Indian major basmati paddy seed varieties images dataset

Area of Science:

• Agricultural Science
• Computer Science
• Data Science

Background:

• Accurate identification of Indian basmati paddy seed varieties is crucial for quality control and trade.
• Existing methods for seed classification may lack efficiency and scalability.
• A comprehensive, labeled dataset is needed for developing advanced AI models.

Purpose of the Study:

• To create and present a novel dataset of Indian basmati paddy seed images.
• To facilitate the development of AI-based models for automated seed classification and adulteration detection.
• To support research in agricultural informatics and machine learning applications in agriculture.

Main Methods:

• Collected and curated 3210 RGB JPG images of 10 Indian basmati paddy seed varieties and an 'Unknown' class.
• Utilized an in-house apparatus and a 5MP tablet camera under standard conditions.
• Pre-processed images for direct use in training and testing machine learning models.

Main Results:

• Developed a dataset comprising 11 classes, including 10 distinct basmati varieties and an 'Unknown' class for differentiating paddy from other grains.
• The dataset enables the training of AI models for accurate paddy seed classification.
• Demonstrated the potential for AI models in detecting adulteration and automating classification during rice threshing.

Conclusions:

• The presented dataset is a valuable resource for advancing AI-driven solutions in basmati paddy seed identification.
• The dataset supports the development of robust models for quality assessment and authenticity verification.
• Future applications include real-time classification systems and expanded variety recognition.