The posture and mobility of workers in industrial production directly affect productivity, safety, and product quality. However, existing posture monitoring technologies for large-scale applications and real-time monitoring are limited.We utilize an improved FastDTW and a gesture-to-video matching algorithm that normalizes feature vectors. An industrial posture detection system was developed to enhance the efficiency and quality of the production line, with a prototype based on the Orange Pi 3B. Our algorithm was compared to advanced gesture recognition algorithms such as OpenPose, DeepLabCut, and TensorFlow. The results show that our method exhibits higher accuracy and faster processing speed due to its efficient feature extraction and optimized similarity calculation. The optimized hardware design and operating system support enable rapid processing and analysis of real-time data streams, combining the advantages of geometric feature extraction and FastDTW. Compared to other components, our system surpasses existing gesture recognition methods in many ways, opening new possibilities for real-time applications. The accuracy rate in the test data reaches 90%, which is comparable to other algorithms but less computationally intensive. This automated posture detection system improves productivity and quality, bringing significant technological innovation to industrial assembly lines.