Industrial Robotic Arm Grasping Action Recognition Dataset

In the current industrial field, the application of robotic arms is increasingly widespread, but recognizing grasping actions in complex environments still faces many challenges such as lighting changes and background interference. Existing solutions often rely on traditional image processing techniques, which are difficult to effectively handle variable environments and complex action patterns. Therefore, this dataset aims to provide researchers with a rich action recognition dataset that covers various grasping actions and their paths, facilitating the training and testing of deep learning models. Data collection uses high-resolution cameras to record the grasping actions of robotic arms in different environments to ensure video quality. We conducted multiple rounds of annotation and consistency checks, and all annotations were reviewed by experienced experts to ensure the accuracy and consistency of the annotations. The data is stored in MP4 format for videos and JSON format for annotations, facilitating subsequent processing and analysis.