Agapanthus Recognition Image Dataset for Garden Flowers

With the advancement of urban greening projects, precise management and diagnosis of garden plants have become increasingly important. However, traditional flower recognition methods require extensive manual involvement, are time-consuming, labor-intensive, and their accuracy is difficult to guarantee. Existing automated recognition systems generally rely on a limited number of feature dimensions, resulting in insufficient performance when encountering different growth environments and cultivar variations. This dataset focuses on Agapanthus, a common horticultural plant, providing systematic image data to support more refined and generalizable automated recognition technology. We used high-definition digital cameras to capture images under different climatic conditions, ensuring data diversity. The data underwent multiple rounds of manual annotation and expert verification to ensure annotation consistency and accuracy. The annotation team consisted of five botany professionals, ensuring scientific and authoritative annotations. In preprocessing, techniques such as edge detection and background denoising were used. Data is stored in JPG format with clear labels, aiding quick feature extraction during training. In terms of data quality, we provided over 95% annotation accuracy and strict annotation consistency, ensuring the dataset's integrity. Technically, we introduced new data augmentation techniques, including random cropping and color transformation, to enhance data diversity. This dataset can significantly improve the performance of current flower recognition systems, with an accuracy increase of up to 15%. Compared to similar products, our dataset has a broader coverage and unique high-resolution images, supporting larger-scale model training and deployment. Additionally, the data format is reasonably designed for easy transplantation and application in different AI models, offering high scalability and versatility.