Subtle Hand Action Recognition in Factory Based on Inertial Sensors

摘要

Recognition of the hand action an important role in the factory. How to improve our method for dynamic both-hand gestures, so as to recognize and read more fingerspelling signs automatically, is an advanced problem. In the paper, we propose an operation action recognition method based on two wearable devices containing inertial sensors. And we using a new idea that using the acceleration and gyroscope data, while segment the action in the same time. This idea will improve the precision for the segmentation technology. The main work for hand action recognition system included data acquisition, hand segmentation, hand feature extraction, and gesture recognition. To systematically preprocess the data, the features which consist of spectral entropy, the sum of acceleration, angular rate and angle are extracted. In order to realize the proposed targets, a massive experimentation was made in a realistic surroundings. In order to corroborate the analysis, we performed four classification algorithms which include Support Vector Machine, k-nearest neighbors, Naive Bayes and Extreme Learning Machine. As verified by the simulation results, ELM tends to have better scalability and similar.