Machine Vision (machines vision, MV) & Computer Vision (computer vision, CV) from the classification of disciplines, both are considered artificial intelligence ( Artificial Intelligence ) subordinate accounts. machine vision is to use machines instead of human eyes to do measurements and judgments. Machine vision system through machine vision products (i.e., image ingestion device, CMOS and CCD two kinds) will be absorbed into the target image signal, transmission to a dedicated image processing system, to obtain the target's morphological information, according to the pixel distribution and brightness, color and other information, transformed into a digital signal; The image system can perform various operations on these signals to extract the characteristics of the target, and then control the equipment action on the spot according to the discriminant results. Computer vision refers to the realization of human visual function by computer--perception, recognition and comprehension of the three-dimensional scene of the objective world. It isa science that studies how to "see" machines, further, refers to the machine vision of using cameras and computers instead of human eyes to identify, track, and measure objects, and to do further graphics processing to make computer processing an image more suitable for human eye observation or transmission to instrument inspection. Computer vision can also be seen as a science of how to make artificial systems "perceive" from images or multidimensional data. The ultimate goal of the study is to enable computers to visually observe and understand the world as people do, with the ability to adapt themselves to the environment. There are two main types of methods: one is bionics, referring to the structure of the human vision system, the corresponding processing module to complete similar functions and work, the other is the engineering method, from the analysis of the function of human visual process, do not deliberately simulate the internal structure of human visual system, Only the input and output of the system are considered, and the system function is realized by any existing feasible means. the two techniques used are similar, the main difference is the use of scenes, focus on the differences,MV Offset Industrial Workshop application, more attention to generalized image signal (laser, camera) and automation control (production line) aspects of the application. Focus on the analysis of the quantity, such as the part diameter. CV-biased software algorithm, more attention to human-related applications, focus on (3D) image signal itself research and image-related interdisciplinary research (medical image analysis, map navigation), focusing on confrontation analysis, such as face recognition, license plate recognition. The application of computer vision is relatively complex, the type of object to be recognized is also many, irregular shape, the regularity is not strong. Sometimes it is difficult to use objective quantities as a basis for recognition, such as identifying age and gender. So deep learning is better suited to computer vision. And the light, distance, angle and other preconditions, are often dynamic, so for accuracy requirements, generally lower. Machine vision is just the opposite, the scene is relatively simple fixed, the type of recognition less (in the same application), rules and laws, but the accuracy, processing speed requirements are relatively high. Regarding the speed, the general machine vision resolution is much higher than the computer vision, and often must be realistic, therefore the processing speed is the key, at present basically is not suitable to use the deep study.
business, the application of computer vision more widely, after all, a lot of business is related to people, such as face recognition, behavioral analysis, many vertical areas have a potential demand for computer vision, relatively speaking, more suitable for entrepreneurship;and machine vision as the name implies, the business is mainly related to the machine, and the accuracy and even security requirements are high, but also in the qualification brand has a higher threshold, so oligopoly serious, generally speaking, more suitable for work rather than entrepreneurship. Links: https://www.zhihu.com/question/23183532/answer/105619829 Links: https://www.zhihu.com/question/23183532/answer/23896265
From for notes (Wiz)
Machine Vision and Computer vision
