Fast curvelet transform based non-uniformity correction for IRFPA

The Curvelet transform was developed from the wavelet transform. The applications of Curve-let transform reveal its great potential in image processing due to its unique characteristics. In this paper, the theory and implementation of Curvelet transform is summarized. The traditional Curvelet transform involves a complicated index structure which makes the mathematics and quantitative analysis especially delicate, and it uses overlapping windows increasing the redundancy. The Fast Curvelet Transform was discussed in this paper, which has the optimal sparse representation. By utilizing Curvelet wrapping algorithm based on translation invariance to the nonuniformity correction of the IRFPA, better MSE compared with traditional methods can be obtained. Great compute and analysis have been realized by using the discussed algorithm to the simulated data and real infrared scene data respectively. The experimental results demonstrate, the corrected image by this fast Curvelet transform algorithm not only yields highest Peak Signal-to-Noise Ratio values (PSNR = 33.803), but also achieves best visual quality.