Injury Prevention of Knee Joint Caused by Cross-step Fall in Beach Sports

In the current beach sports, the research on the injury prevention of knee joint has the problems of poor detection accuracy, long time-consuming, and unsatisfactory treatment effect of knee joint image. In order to solve these problems, the injury prevention method of knee joint caused by beach athlete's cross-step fall based on the PCA-SIFT feature and the Bayesian decision-making is proposed. The TE-WI uncooled focal plane infrared camera is used to capture the knee image of the athlete. The curvelet forward transformation is used to decompose the image into sub-bands of different scales and different directions, and obtain the curvelet coefficients. Generalized cross validation and genetics are used to acquire the asymptotic optimal threshold. The obtained curvelet coefficients are denoised and the image is reconstructed. The denoised image is subjected to the extension processing to enhance the image of the athlete's knee joint. The sample set is randomly selected, and the PCA-SIFT features of each type of image are extracted, and the unstable features are eliminated by the mutual matching rate. Bayesian decision-making is used to classify the images to be classified, and the information about knee joint injuries of athletes is obtained. The prevention and treatment plan is determined according to the damage information. Experiments show that with this method, the image enhancement coefficient is up to 0.99, the athlete's damage detection accuracy is up to 99%, and the damage detection time is the lowest, being 0.42 h. The method has the good image enhancement effect, which improves the accuracy of damage detection and reduces the time required for damage detection.