Missiles Tv Seeker Image Matching Algorithm

The application of TV seeker is prevalent in the modern high-tech weapons, and image matching owns important theoretical and practical values in many military fields, such as pattern recognition, guidance weapon, map matching, etc. These values help image matching become one hot spot in the relating research field. Due to the complex environment, plenty interferential factors, image rotating, scale transform and illuminative difference, the traditional matching algorithm is difficult to realize the rapid recognition and accurate positioning of the target. Therefore, finding one stable, fast and precise matching algorithm is critical for the application purpose. This paper uses the TV seeker for background, contrapose the image preprocessing, the matching algorithm based on grayscale correlation, corner detection algorithm and image partial invariant feature, focuse on the improving of SUSAN, Harris and SURF algorithm.First of all, the image preprocessing technologies includes image enhancement, geometry correction, etc. Contrapose to the influence from the image fuzzy, noise, geometry deformation in image matching performance. It gives the corresponding solutions, restores the image and reduces the margin of error in image matching.Secondly, the paper analyzes the matching algorithm based on grayscale correlation in detail. In the view of the defects of SUSAN corner detection algorithm, the self-adaption threshold of SUSAN algorithm is proposed. In the view of the defects of Harris corner detection algorithm, the paper combines the iterative self-adaption threshold of Harris algorithm with Gaussian space, the Harris corner detection has multiple measure, and provides the characteristics of corner detection.Finally, on the basis of the image partial invariant features matching theory, with SIFT and SURF characteristic descriptor, they solve the extreme value quickly. According to the condition, it is important that the real time in image matching, this paper combines the SURF algorithm with the Kd-tree and Best Bin First algorithm, the results of the theoretical analysis and numerical simulation show that the fast SURF algorithm is real time, validity, accuracy, and supplies the beneficial reference for the engineering application of image matching.