If the industrial robot is an extension of the human hand and the vehicle is an extension of the human leg, then machine vision is equivalent to the extension of human vision on the machine, and is a necessary means to achieve industrial automation and intelligence.
As a key technology that brings visual functions to robots, machine vision is widely used. From industrial vision to computer vision, from human-computer interaction to autonomous driving, from virtual reality to automatic object recognition, machine vision can play an important role.
Industry 4.0 is inseparable from intelligent manufacturing, and intelligent manufacturing is inseparable from machine vision. If the industrial robot is an extension of the human hand and the vehicle is an extension of the human leg, then machine vision is equivalent to the extension of human vision on the machine, and is a necessary means to achieve industrial automation and intelligence. Machine vision has the advantages of high automation, high efficiency, high precision and poor adaptability to the environment, opening a new horizon for industrial automation in China.
Machine vision + robot
The application of machine vision on robots is very necessary. Both industrial robots and service robots need robot vision as their own "eyes", but industrial robots and service robots have different requirements for robot vision.
The difference between machine vision and industrial robots is the application scenario. In the environment of industrial robots, the scenes seen by computer vision are relatively simple, but the application scenarios of service robots are much more complicated. Beyond the scene is the precision aspect. In industrial robots, the recognition accuracy of computer vision must be below the millimeter level, including static identification accuracy and dynamic identification accuracy. Static refers to the relative state of the camera or observation object in a static state. The accuracy depends on the resolution of the camera, including whether the edge of the item is clear, and whether the differentiation is clear. At this time, the accuracy can even reach the micron level. The accuracy of the service robot is not very high, and may be above the centimeter level.
Machine vision + food and medicine
As we all know, food and medicine are related to the health of human beings. If it is a threat to human life because of the quality of medicines, it will be a big disaster. Therefore, various pharmaceutical manufacturers, especially the world-famous large manufacturers, have given great attention to the entire production process of the drug and even the packaging of the latter stage. In the production and packaging process of food and medicine, whether it is blister packaging, liquid filling, or the back cover, labeling, coding, and final boxing detection, machine vision technology can exert its power.
Using the machine vision method, it is possible to quickly and accurately detect whether the pharmaceutical packaging is intact. By setting the image sensor, the image information of the packaged object is obtained, and each medicine or bottle is detected by a preset area parameter. In contrast, the damaged granules or the missing bottles will be detected and passed correctly.
Machine vision + security
Security is the most market-oriented application area of artificial intelligence. We can predict that in the next 5-10 years, the security industry will inevitably develop into a stage of full-scale explosion of artificial intelligence applications, and the turning point of security machine vision has arrived.
In the next 4-5 years, we will enter an era of rapid development of artificial intelligence in the field of security. In 2016, we have basically completed the adaptation of scene intelligence, and in 2017 and 2018, we will enter a foundation based on deep learning. In the intelligent phase, by 2019 and 2020, we believe that we will fully enter the stage of digital intelligence. After the digital intelligence phase, the artificial intelligence of the security industry will be fully integrated with the industry-wide, all-IT big data service platform. This is the next development of the entire industry that we can predict.