<p>Although modern human motion capture systems have been widely used in many sports scenes, there are still some problems in swimming scenes, such as reliability, convenience, and accuracy, which need to be further improved. The existing optical motion capture systems are difficult to automatically segment the swimming phase, which also increases the difficulty and workload of measurement. Based on the existing situation and using optical imaging technology, this study designed a high-resolution pose and motion recognition system that can install cameras at the edge of a swimming pool or underwater, and capture and analyze the swimmer’s pose and motion in real-time using computer vision algorithms. By combining sensor technology, the system can collect body movement data of swimmers. By collecting these sensor data, the system can have a more comprehensive understanding of swimmers’ postures and movements. By training and establishing classification models, the system can identify and distinguish different swimming postures and movements, and provide real-time feedback and guidance based on the recognition results, helping swimmers improve their postures and skills and enhance swimming effectiveness.</p>