Artificial intelligence has become the core driving force of a new round of industrial reform and is having a wide impact on the world economy, social progress and human life. Machine vision will play an important role in the tide of industrial and consumption upgrading. With the introduction of new infrastructure and the development of 5g and other new technologies, machine vision will release huge market potential.
Machine vision is the starting point of digitization in the industry
“Vision” is an important source of information for human beings to perceive the world, and machine vision is a technology to simulate human visual functions through computers, so that computers can obtain relevant detailed information and understand it. Machine vision has the characteristics of low cost, wide application, high accuracy and can well adapt to the poor environment. It has served the industry for more than 30 years. At present, object detection, segmentation and recognition based on a large number of data sets have made outstanding contributions in the fields of smart home, security, industrial detection and so on.
In recent years, with the increasing requirements for image detail accuracy and processing speed in applications, the industrial camera with 10 Gigabit network interface has been replaced by the camera with Gigabit network interface and put into application. Relying on years of deep cultivation and accumulation in the field of machine vision technology, baomeng launched an industrial camera with leading technology level in the industry: LX series 10 Gigabit network interface camera, providing a series of product solutions from 600000 pixels to 65 million pixels. According to Bao Meng, the system uses the same chip XT Series 600000 pixel camera, and the maximum frame rate reaches 1578 frames.
With the semiconductor industry entering a new business cycle, software design companies are springing up. The in-depth learning and algorithm research and development are inseparable from the support of hardware. Chen Tian introduced that baomeng launched an LXT camera integrating JPEG image compression technology to reduce the burden of other hardware and software. “By using baomeng compressed part camera, the size of pictures can be effectively reduced, which reduces the difficulty of PC storage and the requirements for the performance of the system itself. This is also a camera with our unique advantages.”
In combination with the current popular deep learning, edge computing of artificial intelligence, etc., baomeng has recently developed an AI camera integrating software and hardware – ax series AI intelligent camera. Ax smart camera adopts NVIDIA processor and special algorithms in the field of artificial intelligence to form a compact vision system. Chen Tian added that this camera does not need to use other additional software and hardware in the process of deep learning algorithm development and edge computing development, which effectively reduces the system cost. In addition, when this smart camera is deployed to VR and edge computing application scenarios, all its operation results can be connected with the cloud and Internet of things center to directly, quickly and efficiently output executable results without additional image processing.
The future machine has computing power, thinking ability and boundless vision
Machine intelligence needs higher data quality, and it will become a trend to improve the perception and understanding ability of its visual system. Civilian industries, such as intelligent monitoring and vending machines, need the support of face recognition technology. Behind it is the combination of camera architecture and intelligent algorithms. Baomeng proposed the concept of compact and integrated camera, which makes use of the advantages of AI in deep learning to build a smaller and more flexible vision system.
A single algorithm can not solve the small and numerous defect detection of electronic parts in the industrial field. Machine vision introduces deep learning artificial intelligence, which can solve the problem by automatically judging the defect category in the production process. For example, in the pharmaceutical industry and other application scenarios, the product traceability project is huge. The AR camera of machine vision can make the traceability more intelligent and save time and labor cost.
For the future development trend of machine vision, Chen Tian believes that with the continuous promotion of international intelligent manufacturing strategy, machine vision will continue to be hot in the future market and realize technology upgrading due to the addition of new technologies. The details are as follows:
Higher speed. With the improvement of the application scenario scheme, the production capacity will increase, and the requirements for visual judgment speed will increase. High speed has become the focus of competition of machine vision technology.
Combined with smart factory and industry 4.0. At this stage, for the upgrade from gigabit network to 10 Gigabit network, the camera not only needs to improve the bandwidth in terms of technology, but also needs to understand the high-speed application scenarios. On the other hand, unmanned factories are inseparable from machine vision. In the process of promoting industry 4.0, the birth of 25X and 40x cameras will be connected to the Internet of things in the form of network.
From two-dimensional to three-dimensional and multi-dimensional. With the development of technology towards higher speed and greater bandwidth and the deepening of various applications, 3D camera will play an important role, accompanied by the technical challenges of higher precision and faster response time. For example, some new fields brought by cutting-edge technology, such as UAV and server man, are promoting the transformation of this technology.
Testing of special materials. Scratch detection on special materials such as plastics and transparent films is one of the directions of machine vision technology in the future.
For special application scenarios, it is necessary to build a multi domain knowledge base. In addition to the mature application of machine vision in the industrial field, the advantages of machine vision in some non industrial fields are also very obvious, and there is still much room for development and application. Now there are also preliminary cases introduced in the market, subdivided into new industries, such as forest fire prevention, aircraft runway foreign matter detection, etc. In the future, there will be more application scenarios such as smart water conservancy, smart oil field, smart network, smart manufacturing and holographic community.