On the possibilities of using the method of near-periodic analy-sis for image processing

Abstract

The paper presents the application of the method of near-periodic analysis based on the shift function to the pro-cessing of data presented in the form of aerial images of the dynamics of cyclone activity. In the course of the study, spa-tial, temporal and spatiotemporal near-periodic analyses were carried out. The proposed method of near-periodic analysis based on the shift function showed the possibilities for performing spatial segmentation of the image in both Cartesian and polar coordinates, and also determined the existence of periodicity in the time scale of the dataset. Based on the results obtained, a spatiotemporal study of the da-taset with the final image segmentation was carried out. The existence of near-periods in the Cartesian system of spatial coordinates of the image is shown. The existence of near-periods in the Cartesian system of spatial coordinates of the image is shown. The substantiation of the fact that in the Cartesian coordinate system, almost periodic analysis in polar coordinates represents a qualitative rhythmic marking is presented. It is demonstrated that an almost periodic analysis based on a generalized shift function provides the possibility of using the studied data set in a time slice. The article presents a classification model for parts of the image, obtained on the basis of statistical estimates of slices of the generalized shift function of time series groupings, When using near-periodic analysis, the possibility of using it for spatial and temporal analysis of data obtained from aerial photography of the dynamics of cyclone activity is demonstrated