A technique for automatic concealment of confidential targets in fine spatial resolution imagery

Abstract

With the commercial availability of very fine spatial resolution imagery, the detailed image information on various kinds of ground facilities has been revealed to the public. In Korea, distribution of very fine spatial resolution images of the national confidential facilities such as military bases and nuclear power plants to the public is regulated by the government. Image distributors should camouflage such national confidential facilities in the images before the distribution. Therefore, the concealment procedure has been one of the most important and costly procedures for the image distributors in Korea. Because the concealment procedure has involved the manual masking of the target objects, it was very difficult to ensure the quality of concealment effect. A spectral clustering based technique for automatic concealment of the confidential targets in fine spatial resolution imagery was proposed in this study. The proposed technique has been implemented as an executable software program and experimented with various kinds of spatial targets in fine spatial resolution imagery. The experimental results showed that the proposed technique can be used as a practical procedure for those who need to hide image information before public distribution.