Research On The Enhancement And Normalization Of Infrared Image

In this paper, a number of theoretic algorithms of image enhancement on infrared images are studied by analyzing the characteristics of infrared image. Laplacian sharpening is mainly studied. A new algorithm of image sharpening is presented in this paper which combines the algorithm of limited Laplacian sharpening with high-frequency emphasis filter. This algorithm can adjust both the brightness of image and the degree of sharpening. Simulation results proved that the infrared images could be effectively enhanced by this algorithm.Image normalization is very useful in image recognition. In this paper we can resolve the image distortions successfully based on the invariable characters of the moments.A new method of image normalization is provided in the paper, which can normalize the images distorted by translation, scaling, rotation, skew and reversal. In the algorithm, we first compute the covariance matrix of the image, then rotate and scale the image according to the eigenvectors and eigenvalues of the covariance matrix, thus we get the compacting image which is invariant to translation, scaling, and skew. In the next step, we rotate the image with respect to the image ellipse tilt angle. But both the ellipse tilt angle and the rotation of image can bring in an uncertainty of 180 degree rotation. The last step of the algorithm is reversal normalization which can solve the problem of 180 degree rotation simultaneously. The feasibility and effectiveness of the algorithm are proved by simulation experiments.