[Total: 0 Average: 0/5]
The problem to know, as exactly as possible, the number of cardboard sheets delivered to the customer, is an important question for a corrugated cardboard industry because actual solutions don't or partially consider cardboards eliminated, by quality control and internal transport. In this context, try to project solutions based on imaging vision and processing is an idea to solve this computation problem improving the limitations of the actual solutions. Many vision systems, that can solve this problem, are implemented in this thesis, based on the different possibility to divide the total image in more subimage, with the conclusion that, one solution that use the global image at the request resolution is not necessary; furthermore this kind of solutions, for actual price of industrial camera technology, exceed the budget for this dispositive. So, the final projected solution is based on the computed density of cardboard in a segment of the image and the successive exstension of this density to the global image, with a proportional adactive alghoritm. This solution consider quality control and internal transport and solve the presented problem with the best relationship betwen quality and cost.