CASPIAN JOURNAL
MANAGEMENT AND HIGH TECHNOLOGIES
Recognition of imposition flat objects on dimensionless marks of their images contours
 |
Read | Sadykov S.S., Kulkov Ya.Yu. Recognition of imposition flat objects on dimensionless marks of their images contours // Caspian journal : management and high technologies. — 2016. — №4. — pp. 10-20. |
Sadykov S.S. - D.Sc. (Engineering), Professor, Murom Institute (branch) Vladimir State University named after Alexader Grigoryevich and Nickolay Grigoryevich Stoletovs, 23 Orlovskaya St., Murom, 602264, Russian Federation, sadykovss@yandex.ru
Kulkov Ya.Yu. - post-graduate student, Murom Institute (branch) Vladimir State University named after Alexader Grigoryevich and Nickolay Grigoryevich Stoletovs, 23 Orlovskaya St., Murom, 602264, Russian Federation, y_mail@mail.ru
The purpose of the work is pilot study of efficiency of class recognition of the imposed flat objects by method of "the closest neighbors" on the basis of dimensionless marks of their bitmaps contours. For carrying out a research images of flat objects and details were used. The description of process of generation of test selection for each object is provided. At first the image of each initial object rotates by 360 degrees with a step to 1 degree. Further with use of the received "turned" images options of imposing of one object on another in sight of the camera of system of technical sight, on 2000 images for each class are created (i.e. for paired combinations of objects). For each image containing two objects a set of primary parameters created on their contours is calculated. The received parameters are used for calculation of dimensionless signs and forming of a vector of signs from them. The following step is the system grade level. Among the created vectors of marks separately for each class with use of a mean square deviation a set of etalons is selected. The base of reference vectors is created. Recognition of a class of an unknown object consists in receipt of its contour, calculation of primary parameters and forming of a vector of dimensionless marks. Further mean square deviations of its vector of dimensionless marks from all reference are calculated. The minimum value of a deviation will specify probable belonging to the corresponding class. Results of recognition are given in article, and also the possibility of use of this method in systems of technical sight is proved.
Key words: система технического зрения, конвейер, изображение, распознавание, плоский объект, детали, наложенные объекты, безразмерные признаки, среднеквадратичное отклонение, метод ближайших соседей, machine vision system, conveyor, image recognition, flat object,