Limb rehabilitation training detection method and system based on artificial intelligence and control equipment

The invention relates to a limb rehabilitation training detection method and system based on artificial intelligence, and control equipment, and the method comprises the steps: collecting a rehabilitation training image of a target, obtaining human body key points through key point detection, obtaining two-dimensional skeleton key points of the target through combining a component affinity relation vector field, and generating a 3D motion sequence through a time domain convolution model, namelythe three-dimensional coordinate information of human body key points; binding each piece of behaviorinformation with a corresponding patient wristband ID through a KM matching algorithm in combination with the three-axis angular velocity of a gyroscope in the patient wristband to obtain rehabilitation training information of a patient, and comparing the rehabilitation training information of a target with rehabilitation training information of a set standard to obtain rehabilitation training action deviation, judging w