Development of Effective Methods for Structural Image Recognition Using the Principles of Data Granulation and Apparatus of Fuzzy Logic

The processes of intelligent data processing in computer vision systems have been researched. problem structural image recognition is relevant. This a promising way to assess the degree similarity objects. approach provides simplicity construction and high reliability decision making. main an effective description characteristic features distortion fragments analyzed reasons for changing input can be actions geometric transformations, influence background or interference. elements false objects with similar characteristics are formed. ensuring high-quality requires implementation means processing. Methods statistical modeling, granulation fuzzy sets, detection comparison keypoints on image, classification clustering data, simulation modelling used this research. proposed approaches formation concise vector representation unique keypoints. verification theoretical foundations evaluation effectiveness methods real bases performed using OpenCV library. applied significance work substantiated according criterion time without reducing interference immunity. developed allow increase images by several times. Perspectives research may involve identifying optimal number base set.