Analysis of the hottest machine vision application

  • Detail

Market analysis of machine vision applications (Part 2)

PC assisted quality control (QC)

those pharmaceutical manufacturers who want to investigate quality problems in depth can choose the last 100 unqualified products to make image files and store them. Operators or factory inspectors store these documents in PC

Clarke believes that some machine vision software packages, such as Cognex's in sight system, include tracking functions. Using the tracking function, you can find out who made changes to the system, what kind of changes, and when. The software system with 21 CFR Part 11 packaging standard option will track the history of machine vision inspection, and the system will also record some events that occur in the inspection process, store these recorded information and call it. "

most high-end vision systems can track detection records and store these recorded information in the PC database. Some companies are tracking and storing this information for better tracking performance. Using the visual system, we can trace and reproduce the events that have occurred, check the images, find out the events that have occurred, and then solve the problems. In this regard, insiders generally believe that Cognex's in sight system can be used to classify and describe qualified/unqualified products and the characteristics of unqualified products, store images related to unqualified modes, and guide users in selection

of course, the adoption of machine vision system also has its disadvantages, because the more complex machine vision system is very expensive, and it will be difficult to pass the verification due to the verification cost and other reasons. As a result, pharmaceutical manufacturers are unwilling to replace the system they think has been tested and effective

a problem faced by pharmaceutical manufacturers is that the ink on the product trademark containing bar code, product identification and product expiry date may become blurred and illegible. Using the machine vision system, the product information stored in the camera memory can be used to inspect the trademark bar code on each product. The system will prompt the operator to check the bar code of the product, check whether the color band is dry or find out the reason why the bar code is blurred. Usually, when there is no such system, the problem of bar code blurring will not be noticed before the product is sent to the user

hi tech pharmaceutical uses DVT's machine vision system to upload detection data and images to the company's Ethernet. Managers can see the situation of six production lines on the office computer. Similar to the above products, banner engineering's machine vision system provides real-time dynamic video for the PC of operators or quality managers. If the production line stops running for some reason, or a product is incorrectly identified as unqualified, the operator can see this situation through the video and make changes

it is generally believed that it is the most common practice to use the PC in the office of the factory manager or inspector to control the system, and it is inappropriate to place the PC on the production line. If PC is used for control, only software training is needed for personnel to make changes to the system

as the creator of machine vision system, Clarke believes that pharmaceutical manufacturers are still in the early stage of understanding the application of machine vision and its potential use in improving products, automatically recording relevant data, reporting and storage. Identify new uses

although the most common use of machine vision is product detection, some observers believe that other uses of machine vision, especially the identification of materials, products and trademarks, will bring benefits and value to the development of pharmaceutical manufacturers

among the various uses of machine vision, compared with measurement, detection or guidance, pharmaceutical manufacturers are most concerned about clear recognition. It is very important to record bar codes, read characters, identify materials, check whether the number on the identified package is correct and other problems on the trademark, etc. it is absolutely necessary to ensure the trademark of the product and the words on the trademark

for the successful application of machine vision system, other crucial issues are the distance between the machine vision system and the product or package, the background and the relative position with the camera. Next is the packaging design and labeling. The goal is not to read blurred fonts, so the position of the camera is also an important factor. Of course, the position of other machines is also important

in order to make the system work fully, the factory management must participate and make efforts to ensure the successful application of machine vision technology, and must understand why the successful application of machine vision requires certain necessary conditions, otherwise, the fact is to tell them, "you have only completed 90% of the output, and you are throwing away 5% - 15% of the products". Sounds a little scary, doesn't it? Don't worry, because for most companies that install machine vision systems, success is measured by minimizing the number of errors. Overall, the impact of machine vision systems on production is actually only 0.25%

comments by Tan Yonglin of Germany's sick optoelectronics company: machine vision has a wide range of uses. In addition to its main uses (inspection, measurement, recognition, positioning), there are many additional functions, such as using machine vision to establish a product database, realize product traceability, and monitor production conditions in real time to prevent the production of large quantities of defective products

in addition to its application in the pharmaceutical industry, machine vision can also realize related functions in other industries. Many enterprises may think that the establishment of a complex machine vision system will increase a considerable part of the cost, perhaps in the initial stage of investment, but the machine vision system will bring these enterprises more than superficial benefits. It can not only improve the product qualification rate and reduce the cost of manual inspection, but also improve the visibility of enterprises and increase the order volume in the long run, which will benefit enterprises a lot, Especially in a large manufacturing country like China

Xing Jijun, product manager of banner engineering company, commented that the application of machine vision in China is still backward compared with developed countries, even with some developing countries, such as Brazil. From the understanding of detection, to the application of single machine, and then to the integration of vision system into the whole production line, in general, the application of machine vision in China is still in its infancy (excluding imported production lines)

it is too early for China to discuss the application and integration of visual inspection data at a higher stage, because too many basic application development are far from complete. Every time a new detection application is introduced, it needs to carry out a long-term and a large number of reliable on-site verification for users to prove that the user's investment will produce a corresponding return, or how to save a lot of costs for users

only take the pharmaceutical industry as an example, such as pharmaceutical packaging. In China, it can only be said that the inspection of blister packaging is relatively mature; In others, such as label phase III detection, digital anti-counterfeiting confirmation, packaging quantity confirmation, etc., a large number of detection technologies have not been applied yet. In the long run, there is no doubt that the domestic market is very considerable, but the initial stage now will be long-term and require a lot of investment. Therefore, we hope that the majority of visual product providers and system integrators can have a clear understanding

at present, what we should focus on is to promote the application of machine vision through various occasions or opportunities, expand the publicity to end users, tell them what we can do and what effect we can achieve by doing so

Hu Jinlin, Deputy General Manager/chief engineer and manager of system research and Development Department of Beijing micro vision new era technology Co., Ltd.

comment: with the development of electric vehicles in recent years and the continuous promotion and application of domestic manufacturing industry, the connotation of machine vision has been gradually recognized and accepted by most people in the industry. However, there is no unified understanding of the extension of machine vision in the industry

in fact, machine vision is a means. It can be said to be a sensor or a weighing instrument. In fact, it is a standard that can be controlled by computer and is not disturbed by human factors. It is a dangerous trend to unilaterally expand the extension of machine vision, which has just started in China. In a sense, it will even hinder the development of machine vision itself

for example, the application of database, the traceability of product quality, the real-time monitoring of production conditions, etc., for example, if a glass bottle production line has broken bottles for many times within one hour, the system needs to notify the production quality control department to take corresponding measures, such as adjusting parameters, adjusting ingredients, stopping for maintenance, etc., all these processes are no problem. The problem is that the adoption and introduction of these means are not the changes brought about by machine vision itself

before there is no machine vision, the design of the production line also needs to introduce the management of the database. Therefore, it also needs to change the internal structure of the spring (change the size of the center hole, etc.) to monitor the production situation and effectively control the quality. However, the means of detecting glass bottles may be manual observation with the naked eye. Therefore, machine vision is fundamentally just the progress of detection means, that's all. Machine vision itself will not bring about the revolution of the production line, let alone unilaterally expand the extension of machine vision

Yang Wei, Senior Application Engineer of Cognex company, comments: the code reading with visual system mentioned in the article is connected with the centrally controlled database, which has been applied abroad for at least ten years. The reason why machine vision cannot be popularized in China is that the automation level of domestic production lines is generally very low. In order to improve its application status, the first step is to improve the technical level of domestic equipment manufacturers and make them close to the foreign level as soon as possible; The second step is that the visual industry should cooperate closely with equipment manufacturers to form a "win-win" situation

looking at the domestic visual industry, it is easy to find several problems: first, the lack of domestic visual products, especially software and a complete set of visual systems; 2、 The agency link is too long. At present, there are too many companies engaged in agency (supply exceeds demand), which leads to the agency companies entering the stage of vicious price competition too early. In addition, a considerable number of domestic OEMs and end users regard their own visual development ability too highly, resulting in the unreasonable situation that agents directly deal with end users through system integrators; 3、 System integrators are widely used in calendering or extrusion to produce soft film, artificial leather, cables, pipes, soft profiles, etc., and OEM customers are seriously insufficient. According to the current situation of the Chinese market, system integrators are an important link to ensure the normal operation of the vision system on the end-user production line, and OEM customers are the real driving force of the entire vision industry. The lack of these two links is the main reason for the difficulty of the vision industry in China at present; 4、 A large number of end users need to be developed. At present, the practical application projects of domestic vision are mainly concentrated in the tobacco and electronic industries, and the three industries of electronics, automobile and packaging will be the most powerful direction for the development and application of machine vision in China in the future

source: Port spot

Copyright © 2011 JIN SHI