Hello, welcome to Shenzhen Fasttest Electronic Technology Co., Ltd. official website

+86 138-23138575

Current Position:Home > Fasttest News > Industry Information

  • News Center

  • Recommended information

    The application and development trend of machine vision system

    Update time:2019-06-05 17:07:29

    Machine vision inspection


      Machine vision systems are highly reliable in high-speed, subtle and repetitive manufacturing processes, and are therefore widely used in manufacturing and manufacturing companies to perform repetitive inspection tasks in high-volume production processes. The use of machine vision in quality inspection accounts for nearly 80% of the total industrial use. The largest industries used are: car, pharmaceutical, electronics and electrical, manufacturing, packaging, food, beverages, etc. Machine vision detection for fault contact non-destructive testing, compared with traditional detection methods, it has irreplaceable superiority, and thus has been widely used. Use the linear array CCD to match the one-dimensional motion of the package to obtain the policy picture, and then process the picture by the computer, which can detect the omission and correctness of the date number and other information; use the bright light as the illumination source, use the area array or line array CDC is used as a detector material for the outline of rebar profile. It performs on-line measurement of several parameters of hot-rolled rebar, and it is also used in the detection of defects in various products.


    The development of the machine vision system


         In recent years, the development of computer vision (ie machine vision) has been manifested in the following three aspects:


    1. According to several methods, the accounting visual accounting theory system is one of the research guidelines for the perfect computer vision. It is to make the machine can sense the information of the object in the three-dimensional environment, including its shape, orientation, posture, motion and so on. Since the mid-1990s, the computer vision community will correspond to the projections, affine and Euclid, and systematically introduce visual accounting methods, which correspond perfectly to the object in the visual system. The detailed depiction reduces the requirements for understanding the parameters of the camera system in the vision system of the computer, and improves the robustness of the system to noise.


    2. Machine learning methods are receiving more and more attention. All categories of pattern recognition always exist according to the two branches of structure and calculation. If it is said that the three-dimensional structure of the object and its motion is mainly based on a few kinds of computer vision, it belongs to the structural method and has been studied systematically; the calculation method in the computer vision is better used for the bottom layer of the picture. In addition to handling, it has always been imperfect, not to mention systematization.


    3. The research on the use of many specific categories has been deepened, and the large-scale use system has gradually become commercialized. With the rapid development of the performance and price ratio of the computer at that time, the commercialization of real-time use systems of many specific categories of computers has become possible. For example, fingerprint, iris, face, voice and other recognition skills, behavior recognition skills and motion tracking skills, multi-camera fusion skills constitute a visual monitoring system for information security, intelligent transportation, anti-terrorism theft, identity identification.