The analysis of technical details in the teaching process of college volleyball cannot be separated from the analysis and evaluation of its movements. In this regard, this paper is based on the deep learning algorithm for volleyball basic action recognition, and then construct a training assistance system that can analyze volleyball video action in detail, so as to improve the athletes’ athletic ability. The technical action recognition model is a deep learning neural network model containing two layers of one-dimensional convolutional neural network (CNN) and one layer of long- and short-term memory network (LSTM), which can realize the accurate recognition of volleyball actions. The evaluation and recognition accuracy of the four types of movements, namely pad, serve, dunk and block, reached 84.25% after analysis. Based on the algorithm design of this paper, the training assistance system can significantly improve the athletes’ athletic ability after application and teaching practice (P>0.05). It shows that the system in this paper can effectively improve the quality and effect of volleyball teaching and promote the modernization of volleyball teaching.
