Briefly describe the requirements for machine vision in smart factories

 When choosing a machine vision system , the advantages of traditional industrial intelligent vision equipment are small size, high integration, and ease of development and use; the advantage of embedded machine vision systems lies in their flexible configuration and high scalability. In the construction process of smart factories, the requirements for machine vision systems are to not only reduce costs, but also meet the rapidly changing needs of vision applications.

What are the needs for machine vision in smart factories?
Performance and processing power. How to improve production capacity, system performance and processing capacity (throughput) play a key role. As far as general machine vision systems are concerned, high resolution and high frame rate (frames per second) are like fish and bear's paw, you cannot have both. In general practical applications, it is usually a combination of high resolution but low frame rate or low resolution but high frame rate. If you want to have both, the only way out is to use a high-end CPU processor to make up for the multiplication of resolution and frame rate. How to obtain the best processing performance at a reasonable cost is what system developers are concerned about.
Production line environment. The factory environment is usually harsh. For example, in beverage production and packaging production lines, the system may be in direct contact with liquids. In the environment of tool machining, it is a harsh environment full of cutting workpieces. If the machine vision system must be deployed nearby in a harsh production line environment, then choosing products with waterproof and dustproof capabilities can meet this requirement.
Multiple groups of production workstations. In a factory environment, when a finished product is put on the market, it must go through different workstations from the manufacturing of components, the picking and placing of semi-finished products, quality inspection to packaging for shipment. For example, the CNC machine is responsible for the turning processing of components. After the industrial robot picks up the workpiece and uses the industrial camera to position the workpiece, it starts cutting the workpiece. After completion, it enters the inspection station for defect detection; the finished product passes the test. Scan and read shipping barcodes in the packaging area. How to facilitate the integration and communication between systems among multiple groups of production workstations is a key to whether the factory is intelligent.
Software development environment. The difficulty and integration of software solution development are a major concern for all engineers who introduce intelligent systems, and are often the most important factors that determine the success or failure of the project. How to shorten development time and reduce system development costs is an important key.
Current status of the development of machine vision in China.
The machine vision industry in China started relatively late, and its concentration is not very high. At the beginning, it mainly acted as an agent for foreign brands. In recent years, many dealers have begun to develop products independently, but there is still a certain gap between them and foreign countries in terms of industry distribution, channel distribution, and mature automation products. The quality and technical content of domestic machine vision's relatively mature automation products are low, and the market is far from saturated.
Machine vision companies can be roughly divided into first-tier development manufacturers, secondary development manufacturers and product agents. Domestic machine vision companies are mainly foreign machine vision product agents and system secondary development manufacturers. At present, there are more than 100 foreign brands entering my country's machine vision market, more than 200 local companies in my country are responsible for sales agents, and there are more than 50 professional system integrators. There are very few truly professional bottom-level machine vision manufacturers in my country, and there are deficiencies in local machine vision system manufacturers and machine vision system component manufacturers.
The development trend of machine vision in China
, Industry 4.0, is inseparable from intelligent manufacturing, and intelligent manufacturing is inseparable from machine vision. Machine vision is a necessary means to achieve industrial automation and intelligence, which is equivalent to the extension of human vision on machines. Machine vision has the advantages of high automation, high efficiency, high precision and adaptability to poor environments, and will play an important role in the realization of industrial automation in my country.
Visual image technology needs to focus on building four core capabilities:
first, intelligent recognition. Massive information converges quickly, and accuracy and reliability are key to finding key features from a large amount of information.
Second, intelligent measurement. Measurement is the basis of industry and requires accuracy.
Third, intelligent detection. On the basis of measurement, multiple information and multiple indicators are comprehensively analyzed and judged. The key point is intelligent judgment based on complex logic.
Fourth, intelligent interconnection. Massive data of images are collected and interconnected at multiple nodes. At the same time, data on personnel, equipment, production materials, environment, processes, etc. are interconnected to derive innovative capabilities such as deep learning, intelligent optimization, intelligent prediction, etc., truly demonstrating the power of Industry 4.0.
Machine vision related products
A typical industrial machine vision system includes: light source, lens, camera (including CCD camera and COMS camera), image processing unit (or image capture card), image processing software, monitor, communication/input and output unit wait.
(1) Core components of machine vision:
smart cameras: black and white smart cameras, line scan smart cameras, color smart cameras, CMOS smart cameras, ID readers, etc.; boards
: black and white capture cards, image compression/decompression boards, color capture cards cards, etc.;
software packages: image processing software, machine vision tool software;
accessories: industrial cameras, CMOS cameras, CCD cameras, area scan cameras, line scan cameras, infrared cameras, 1394 interface cameras;
industrial lenses: FA lenses, high resolution Lenses, image scanning lenses, condenser lenses, telecentric lenses, etc.;
Light sources: LED light sources, UV lighting systems, infrared light sources, fiber optic lighting systems, etc.;
Auxiliary products: sensors, calibration blocks, gratings, washers, wiring and connectors, Power supply, baseboard.
(2) Machine vision accessories:
image processing system: optical text, recognition system, automation/robotics technology, infrared imaging system;
machine vision integration: character processing and recognition system, automation/robotics technology, infrared imaging system, tobacco, banknote printing , Electronic assembly, quality inspection, automatic recognition (OCR/OCV), measurement, intelligent vision, surface inspection, printing, packaging, complex industrial object vision online, automobile manufacturing, license plate, intelligent transportation, biometric identification, monitoring, medical testing, Optical inspection and other systems.
Since the machine vision system can quickly acquire a large amount of information, is easy to process automatically, and is easy to integrate with design information and processing control information, in the modern automated production process, machine vision systems are widely used for working condition monitoring and finished product inspection. and quality control.
However, machine vision technology is relatively complex, and the biggest difficulty is that the human visual mechanism is not yet clear. People can use introspection to describe the process of solving a certain problem, and then use computers to simulate it. But even though every normal person is a "visual expert", it is impossible to use introspection to describe his or her visual process. Therefore, building a machine vision system is a very difficult task.