Automatic Umbrella Button Detection Dataset

The current retail e-commerce industry is facing challenges of automation and intelligence transformation, especially in object recognition and automated processing. Existing solutions have certain deficiencies in the accuracy and real-time performance of object detection, resulting in suboptimal application of automation equipment. This dataset aims to address the accurate detection needs of robotic arms for umbrella key points in opening tests, enhancing automated processing capabilities. Data collection uses high-resolution cameras in various environments, ensuring coverage of different angles and lighting conditions. We have implemented multiple rounds of annotation and expert review to ensure annotation precision and consistency. Data is stored in JPG format with structured storage for each image and its corresponding annotation information. The core advantage of the dataset is high-quality annotation, with annotation precision of bounding boxes and key points reaching over 95%, and consistency over 90%. We applied new data enhancement techniques to ensure the model's generalization ability in different scenarios. By using this dataset, the accuracy of object detection tasks is expected to increase by more than 20%, significantly improving the performance of automation equipment.