Automatic mouth sampling based on computer vision and robot control

With the increasing application of robots, the automated and unmanned operation is becoming a new trend. In the scenario of nucleic acid sampling, the medical staff are under great exposure to high-risk substances. Compared to humans, unmanned sampling has higher accuracy and round-the-clock service capabilities, most importantly, it protects the medical staff from getting infected. To accomplish the benefit reached by the robot, an automatic nucleic acid sampling method based on computer vision and robot control is proposed in this paper. The pipeline first obtains 2D image coordinates of the target sampling points by YOLO detection algorithm, then translates the 2D image coordinates to 3D robot base coordinates, and finally controls the robot to achieve automatic mouth sampling. Experiment results demonstrate the feasibility and accuracy of the pipeline.