Sub-station

ShIP to

Shallow of machine vision detection system can fully replace artificial visual inspection?

by:Fugen      2020-10-16
In the application, often be given some specific defect mode, using machine vision to identify them exactly have occurred. But things are often encountered, many obvious flaws, because never happened before, or the mode too diverse, and was missing. With the development of science and technology changes with each passing day, machine vision slowly being familiar, then use the industrial camera and analysis software as the main body of machine vision detection system, can fully replace artificial visual inspection? If so, what are the range of application? If not, what is the lack of conditions, the difficulty in where? Idea 1: alternative? Is only a part can replace precisely. The small and medium-sized enterprise survival space is limited, automatic upgrade need capital, talent, due to various limitations, is not willing to invest a lot of money, so the machine visual inspection or in the strength of large enterprises. Has the highest current labor costs, companies are also high cost of operations, capital pressure, unless it is a must buy, not all live with. Idea 2: in the application of various kinds of defect detection, lighting is a difficulty. If the picture let a person see more carefully to give as a result, the algorithm is too difficult. Conversely if the good lighting, prominent feature to detect, algorithm is not difficult things. Question also has, however, is not to say that the algorithm has good enough, machine vision and human gap is very big! There's the biggest gap is intelligence. Yes, smart camera two words are far from intelligence, mainly reflects in: identification of defects for unexpected. In the application, often be given some specific defect mode, using machine vision to identify them exactly have occurred. But things are often encountered, many obvious flaws, because never happened before, or the mode too diverse, and was missing. If in person, while in operation process documents didn't let him to detect the defects, but he will notice, which has a greater chance to grab it. Currently, there are many smart camera supplier, there are also many analysis software vendors, although released software algorithm each have their respective characteristics, actually it's the same, to use functions are very similar. But all is according to fixed patterns and steps to deal with the camera to obtain images, went up from image analysis of a certain expected characteristics, thus gives the discriminant result, not a revolutionary intelligent algorithm. In aggregate, a short period of time ( 10 years) , we may not see as image analysis algorithm of human wisdom, because the related theory are still not ready. But machine vision detection still have a lot of the market, because it solves a very serious problem: the human is not stable. Operators of artificial visual inspection, no matter how you design the system of rewards and punishments, relatively high miss rate will occur. But machine vision detection equipment no problem, as long as you write good things in the algorithm, every game seriously. For quality control of the factory, we are more willing to give up the benefits of wisdom brought by the artificial visual inspection, and proved to be more stupid but meticulous work of machine vision. Because for the factory, the most important thing is 'controlled', we can accept a few special not happened in defect was leaking, but must know there are some frequent unacceptable defect will be found. Cost, machine vision is much cheaper than the human cost. General plant are four divisions, four operator's annual cost of more than 200000. The 200000 basically enough to make a set of machine vision. So, can such conclusion: visual inspection system in the short term can not completely replace manual visual inspection, the main reason is smart enough. But for the factory's quality control, machine vision than people.
Custom message
Contact