Features of correlation analysis of images and video sequences

The paper deals with 2D-dimensional autocorrelation functions of 2D-dimensional video signals with allowance for the finiteness of the field of view, and it is established that the limits of integration need to be adjusted. It is shown that the autocorrelation function can be used to estimate the location of the object relative to the edges of the image. Possible variants of the object displacement along the image field are considered, the autocorrelation function is determined through the change in the area of the object when it leaves the field of view. Variable substitutions were introduced to find the difference between the autocorrelation functions in the processing of video sequences. Stable areas are identified that change only their location depending on the parameters of the object’s movement and its position relative to the center of the frame. The results are consistent for continuous and discrete models. The stability of the proposed method for identifying the motion on the difference in the autocorrelation functions to the change in the shape of the object is shown.