Detail enhancement of blurred infrared images based on frequency extrapolation

A novel algorithm for enhancing the details of the blurred infrared images based on frequency extrapolation has been raised in this paper. Unlike other researchers' work, this algorithm mainly focuses on how to predict the higher frequency information based on the Laplacian pyramid separation of the blurred image. This algorithm uses the first level of the high frequency component of the pyramid of the blurred image to reverse-generate a higher, non-existing frequency component, and adds back to the histogram equalized input blurred image. A simple nonlinear operator is used to analyze the extracted first level high frequency component of the pyramid. Two critical parameters are participated in the calculation known as the clipping parameter C and the scaling parameter S. The detailed analysis of how these two parameters work during the procedure is figure demonstrated in this paper. The blurred image will become clear, and the detail will be enhanced due to the added higher frequency information. This algorithm has the advantages of computational simplicity and great performance, and it can definitely be deployed in the real-time industrial applications. We have done lots of experiments and gave illustrations of the algorithm's performance in this paper to convince its effectiveness.