Image Super-Resolution Reconstruction Algorithm Research Based On Directionlet Transform

Image super-resolution reconstruction has always been a very difficult and challenging research topic in the field of image processing.With the rapid development of the Internet,the screens of display devices such as mobile phones,pads,and computers continue to increase,and people have increasingly higher requirements for image resolution.The solution to increase the area of the photosensitive device of the photographing equipment to obtain a higher resolution image is too expensive and subject to the development of device technology,so it cannot be widely adopted.Therefore,a large number of researchers are engaged in the algorithm research of reconstructing high-resolution images from low-resolution images,trying to use the software to achieve the enlargement of image size on the image display side,so as to avoid the constraints of the development of shooting equipment.According to the number of reference images,the image super-resolution reconstruction algorithm can be divided into two types: the reconstruction based on multiple low-resolution reference images and the reconstruction based on a single low-resolution reference image.Among them,the super-resolution reconstruction algorithm based on a single low-resolution reference image is more difficult and widely used.Starting from the wavelet domain,this topic proposes a super-resolution reconstruction algorithm based on a single low-resolution reference image,and obtains an ideal reconstruction quality.This article first analyzes and compares the advantages and disadvantages of the basic theories of common super-resolution image reconstruction techniques and various algorithms.Then,according to the advantages of wavelet transform in the sparse representation of the image,the principle of super-resolution image reconstruction in the wavelet domain is analyzed in detail,and two algorithm frameworks for reconstructing super-resolution images in the wavelet domain are compared: edge-directed and edge-corrected,and under the framework of edge correction reconstruction,an edge correction super-resolution image reconstruction algorithm based on sparse interpolation and wavelet transform is proposed.Experimental simulation results show that the algorithm has low computational complexity and the image reconstruction quality is better than the original algorithm framework.Finally,in view of the shortcomings of wavelet transform in the field of image processing technology,that is,the wavelet analysis transform will produce a large number of coefficients with larger amplitudes in the representation of the image contour,this paper deeply studies the Directionlet theory,and compares the Directionlet transform with the wavelet transform experimentally The advantages.Finally,combining the Directionlet coefficients with the hidden Markov tree(HMT)model,a super-resolution image reconstruction algorithm based on the Directionlet domain HMT model is proposed.The method achieves the expected results in performance and computational complexity.The quality of the reconstructed image is greatly improved compared with the sparse interpolation based wavelet domain edge-corrected super-resolution image reconstruction algorithm and the original edge-modified super-resolution reconstruction algorithm.