Research On Detection And Grasping System Of Glass Bottle Defects Based On Machine Vision

In the manufacturing process of glass, it is inevitable to cause some broken bottles, bottle scratches and other defects.Especially for these products of high economic value such as perfumes, essential oils, and oral liquid, the appearance quality of bottles are of higher requirements.Taking the control of oil filling bottles as the control object, the research and design work of the glass bottles defect detection and robot crawling system is carried out by using machine vision technology, robot technology and industrial control technology. The specific content of the paper include:(1) The function of the system is defined, the design of the system based on defect detection machine vision and robot crawling is completed, and the related technologies of the machine vision system is introduced in detail. Then with these work, the selection of hardware and software design is completed.(2)According to Blob analysis procedure, a comprehensive theoretical analysis of image acquisition, preprocessing, segmentation, morphology, connectivity analysis and feature extraction techniques is carried out; And with the comparison of the image processing results of various image processing algorithms, the image processing method of this system is determined.(3)Based on Halcon software, the glass bottle, bottle, bottle bottom defect detection experiments is carried out respectively, and detection methods can be adjusted according to the different characteristics of the detection site in order to obtain an ideal detection effect.(4)On the basis of visual guidance on, the grasping experiment of defective bottle is finished by using FVRTS platform six degrees of freedom manipulator path planning and control experiments, and finally, the removing function of defect bottles is achieved.(5)The human-computer interaction system for glass defects detection is designed based on VC++, the the detection and control functions of the system are achieved by calling machine vision algorithms and robotic manipulation instruction.The interface of system is intuitive and simply operated, which greatly facilitate the users.The system has been successfully applied in the oil filling system. And to some extent, it reduces labor intensity and avoids raw material waste and equipment contamination.so that the efficiency of the original filling line has been greatly improved. Meanwhile, the main results of this paper,removing defective detection and the other colors, shapes and material object defects, also has certain reference significance.