Effective identification and correction of swimmers’ improper postures can significantly improve athletes’ weekday swimming training quality. The human body’s affine deformation is prone to occur during swimming movements when performing posture recognition and correction, resulting in the creation of low-brightness action feature locations. The inability of coaches to identify and correct athletes’ improper posture in real time is a result of a lack of detection and correction. Additionally, the human skeleton motion data from the depth camera Kinect contains a high amount of noise and fewer skeleton nodes, and the data level of detail is low. To overcome this issue, this research proposes a network for enhancing Kinect skeleton motion data. The network is composed of six bidirectional cyclic autoencoder stacks. The stacking structure improves the smoothness and naturalness of the data, and the training phase includes hidden variable limitations to ensure that the bone motion data preserve a genuine bone shape when the degree of detail is raised. The trials demonstrate that the optimized data from the network have a better degree of smoothness and can keep a more realistic bone structure, enabling the goal of obtaining high-precision motion capture data with low-precision Kinect equipment to be met.

Swimming is predominantly a long-distance endurance sport. In this sport, like in many others, monitoring and tracking swimming attitude error correction and how it changes over time is critical. Besides, in swimming posture measurement, due to the absorption and refraction of various signals of water, sensors relying on external information cannot provide accurate information. In addition, the inertial technology is not dependent on external information, suitable for the field of swimming posture measurement. Inertial attitude measurement requires initial alignment technology to provide initial values to calculate the attitude in the swimming movement. However, the swimmer’s jump time is uncertain and the traditional initial alignment algorithm needs a long time to get a high-precision result. To solve this problem, in this paper, we proposed a fast alignment method by the Mahony algorithm. In our work, we have used nine-axis inertial measurement unit of micro-electro-mechanical systems (MEMS) to collect information. Furthermore, we calculated the attitude angle by angular velocity information, horizontal attitude angle, and yaw angle, and these were corrected by acceleration information and magnetic field intensity information, respectively, and multisource information was integrated by a complementary filtering method. It is simple to acquire the starting value of inertial attitude measurement. Laboratory experiments verify that the horizontal accuracy of attitude angle can reach the angle classification within 3 s, which meets the requirements of swimming sports. The algorithm’s viability is further confirmed through experiments in real-world sporting scenarios.