More than 75% of information acquisition comes from vision, and machine vision is a further improvement of human vision. Machine vision detection equipment is undoubtedly more powerful than human eyes. In many environments, human visual changes have become increasingly difficult to meet the requirements. For example, high-speed and high-precision production lines in electronic manufacturing and other industries tend to be fast, with small workpieces and many non-standard parts. If a large number of workers rely on labor, objective factors such as fatigue that workers must avoid will bring very high errors, resulting in unstable product quality. Therefore, it is very important to replace human eyes with machine vision detection equipment.
With the inevitable increase of labor costs and the exposure of the problem of workers' instability affecting quality, human uncertainty has become the bottleneck of the development of labor-intensive manufacturing. At the same time, another dilemma of China's product manufacturing model is also emerging, that is, no one can do high-end products, while low-end products are of poor quality, easy to overstock and cannot be sold, and the industrial structure is very unreasonable.
The fundamental way out of these difficulties lies in transformation and upgrading, and at the same time, improving the degree of intellectualization. Using artificial intelligence combined with manufacturing, using artificial intelligence + robots to reduce artificial, and reducing the problems of unstable quality and poor quality caused by artificial, is undoubtedly a good way to promote the transformation and upgrading of manufacturing.
New changes brought about by machine vision. Machine vision is the most basic unit technology and a non-contact sensing device. In the actual manufacturing industry, it is mainly used to automatically obtain and interpret the image of real things, and obtain the image for machine recognition, feedback and control the manufacturing process.
At present, the typical implementation technology of machine vision is to use industrial camera. Through the integrated motion control unit, industrial camera can collect the moving image to the computer, and then return to the machine through computer recognition and feedback to perform action adjustment. In the past, the automatic equipment was implemented according to the existing predetermined instructions, but after adding visual feedback, it can also make real-time adjustments quickly according to the environmental perception, which improves the intelligence of the equipment and the efficiency and accuracy of implementation.
From the perspective of production lines and intelligent factories, the addition of machine vision can well improve the manufacturing level. At present, many factory equipment have achieved basic automation, which is manifested in the mechanical repetitive movement according to human programming. However, if the machine is added with vision, the equipment can actively adjust according to different environmental conditions, so as to realize intellectualization on the basis of automatic equipment.
It is very feasible to replace the artificial industry with machine vision technology. At present, machine vision has been widely used in the Pearl River Delta. As the Yangtze River Delta and Pearl River Delta are the concentration of electronic manufacturing industry, the manufacturing industries such as processing electronic components, PCB boards and mobile phone assembly are fine and relatively small, and there are a large number of strong lights as backlights, which do great harm to human body, these scenes are undoubtedly very suitable for the application of machine vision.
In the future, the development of machine vision is also expected to replace labor in a large number. With the help of machine recognition and information processing, there are still many scenes that can be developed. Machine vision can also improve the product consistency and quality stability of workpieces. Machine vision can well complete key manufacturing processes such as inspection and improve the overall efficiency of the production line. For example, machine vision has more room for development in the detection of workpiece length, width and thickness, as well as in the detection of screen stickiness and scratches in the mobile phone industry.
In addition, machine vision can also achieve more accurate positioning, so that the robot can achieve visual positioning and error compensation in the process of grasping. In OCR recognition, machine vision also has great potential, such as recognizing board characters, judging omissions, and judging whether the installation is correct.
However, the application scenario of single machine vision is also very limited. Machine vision has developed for 10-20 years now, and it has been less used in the past. In recent years, with the continuous reduction of hardware costs such as cameras and light sources, as well as the improvement of computer processing capacity, artificial intelligence technology has entered the industrial field more and more widely.
With the development of software and intelligent technology, the application of machine vision has increased significantly. In the electronic manufacturing equipment industry, machine vision has almost become a standard configuration, but in specific applications, traditional machine vision algorithms are often used, and machine vision algorithms with AI are still rarely used. Therefore, the field of machine vision + artificial intelligence still has a very large market space.
Using artificial intelligence + machine vision to do machine recognition has obvious advantages. The most direct embodiment is that it can greatly increase manufacturing flexibility.
Flexible manufacturing often shows that robots and other industrial equipment can show better adaptability to different positions, lighting, and complex environments. For example, for the color detection of objects, if traditional machine vision is used, the subtle recognition of different colors is not high. If there are slight differences such as depth and color deviation in a certain color, traditional machine recognition is easy to make mistakes, and the method of artificial intelligence + machine vision can be more accurate to improve the detection rate and accuracy.
Using AI machine vision for detection can also facilitate human-computer interaction and improve the convenience of detection. In the past, to realize the specified position detection of precise points, lines and surfaces, people often needed to click the mouse and tell the vision to detect which circle and which line. But now, after adding AI, the circuit board is placed on the production line, and the problems in all aspects of the original can be automatically identified by means of reinforcement learning, so as to complete visual + intelligent calibration, intelligent dimension measurement and real-time position judgment, Efficiency and human-computer interaction intelligence have been greatly improved.