Research On Key Algorithms Of Quantum Image Processing

Quantum computation has prominent advantages in storage performance,parallel computation and uncertainty computation.It is the most possible alternative to classical computation and a new computing mode to solve the Moore law failure problem.In recent years,with the rapid development of optical imaging,Internet and high-performance computing,the visual data based on image and video have greatly reduced the cost of acquisition,calculation and exchange,which caused the explosive growth of visual data scale,resulting in visual big data problems.How to store and efficiently deal with massive visual data is the key technology that needs to be broken through.Quantum image processing is a cross subject of quantum computation and visual information processing.It combines the advantages of quantum computation and traditional visual information processing.It is an important approach to solve the problem of storage and efficient processing of massive visual data,and has great potential applications.It has gradually become a new research focus in the field of visual information processing.In this paper,the fast quantum Fourier transform algorithm,quantum edge extraction algorithm and quantum morphological gradient algorithm are studied using the unique advantages of quantum computation,and the application of quantum computation theory in the field of visual information processing is extended.It provides a new concept and idea for the development and application of visual data processing,and has certain practical value and guiding significance for the perfection and application of quantum computation theory.The innovative achievements of this paper are summarized as follows:1.A quantum vision representation model(QVRM)based on a normal arbitrary superposition state(NASS)is proposed.In order to exploit the unique performances of quantum computing in storage capacity and parallel computing to process visual data,the quantum image representation based on NASS is extended to quantum vision representation.Thus,a QVRM is established.2.A fast quantum Fourier transform(QFT)algorithm based on quantum vision representation is proposed.Using the iterative formula of Fourier transform established by extended tensor product,four quantum implementation circuits and corresponding inverse conversion circuits for fast one dimensional QFT are designed.On this basis,the 2D and 3D QFT based on quantum vision representation is proposed for the first time,which provides a feasible scheme for the application of QFT in visual information processing.3.Two quantum image edge extraction algorithms based on NEQR are proposed.One is edge extraction algorithm of quantum image based on Sobel operator,the other is edge extraction algorithm of quantum image based on Laplacian operator and zero crossing method.Firstly,the quantum implementation circuits of threshold operation comparison,addition,subtraction and shift operation based on NEQR images are designed.Then,the convolution operation is realized by addition and subtraction and shift operation,thus the Sobel operator and the Laplacian operator are realized.Finally,the two quantum image edge extraction algorithms are realized using quantum circuits of image threshold operation comparison and zero crossing method respectively.Compared to other quantum-inspired image segmentation algorithms,the proposed two quantum image edge extraction algorithms are suitable for the operation of the quantum systems,and the operation performance is exponentially higher than the corresponding classical algorithm.4.The quantum circuits of dilation and erosion operation for gray image morphology are designed,and the quantum morphological gradient algorithm of gray image is realized.It is a useful exploration on the field of quantum image morphology.