Identification Method of Higher Vocational Basketball Course Teaching and Training Action Violation Based on Deep Learning

When identifying violations in basketball training movements, the accuracy and stability of the Kinect device affect the fusion of joint angle space in the obtained three-dimensional coordinate information of human skeletal points, making it difficult to achieve good recognition results. Therefore, a deep learning based method for identifying action violations in vocational basketball course teaching and training is proposed. The article first extracts the human skeleton contour characteristics of teaching and training actions, then uses in-depth learning to construct the recognition model of teaching and training action violations in higher vocational basketball courses, so as to realize the recognition of teaching and training action violations. Finally, the application of experiments verifies the progressiveness of the proposed method. The experimental results show that the designed recognition method of teaching and training action violations in higher vocational basketball courses has good recognition effect, excellent indicators and reliability, which can make a certain contribution to improving the teaching quality of higher vocational basketball courses.