Garden and Flower Mulberry Recognition Image Dataset

Currently, the agricultural and forestry industries face challenges in precise recognition and efficient monitoring of flower and fruit tree health management. Existing solutions often rely on manual detection, which is inefficient, costly, and limited by expert experience. To enhance recognition efficiency and accuracy, the Garden and Flower Mulberry Recognition Image Dataset was developed to address pest and disease recognition and health assessment through machine learning. The dataset is collected using high-definition cameras under various seasonal and lighting conditions, including images of various flowers and mulberries in different health states. Quality control ensures data accuracy and consistency through multiple rounds of annotation and expert consistency review. The annotation team consists of over 20 members, including plant pathology experts and computer vision engineers. Data preprocessing includes image enhancement, denoising, and size standardization, ultimately stored in a hierarchical JPG format for quick access and retrieval. The core advantage of this dataset is its high annotation accuracy and consistency, exceeding 95%. Advanced data augmentation techniques are employed to enhance model training performance, significantly improving the accuracy of pest and disease recognition. Compared to similar datasets, we not only increased the data volume to 100,000 images but also included rare diversity-indicating feature images, suitable for extended research on various regional environments and supporting multi-model compatible applications.