Image preprocessing based on spatial statistics for the optical correlation recognition of aviation photographs

A new method of digital image processing is developed for the correlation recognition of aviation photographs under volume holographic correlator. The physical characteristics of semivariogram of the image are analyzed according to spatial statistics, which suggests the principle of the image segmentation, including the dimension of the templates, the space between each template and the image restoration. Then the residual space of each template is calculated directly based on the 2D-images to enlarge and to scatter the differences between the correlativity. After that, the residual spaces are reconstructed to form the final set of templates. Finally the proposed method is tested on a series of aviation photographs. The experimental results come with rather higher precision as well as rather lower computational complexity, which supports the real-time capability of the optical correlation recognition of aviation photographs under volume holographic correlator.