Swimming training evaluation system and method based on convolutional neural network

Abstract

The invention discloses a swimming training evaluation system and method based on a convolutional neural network. An underwater high-definition camera is arranged on a swimming pool bottom guide railand shoots upwards from the bottom of a swimming pool; a water surface high-definition camera is arranged on a swimming pool side wall track and is used for horizontally shooting from the side surfaceof the swimming pool; a high-altitude high-definition camera is installed on a rail above the swimming pool to shoot downwards and vertically and moves forwards along with swimming; the method comprises the following steps: sampling images at intervals, performing analyzing and processing to obtain swimming action data, taking multiple times of sampling in a complete swimming action period as input, taking swimming scores as output, establishing a swimming evaluation model by using a convolutional neural network, simulating and adjusting swimming postures of athletes, calculating scores aftersimulating and adjusting the postures, and further improving the swimming postures. The method is used in swimming training assistance, interval images of swimming postures and actions are captured through a camera, sampled swimming action data are adjusted and predicted through a neural network model, and feedback is performed to expect to help to improve the swimming performance.

Description

technical field [0001] The invention relates to a swimming hardware system and a swimming image data processing method, in particular to a swimming training evaluation system and method based on an image algorithm and a convolutional neural network algorithm. Background technique [0002] The traditional swimming training evaluation is mainly based on the experience of the coach and the judgment of the naked eye. The effectiveness of the swimming is evaluated through the athlete's water surface posture, movement, position and final score, which is relatively inefficient. [0003] The more advanced method on the market is to use a camera to shoot the whole swimming process of the athletes. After the competition, the coaching staff will manually analyze and judge the movements and postures during the swimming process, and propose a movement correction plan based on the results. Continue to shoot with the camera, and then analyze it based on the captured images, and repeat the ...

AI Technical Summary

Problems solved by technology

This method is time-consuming and labor-intensive, subject to the constraints of the coach's training experience, and the artificial subjective judgment has a great influence

[0004] In the prior art, there is a lack of hardware equipment and software methods for effectively helping swimming training

Images