Hawthorn Recognition Image Dataset in Garden Flowers

With the development of urban gardening and the agriculture and forestry industry, accurate plant recognition has become increasingly important. However, as a type of garden flower, hawthorn has a complex and variable appearance, posing a significant challenge for ordinary growers and horticulturists to identify. Current recognition technologies largely rely on manual experience or simple image libraries, which have low accuracy and efficiency. This dataset aims to improve the accuracy and efficiency of plant recognition algorithms through high-quality image data, meeting the needs for precise classification. The data collection process involved taking photographs with professional cameras under natural light to ensure clear image quality and coverage of various growth environments. Through multiple rounds of annotation and consistency checks, the annotation accuracy exceeds 95%, and all data has been reviewed by botanical experts. The annotation team consists of professionals with a background in biology and information technology, totaling 30 participants. Data preprocessing includes image denoising, enhancement, and standardization to improve model training outcomes. The data is stored in a categorized directory structure for easy retrieval and use. This dataset features high-quality annotations, ensuring accuracy and consistency, and innovatively introduces data enhancement techniques to improve algorithm adaptability to complex environments. Its application value lies in increasing the accuracy and real-time performance of plant recognition systems. Experiments show that models using this dataset have improved accuracy by 20%. Compared with other datasets, it provides more comprehensive data on hawthorn appearance variations and has strong scalability, applicable to tasks for the recognition of other plants.