Reseach On Visual System For The Arm Of Power Pharmacy Robot

Intravenous drug therapy is now commonly used in clinical treatment, theenvironment configuration and the nursing labor intensity for intravenous drugconfiguration are highly required. PIVAS (intravenous drug distribution center) ispresently the best solution to improve the quality of the intravenous drugconfiguration, but there still exist some significant problems in the nurse laborintensity and body protection, such as greater injury for operator when configuringchemotherapy drugs for cancer patients. Meanwhile, the accuracy of chemotherapydrugs configuration has a great influence on treatment. In order to enhance thequality of the chemotherapy drugs configuration, and to protect health care workers,we propose the chemotherapy drugs configuration robot: robot manipulator is doneautomatically, according to the dispensing needs of the precise configuration of thevarious chemotherapy drugs. Because of the variety of chemotherapy drugs anddispensing process, the high security requirements for the choice of drugs and thedispensing accuracy higher than95%, the visual system of the robot arm is crucial,especially in high-precision positioning of the dispensing module, which shows thedecisive significance for the robot’s accuracy and safety.The paper first analyzes the dispensing process of the chemotherapy drugsconfiguration robot, completes the selection of the main components of the visualsystem, and on this basis, proposes an equidistant pseudo binocular imageacquisition method based on the monocular vision system, establishes a targetfeature position and orientation conversion algorithm by researching the inherentrelationship between image coordinates and space coordinates to avoid the complexspatial matrix conversion and to improve the positioning accuracy. In addition, theexternal parameters of the camera are determined according to the active vision-based camera calibration method in the equidistant pseudo binocular imageacquisition mode.Then introduces the key technologies in the image processing algorithm. Hostcomputer software preprocesses the received image, determines the gray-scaletransformation approach in gray histogram stretching mode according to the gray-scale transformation effect of the gray histogram equalization and gray-scalehistogram stretch. On this basis, completes grayscale image segmentation based onthe threshold method and the fits straight line contour based on the least squaresmethod to acquire the location information of the target feature point in the image.In order to enhance the real-time of the algorithm and shorten the response time, the paper preprocesses to get region of interest in the image and integrates thepointer method in the image searching process.The paper also sets up the dispensing experimental platform of chemotherapydrugs configuration robot and uses C#language programming on the Point GreyFlea2camera, to achieve the automatic identification of the syringe needlepositioning and successfully guide the Staubli TX40robot arm finishing thedispensing experiments. The program runs less than200ms and the positioningerror is in the circle with0.2mm diameter, which all these verify the speed andprecision characteristics of the algorithm.