Research on 3D gesture recognition in virtual maintenance

Abstract

In the process of virtual maintenance training, in order to achieve more natural human-computer interaction and enable trainees get the best immersion, it is particularly important to recognize 3D gestures in the real environment. By analyzing the development status of today's three-dimensional gesture recognition technology, gesture recognition device Leap Motion Controller is used to collect gestures in virtual maintenance training, and simulation experiments are carried out by SVM and PNN algorithm. The results show that the static gesture recognition rate of both algorithms can reach 100%, but SVM has a higher computing efficiency. The optimized PNN can achieve higher recognition rate for dynamic gestures, and the processing efficiency after PCA processing is higher.