Joint Insect Pest and Disease Detection Dataset

The current agricultural sector faces threats from pests and diseases to crops, leading to decreased yields and quality. Traditional manual recognition methods are inefficient and prone to errors, unable to meet the high-efficiency demands of modern agriculture. Existing solutions often focus only on a single factor, lacking comprehensive evaluation of pests and diseases. This dataset aims to support AI in joint recognition by including images of various disease spots and insect holes, enhancing recognition accuracy and efficiency. Data collection employs high-resolution cameras under natural light to ensure image clarity. In terms of quality control, multiple rounds of annotation and expert review ensure consistency and accuracy of annotations. Data is stored in JPEG format, organized by category for ease of use and analysis. The core advantage of the dataset lies in its high-quality annotations, with annotation accuracy over 95%. Its innovative methods in multi-factor recognition significantly improve recognition accuracy, potentially increasing pest and disease recognition efficiency by over 30%. Through data augmentation techniques, this dataset can effectively expand sample diversity and improve model robustness.