction and anomaly detection.
• Accurate and non-destructive monitoring of individual plant traits is crucial for plant breeding and crop management.

Purpose of the Study:

• To propose a novel two-step image-based online detection framework for real-time monitoring and change detection of individual plant leaf area.
• To address the challenges of massive image data in HTPP by enabling efficient plant trait prediction and anomaly detection.

Main Methods:

• Development of a two-step image-based online detection framework utilizing real-time imaging data.
• Implementation of a system for monitoring and detecting changes in individual plant leaf area with minimal detection delay.
• Validation of the framework's efficiency using real-world data analysis.

Main Results:

• The proposed framework achieves a smaller detection delay compared to baseline methods under a predefined false alarm rate.
• The method operates in real time without the need to store all past image information.
• Demonstrated efficiency and applicability through real data analysis.

Conclusions:

• The developed framework offers an efficient solution for real-time monitoring and anomaly detection in HTPP.
• This approach facilitates faster and more accurate plant trait analysis, supporting advancements in plant breeding and crop management.
• The system's ability to process data in real time makes it suitable for dynamic agricultural monitoring applications.