Visual Feedback System for Traditional Chinese Medical Massage Robot

摘要

Traditional Chinese medicine physiotherapy is an important part of Chinese medical culture. It can prevent and treat diseases by acupoints massage. However, the number of doctors who are proficient in acupoints is insufficient to meet the demand. It is necessary to use traditional Chinese medicine massage robot to complete this work. The method of teaching or offline programming is widely used in the general massage robot, which greatly reduces the flexibility of robots. Once the human body moves, the location is not accurate, and the real-time detection and recognition of acupoints cannot be performed. Considering the small size and the complex information of acupoints, this paper uses QR code for acupoints identification, which is convenient and laconic. The QR code is printed on A4 paper at 1x1 cm scale, and captured though camera. As for recognition, some methods of open source, such as Zbar, Zxing, opencv4.0, even Wechat are used, but the QR code recognition is inaccurate. We design a new two-dimensional code dedicated to acupoints location, with CRC check function, and a new two-dimensional code detection and recognition algorithm is proposed. In this paper, the camera calibration algorithm and the distance measurement model are designed to convert the detected recognition results to the world coordinate system. In the experiment, the measurement error of depth is controlled within ± 1.7 mm, and the measurement error of horizontal distance is controlled within ± 0.8 mm. The experiment verifies the effectiveness of the algorithm.