Research On Positioning Technology Of Medical Filling And Sealing Robot Based On Machine Vision

In recent years,intelligent manufacturing,with its remarkable advantages of intelligence,collaboration and flexibility,has gradually replaced the traditional manufacturing industry and become the focus of the development of manufacturing industry.Pharmaceutical manufacturing is the key field of intelligent manufacturing,which is closely related to people’s health and well-being.With the continuous growth of the global pharmaceutical market,people have higher requirements for the quality,production efficiency and technology of drugs.In recent years,China has launched the "Made in China 2025" policy,and continuously and deeply promoted the pharmaceutical reform,and promoted the development of pharmaceutical manufacturing industry from mechanization and electrification to automation and intelligence.High-end intelligent features such as autonomous perception,autonomous decision-making and autonomous execution have become the urgent needs of pharmaceutical manufacturing industry.In this paper,the medical filling and sealing robot system is taken as the object,and the visual positioning technology of penicillin bottle in the process of medical filling and sealing is studied.The main research work of this paper is as follows:1.The overall structure,working method and workflow of the medical filling and sealing robot system are introduced in detail.According to the requirements of medical filling robot system,in order to obtain high-quality image information of bottle mouth,bottle stopper and bottle cap,the selection of camera,light source and other key equipment of vision system is completed.2.Several denoising,enhancement and edge detection algorithms are compared and analyzed.Finally,the method suitable for this paper is selected to preprocess the image of bottle mouth,bottle stopper and bottle cap.Aiming at the problems and shortcomings of traditional Canny edge detection algorithm,an improved Canny edge detection algorithm is proposed to extract the edge of the bottle mouth and other objects.In this algorithm,the noise of image is removed by steering filter,and the gradient amplitude and direction of image are accurately obtained by Kirsch operator.After the non-maximum suppression,the Ostu threshold segmentation method is used to obtain the best high and low thresholds.Finally,threshold screening is carried out to complete the edge detection of bottle mouth,bottle stopper and bottle cap images.The experimental results show that the algorithm can effectively obtain the edge information of the target and eliminate the interference of other invalid edges.3.In order to complete the accurate positioning of the bottle mouth and other target objects,a multi-target localization algorithm based on machine learning is proposed.Firstly,the advantages and disadvantages of the traditional DBSCAN clustering algorithm applied to the target segmentation of bottle mouth and other images are analyzed,and then the DBSCAN algorithm has been improved.The improved DBSCAN algorithm takes advantage of the characteristics of image neighborhood,greatly reduces the complexity of the algorithm,and automatically calculates the core parameters,which can achieve fast and accurate target segmentation.Then,a fast algorithm for locating the center of circle based on DBSCAN is proposed.The specific steps of the algorithm are as follows: firstly,the least square fitting circle method is used to roughly locate the segmented target and obtain the center;secondly,the outer edge point set is obtained by radial scanning method;secondly,the random three-point fitting circle method is used to obtain the candidate center set composed of several centers;finally,the DBSCAN clustering algorithm is used to extract the real center set and estimate the center to complete the location.The experimental results show that the average positioning error of the DBSCAN based circular positioning algorithm is about 0.71 pixels,and the positioning accuracy is better than the traditional algorithm.At the same time,the average execution speed of the algorithm is only about 1.20 ms,which can meet the requirements of accuracy and real-time of medical filling and sealing production line.4.A software system is designed according to the actual needs of the medical potting robot in the production process.The function,composition and characteristics of the software system are described.The design and implementation of login module,main control module,camera setting module and light source setting module of the software system are introduced in detail.The software system can meet the requirements of rapid and accurate positioning of the bottle mouth,bottle stopper and bottle cap,as well as stable and reliable operation.