A brief discussion on three development trends of machine vision

 In the machine vision industry, a typical 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 and other hardware accessories. So what are the development trends of machine vision? Below, the editor of Chengdu Xinxiwang will explain to you:

(1) Make the machine vision product interface easy to use

The user interface of machine vision systems will attract more and more attention. The user interface is often the last consideration and needs to be built with ongoing contributions from algorithm designers. The results may be meaningful to designers, but they take a long time to learn for users. With IPD technology, we start with the interface and build in a way that makes our vision systems easy to use with acceptable human factors.

(2) Make machine vision tools simple-to-use special tools

In contrast to a complex, multifunctional basic vision system, our tool targets a single class of problems. This builds on the specific knowledge domain and will greatly reduce the expert knowledge required to use a machine vision tool. For example, a traditional machine vision system has many different detection algorithms. As far as developers are concerned, they know which tool to use to detect the project that suits them. Because the developers know their product, they find it easy to operate. However, for most users, they do not know where to start detection and have no clue. Therefore, the use of machine vision systems is incomplete. We turned the knowledge of vision experts into a toolkit that provides guidance on what needs to be done and the selection of algorithms for measurement in scale detection tools, so that the user can just focus on their task rather than becoming a vision expert.

(3) Increase the intelligence of the visual system so that it can overcome more environmental changes

For example, instead of asking the user to mount a part to a specific location, we use visual search to find the part within the field of view. Or for another example, we use algorithms that are immune to changes in brightness to simplify part lighting. Combining trends one and two, we have vision systems designed for specialized applications, such as monitoring tags. In this case, the vision system understands the final task and uses an interface that is familiar and suitable for this task. Visual merchants must balance ease of use with market size and the cost of carrying multiple products.