Leaf Health Status Classification Dataset

The current agricultural industry faces challenges in pest and disease monitoring and management, especially in large-scale plantations, where manual inspection is costly and inefficient. The application of existing machine vision technology in target detection and classification is not yet widespread, leading to an inability to promptly respond to pest invasions. This dataset aims to provide high-quality leaf health status classification data to support automatic pest and disease recognition and monitoring in the agricultural field. Data collection uses high-resolution cameras in different field environments to ensure coverage of various lighting and weather conditions. To ensure data quality, multiple rounds of labeling and consistency checks are conducted, reviewed by agricultural experts. Data storage uses JPG format, organized by folder structure, ensuring easy access and use. This dataset not only improves labeling accuracy (up to 95%) but also optimizes data consistency and integrity, ensuring the key features of each sample are preserved. By introducing new data augmentation techniques, the generalization ability of models is enhanced, expecting to improve performance metrics in pest and disease identification tasks by over 20%.