Gesture Recognition Technology In The Study Of The Application In Virtual Simulation Experiments

Experimental education plays an important role in the field of teaching;however, due to the uneven development of domestic education, such as lack of educational resources, people can't learn knowledge effectively. By the help of virtual experiment, people can control and operate experiments freely in the virtual environment increasing the interest in self-study; meanwhile, it can solve the problem effectively which is mentioned above. This paper studies the gesture recognition technique that is part of the field of computer vision so as to apply it to virtual experiments. This paper is mainly divided into four parts.Firstly, the current state of the development in the field of human-computer interaction at home and abroad is discussed in detail; also, the trend of virtual reality is presented. In order to develop a virtual experiment application which supports natural human-computer interaction, we discussed and studied the approach to applying gesture recognition technique into it.Secondly, for the purpose of improving the accuracy of gesture recognition,we optimized the algorithm of skin segmentation by combining with the depth information; thus, some distractions such as light and complex background were avoided. In regard to extracting features, we took the position and number of fingers into account. We made a foundation for the following research by these methods.Thirdly, the Cam Shift algorithm was used for tracking the target gestures continuously, and basic gesture features were used in the dynamic recognition stage.Finally, to verify the feasibility of the techniques applied in virtual experiments mentioned above, a virtual optics experiment which supports gesture recognition was designed and developed.The feasibility of gesture recognition technique used in virtual experiment has been proven successfully; meanwhile, virtual experimental apparatus can be operated freely by the operation gestures defined in the paper. This virtualexperiment has the advantages of lifelike experimental effect, stable natural human-computer interaction, and simple operation mode. The work of this paper offers a feasible approach to developing a better virtual experiment instead of traditional experiments.