Digital Area Correlation Tracking by Sequential Similarity Detection.

摘要

Area correlation tracking is an effective method of tracking targets that do not have salient characteristics. This involves detecting a target within a scene by means of its similarity with a stored image of the target. This report presents an evaluation of the performance of a sequential similarity detection algorithm (SSDA) in the task of digital image detection. It is shown that the performance of the SSDA is highly competitive with that of the widely used mathematical correlation method. Both techniques are reviewed and tested on still photographs. An analytical evaluation of the SSDA applied to binary images is presented. (Author)