Real-Time AI-Assisted Push-Broom Hyperspectral System for Precision Agriculture

Area of Science:

• Agricultural Technology
• Spectroscopy
• Artificial Intelligence

Background:

• Modern agriculture requires advanced technologies for crop management and sustainable food production.
• Hyperspectral imaging offers detailed crop monitoring capabilities.
• Artificial intelligence enhances the analysis of complex spectral data.

Purpose of the Study:

• To develop and implement a real-time AI-assisted push-broom hyperspectral system for plant identification.
• To demonstrate the system's capability in accurately capturing and analyzing spectral data for crop monitoring.
• To apply the system for plant leaf classification using neural networks.

Main Methods:

• Design and construction of a push-broom hyperspectral spectrometer, including optical assembly and system integration.
• Development of a real-time acquisition and classification system using an embedded computing solution.
• Calibration and resolution analysis of the hyperspectral system.
• Application of a neural network-based AI algorithm for continuous hyperspectral data analysis.

Main Results:

• The developed hyperspectral system accurately captures spectral data, validated through calibration and resolution analysis.
• The AI algorithm effectively analyzes hyperspectral data for plant classification.
• The system demonstrated real-time analysis of up to 720 ground positions at 50 frames per second.

Conclusions:

• The real-time AI-assisted push-broom hyperspectral system is a viable tool for accurate plant identification and crop monitoring.
• Integration of AI with hyperspectral technology significantly enhances agricultural management capabilities.
• This technology contributes to optimizing crop management and supporting sustainable food production.