Industry 4.0: An AI breakthrough in assessing the quality of Vietnamese agricultural products
Originating from practical needs, a research project at the Vietnam Academy of Science and Technology (VAST) entitled "Research on the Application of Artificial Intelligence Technology in the Assessment of Agricultural Produce Quality" (Project Code: CN4000.02/22-25) has been implemented under the leadership of Dr Bui Quang Minh from the High-Tech Innovation Centre. The project focuses on developing an AI system for real-time monitoring, assessment and quality warning of agricultural produce. By combining deep learning, image analysis and physicochemical data, the research team has created an integrated solution that supports the application of AI in post-harvest preservation while opening a new pathway for digital transformation in the agricultural sector.
Dr. Bui Quang Minh in his office
As part of the project, the research team selected mangoes and dragon fruit - two high-value export fruits - as the research subjects. They also developed a comprehensive AI system comprising a standardised database, deep-learning models, management software and an operational process capable of being deployed under real-world conditions.
Samples of mangoes and dragon fruit - the two research subjects of the project
One of the project's most significant achievements is the successful development of a standardised database containing more than 14,400 images of mangoes and dragon fruit at different quality levels.
Building on this database, the team applied YOLOv8 - the latest version of the YOLO image recognition model -combined with a Convolutional Neural Network (CNN) to automatically detect, identify and classify agricultural produce.
According to Dr Bui Quang Minh, "Instead of requiring multiple rounds of analysis as in the past, the new AI model enables users to process an image only once to determine the location and characteristics of each fruit, thereby increasing processing speed and improving the accuracy of quality assessment."
AI-based fruit identification and quality assessment results.
The AI model detects and classifies the quality of mangoes and dragon fruit with an accuracy exceeding 90%
A particularly noteworthy aspect of the research lies not only in the application of AI for image recognition but also in its integrated approach, combining multiple data sources to provide a comprehensive assessment of agricultural produce quality. The research findings open up significant opportunities for practical applications.
Translated by Tuyet Nhung
Link to Vietnamese version
