Give the machine a pair of insight

Giving machines a pair of insight

vision is an important means for human beings to recognize the world. The information obtained through vision accounts for more than 80% of the external information obtained by human beings. Human beings perceive the information of the external world through vision, and then make corresponding judgments and actions through the brain. It has long been a human dream to make machines have visual functions like humans. With the development of computer, artificial intelligence, optics, electronics and control technology, especially the rapid progress of CCD, DSP, image processing, pattern recognition and computer vision technology, a new technology, machinevision, is turning this dream into reality

machine vision technology enables the computer system to recognize the surrounding environment through images, make corresponding judgments, and control the behavior of the system according to the set rules. This makes the machine system have the ability of human vision and judgment. The machine vision branch of the American Association of manufacturing engineers and the roboticindustriesassociation define machine vision as "machine vision is a device that automatically receives and processes the image of a real object through optical devices and non-contact sensors to obtain the required information or control the motion of robots". In short, machine vision is to use machines to replace human eyes and brain to complete various measurements and judgments. Machine vision first processes image (vision) information, and controls the next action according to the processing results. A machine vision system is a system that can automatically obtain the image of the target object, process, analyze, measure and judge various feature quantities of the obtained image, and make qualitative analysis and quantitative interpretation, so as to obtain the understanding of the target object and make corresponding action decisions. The functions of machine vision system include object location, feature detection, defect judgment, target recognition, counting and motion tracking

the development of machine vision

machine vision technology appeared in the mid-20th century. It was initially used for the analysis and recognition of two-dimensional images, such as optical character recognition, workpiece surface detection, aerial photo interpretation, microscopic image interpretation, etc., which basically belongs to the scope of image processing and recognition. Bert published the famous paper "machine perception of three-dimensional objects" in 1965, which expounds the possibility of using the means provided by computer science to obtain three-dimensional information from two-dimensional images, and pioneered the research of three-dimensional machine vision for the purpose of understanding three-dimensional scenes. By the 1970s, some machine vision application systems had appeared one after another

the visual computing theory proposed by Marr in the 1970s is a milestone in the theoretical research of machine vision, which has become the basic theoretical framework of machine vision research. Since then, various theoretical researches on machine vision have developed rapidly, and new methods and theories in the fields of image processing, pattern recognition, computer vision and so on have been emerging, which has laid a theoretical foundation for the development of machine vision. At the same time, the progress of computer, electronic technology, CCD, DSP, optics, machinery and other hardware technology is also changing with each passing day. From light source, camera, image acquisition card to computer and embedded system, almost every relevant link of machine vision system has seen a substantial improvement or even a qualitative leap in performance. Higher resolution images, faster processors, better light sources, along with the reduction of cost and volume, all of which have prepared a solid hardware platform for the development of machine vision

a typical machine vision system is computer centered, which is composed of vision sensor, image acquisition system, image processing system and control mechanism. The visual sensor and image acquisition system collect images, and convert optical images into digital images. The image processing system processes and analyzes the images, obtains various required information, and makes corresponding action decisions. Machine vision system can replace human vision system to a considerable extent, and even has higher resolution and faster speed than human vision. It can obtain the information of wavelength range that human eyes cannot see, and complete various tasks such as rapid detection, positioning, recognition, navigation, measurement, etc

characteristics of machine vision system

machine vision combines the high-speed computing ability, repeatability and rapid response ability of computers and electronic systems with the highly intelligent, abstract and judgment ability of human vision. It is an effective way to realize the precision control, intelligence and automation of instruments and equipment, and can be called the "machine eye" of modern industrial production. Compared with traditional sensors and manual operation, it has unique advantages:

● high precision. Using high-resolution optical imaging system and image acquisition system, machine vision can achieve much higher detection accuracy than human eyes

● non contact. Machine vision system can realize non-contact measurement and detection, and will not cause any damage and change to the measured object

● continuity and stability. Machine vision system can avoid the fatigue and instability of human eyes, and realize long-term continuous measurement, recognition and analysis

● wide spectral response range. By using spectral bands invisible to human eyes, such as infrared, ultraviolet and X-ray, the recognition and observation of special features are realized, and the visual range is expanded

● fast. Combined with the high-speed computing ability of the computer and the high-speed response ability of imaging devices, the machine vision system can realize the capture, recognition, inspection and corresponding actions of large quantities of high-speed moving objects, which is 2. It is difficult for people to place the measured samples on the friction table. This characteristic is of great significance for military and modern manufacturing

● work in harsh environment. It can replace people to work in harsh environment

● combine with massive knowledge and rule base for rapid inspection and judgment. The obtained image can be quickly compared with a large number of pre stored rules and standards

● rapid quantitative analysis. Machine vision system can not only simulate human vision to make judgments, but also has the ability of rapid and accurate quantitative measurement and analysis that human vision does not have, which is of great help to the regulation and control of the production process

● certain flexibility. By adjusting the software and parameters, it can adapt to the changing objects in a certain range, which is better than the traditional optical, electronic and mechanical sensors

● save labor costs and protect the health of staff

application field of machine vision

the unique advantages of machine vision make it applied in many fields, and even play an irreplaceable role, which greatly improves the technical level of these industries

1. military. Military field is the most eager and sensitive field to new technology, and machine vision is no exception. The earliest vision and image analysis system was used for reconnaissance image processing and analysis and weapon guidance. The guidance based on visible light and infrared ray in modern high-precision guidance weapons is a complete machine vision system. Through sensor imaging and missile borne computer analysis, the target is identified in the complex background and guidance instructions are given. In the Gulf War, the effect of missiles flying into the building from the window is created. Pilotless aircraft and unmanned combat vehicles also use the vision system of aircraft to conduct environmental analysis, path guidance and attack guidance. Even in the rear, dangerous magazine handling operations can be carried out with the help of a manipulator equipped with a vision system. Robot soldiers who may appear on the battlefield in the future will undoubtedly be equipped with a sharp vision system

2. Semiconductor/electronics. About 40% - 50% of the applications of machine vision are concentrated in the manufacturing of semiconductors and electronic products. In this field, machine vision is widely used in surface detection, automatic positioning, character detection and recognition, and size measurement. Different from the "high-end" technology in many industries, the application of machine vision is becoming more and more popular and diversified in the field of semiconductor/electronic manufacturing. In the face of higher and higher integration, more and more dense lines and components, and faster and faster manufacturing speed, machine vision detection, positioning, and guidance is almost the only choice. At present, the application of machine vision can be seen in the front and back processes of semiconductors, the manufacture of electronic components and terminal electronic products; In the IC manufacturing industry, such as wafer carving and wafer cutting, positioning and testing are needed; In terms of packaging, it mainly focuses on the quality inspection of packaged devices, mainly including some measurement and detection functions, and the character quality inspection after laser marking may be the most well-known application. PCB manufacturing is a field that has developed rapidly in recent years. In addition to the traditional automatic positioning (punching machine, etc.) and automatic detection (printing machine, etc.), AOI (automatic optical detection) based on machine vision technology has become an indispensable detection equipment for high-precision PCB board making, which is used for board making negative inspection, PCB bare board inspection, component post installation inspection, etc. In the manufacturing of electronic components, the use of machine vision for detection and measurement, in addition to the original measurement and inspection of ordinary components, the detection of LCD screen will become the focus

3. medical care. It is used for medical image analysis, automatic classification and counting of blood cells, chromosome analysis, cancer cell recognition, endoscopy, etc

4. pharmacy. Appearance quality inspection, size inspection and quantity statistics of drug shape and thickness in drug production

5. Post Office. The postal code is automatically recognized by the machine vision system to realize the automatic sorting of letters

6. Machinery manufacturing. In the manufacturing industry, machine vision technology can be effectively used for product quality inspection, rapid inspection of product surface defects, inspection of product manufacturing dimensions. Machine vision technology can realize three-dimensional measurement and detection, and can also be used for assembly guidance

7. Printing money and coinage. There are strict requirements for the printing quality and quantity of banknotes. Machine vision technology can replace people to carefully check the printing quality of banknotes to ensure the printing quality of banknotes. It can also automatically recheck the number of banknotes and quickly count the number of banknotes without contact

8. Intelligent transportation. Machine vision can be used for traffic flow analysis and license plate recognition. In the electronic police system, visual detection technology is used to analyze vehicle behavior, judge violations, record images, recognize license plate numbers and record

9. Logistics. Identify bar codes, labels, items, sort items, and identify container numbers

10. Tobacco industry. Machine vision can be used for tobacco grading, foreign matter removal and packaging quality inspection

11. Sorting of agricultural products. Using machine vision technology, agricultural products can be quickly and automatically classified and sorted. For example, using the color selection technology in machine vision can remove the black particles and foreign matters in rice, so as to improve the grade of rice, which can bring great benefits to enterprises for long-term preservation