Título del libro: Computer Vision
Título del capítulo: Some applications of computer vision systems in micromechanics
ERNST KUSSUL;
Autores externos:
Idioma:
Inglés
Año de publicación:
2011
Resumen:
For general purpose image recognition systems we developed three neural classifiers: LIRA, PCNC, and Pairwise Coding Classifier. We tested the neural classifiers in various tasks of image recognition in the area of micro mechanics. The first task was the texture recognition problem. Different methods of mechanical metal treatment give different surface textures. It was made the computer vision system for recognition of 4 different types of workpiece treatment: milled surface, polished with sandpaper, turned with lathe, and polished with file. The LIRA classifier was used to solve this problem. 20 samples of each texture were prepared for the experiments. Some randomly selected samples were used for classifier training, and other samples were used for testing. In the case of 10 training samples 99.8% recognition rate was obtained. The second task was connected with recognition of the shape of small screws. When small screws are manufactured on computer numerical control (CNC) machine tool the cylinder part of the screw is turned with one cutter and the thread is made with another cutter. The mutual position of the cutters sometimes is not aligned precisely. In this case the shape of the thread will be incorrect. The recognition of the thread shape gives the information aboutmutual position of the cutters. This information allows us to correct the cutter positions. We tested the LIRA classifier in the task of thread shape recognition. The recognition rate of 98.9% was obtained. The cost of micromechanical devices will depend not only on the cost of manufacture of components but also on the cost of the device assembly. So, the assembly automation is very important. Many assembly procedures can be made with the aid of computer vision systems. Earlier we proved the LIRA classifier to obtain adaptive feedback in the pin-hole alignment task. Now we consider the problem of component selection from the depository, detection of their orientation and transportation to the assembly device. This is the third task that we describe in this chapter. This system will include the micromanipulator and camera, and will permit us to work with different micromechanical components without the development of new mechanical feeders. The manipulator using computer vision system must recognize the types of componen
Entidades citadas de la UNAM:
ISBN:
9781612093994
Editorial:
Nova Science Publishers, Inc. Nombre de la publicación:
Some applications of computer vision systems in micromechanics
Página inicial:
1
Página final:
39
Tipo de producto:
Capítulo de un Libro
