Research On Hand Gesture Recognition And Interaction Technique In Augmented Reality System

Augmented Reality system is an important branch of Virtual Reality technology, which combines virtual computer-generated material with the surrounding physical world, with the help of vision facilities it integrates the virtual objects with the real world circumstances and presents a real-world perception to the user.Hand-gestures itself contain massive interaction information and tally with human perception habits, therefore, it is intuitionistic and the natural way of human social communication mode in people's life. With the gradual transfer of human-computer interface towards user-orientated, study of hand-gestures recognition become the hotspot of people's study. Hand-gestures-recognition-based human-computer interface provides an effective and accurate mechanism through computer that introduces hand-gestures, which is intuitionistic and natural communication mode, into human-computer interface, human hands are then acted directly as the computer input equipment by users and no medium needed between human-computer communication. The implementation of human-computer interaction by using hand-gestures recognition can help deaf and mute person communicate with normal people, and enable the smooth transition between human perception and computer calculations, therefore, improves the human-computer interaction modes.This essay deals both with how to enhance the sense of third dimension in simulation scenes and how to apply new and convenient interaction methods in maintenance training emulator. It investigates the Augmented Reality and Hand Gesture Recognition techniques. And through the improvement of hand recognition method, an Augmented Reality system with hand-gesture interaction capability is designed. Such system uses hand-gesture image attributes to carry through the pretreatment of input hand-gestures, calculates the Fourier descriptors of the hand-gestures' periphery sequences and therefore obtains the hand-gestures' eigenvectors; and then through calculating the Euclidean distance and analogousness between input hand-gestures and the database samples of hand-gestures to implement the system function of input hand-gestures' recognition; simultaneously assistant images are made in accordance with the recognition results by the system. And then Augmented Reality registration technique is applied to implement the system virtual superposing function.By testing and experiments, such system would enhance the Augmented Reality technique and implement the hand-gestures interaction function with the technique of Hand Gesture Recognition. It boosts the training effects, and meets the requirements of simulation maintenance training.