Computer Vision Methods in Applied Problems of Classifying Objects in Images
Yuri Leokhin, Timur Fatkhulin, Konstantin Boitsov
Abstract
This research is devoted to solving problems at enterprises related to the classification of objects in various types of images, including pictures of pipes, using computer vision methods. The purpose of the research is to improve the quality of pipe classification using computer vision methods. The relevance of the work is due to the fact that control of all operations at the enterprise from the standpoint of quality and efficiency requires automation of both key procedures and individual (specific) situations that arise daily or on a specific order from time to time. At present, all enterprises and operational procedures are required to undergo automation using modern methods, engineering and various software products. Accounting for the number of pipes during their storage and transfer to production is one of the problems that require full or partial automation using computer vision technologies. The object of the research is computer vision methods used to classify objects in an image. The subject of the research is the performance indicators of computer vision methods used to classify pipes. The following research methods served as the methodological basis for the work: comparison, description, measurement, scientific abstraction method, as well as analysis and generalization. The conclusion provides the main findings obtained as a result of the research.