This study intends to establish a human action comparison analysis framework that can be practically applied to factory production lines. This research is based on a human skeleton motion recognition technology (OpenPose), which captures the hand motion skeletons of the operators on the production line when assembling products, and then uses the dynamic time warping algorithm (DTW) in the time series data analysis, and K-Means clustering to distinguish the assembly proficiency of the operator for the assembly operation, as a reference for planning education and training. It is expected that the tools provided can help the factory quickly and accurately grasp the movement differences of each operator and obtain effective improvement suggestions. It is also hoped that the research results will help reduce the cost of education and training, improve the speed, accuracy, and quality of products produced by operators, thereby greatly increasing the overall performance of the manufacturing plant.
Kung, Ching Yuan and CHOU, TING-YANG, "Skeleton Based Action Analysis on Manufacturing Assembly Site" (2022). PACIS 2022 Proceedings. 87.
When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.