CCMT: Dataset for crop pest and disease detection.

Area of Science:

• Agricultural Science
• Computer Science
• Data Science

Background:

• The agricultural sector in developing countries faces challenges like knowledge gaps and pest infestations.
• Artificial Intelligence (AI) offers potential solutions for pest/disease control, cost reduction, and yield improvement in agriculture.
• Addressing these challenges requires accessible, high-quality data for AI model development.

Purpose of the Study:

• To present a novel dataset of crop pest and disease images from Ghana.
• To facilitate the development and application of AI technologies in agriculture.
• To bridge the technology gap for farmers in developing nations.

Main Methods:

• Collected raw images of crop pests and diseases from local farms in Ghana.
• Augmented the raw image dataset and split it into training and testing sets.
• Ensured all images were de-identified and validated by expert plant virologists.

Main Results:

• The dataset comprises 24,881 raw images across Cashew, Cassava, Maize, and Tomato crops.
• An augmented dataset contains 102,976 images across 22 distinct classes.
• The dataset is validated and made freely available for research purposes.

Conclusions:

• The presented dataset is a valuable resource for training AI models for crop pest and disease identification.
• This initiative can significantly contribute to improving agricultural practices and food security.
• The open availability of the data promotes collaborative research and innovation in agricultural AI.