Research Of Image Processing Algorithm And Simulation Based On Quantum Technology

Computer image processing is an indispensable part in the field of information processing,especially in aerospace technology,remote sensing,biomedicine and other fields.With the rapid development of digital image processing technology,the computing performance of classical computers has basically reached saturation,and the computer field is facing severe challenges in the future.In order to seek higher computing performance,in the early 1980 s,P.Benioff and others of Argonne National Laboratory in the United States put forward the concept of quantum computing for the first time by combining the characteristics of quantum physics and computer technology.After years of development,quantum technology has gradually matured.At present,many countries have put forward the development plan of quantum technology.However,in the field of image processing,a variety of algorithms proposed are not suitable for quantum computers.In order to enrich the research of quantum technology in the field of image processing,this paper studies three key technologies of image processing: image encryption,image recognition and image watermarking.1.Aiming at the encryption problem of grayscale image,an image encryption scheme based on discrete-time alternating quantum walk and advanced encryption standard(AES)is proposed.The AES algorithm is improved by using the quantum properties.The algorithm uses the key stream generator related to the alternating quantum walk parameters to generate the probability distribution matrix.Extracting some singular values of the matrix as the key of the AES algorithm.Replacing the Rcon of AES algorithm with the element of probability distribution matrix.Then,the ascending order of the elements of the clone probability distribution matrix disturbs the mapping rules of the S-box and Shift Row transformation in the AES algorithm.The algorithm uses probability distribution matrix and explicit XOR operation to complete the preprocessing,and uses the improved AES algorithm to complete the encryption process.The technology is based on simulation verification,including pixel correlation,histogram,differential attack and noise attack.The experimental results demonstrate that the encryption effect of the algorithm is remarkable.2.In order to solve the problem that convolutional neural network requires higher and higher memory and time efficiency,a new model for digital image classification is proposed.The model is a quantum convolution neural network based on strongly entangled parameterized circuits.Firstly,the classical image is preprocessed and qubit coded,and the feature information of the image is extracted,which is prepared as a quantum state as the input of the quantum convolution neural network model.By designing the quantum convolution layer,quantum pooling layer and quantum fully connected layer structure of the model,the main feature information is extracted efficiently.Finally,the Z-basis measurement is performed on the model output,and the image classification is completed according to the expected value.In this work,the experimental dataset is MNIST.The classification accuracy of {0,1} and {2,7} is as high as 100%.The comparison results demonstrate that the quantum convolution neural network model with three-layer network structure with average pool sampling has higher testing accuracy.3.In view of the lack of quantum image watermarking algorithms at present,a new circuit structure for embedding and extracting quantum LSB watermarking is constructed.The algorithm quantizes the watermark image and carrier image using GQIR quantum image representation method,and then inputs the quantum watermark image and the quantum carrier image into the constructed quantum circuit.Firstly,the quantum watermark image scrambles the position information through the quantum Fibonacci scrambling circuit.Secondly,the scrambled image information is embedded into the quantum LSB circuit,and the pixel information of the quantum watermark image is repeatedly embedded into the quantum carrier image.The extraction circuit of quantum LSB watermark is divided into four parts: scrambling,block,counting and comparison.In this experiment,a group of carrier images and information images are simulated and tested.The experimental results demonstrate that the carrier image with watermark is highly similar to the carrier image through visual effect and PSNR measurement.In addition,by adding different intensity of salt and pepper noise to the carrier image with watermark,the visual effect of the watermark image extracted by observation is still very good,which proves that the algorithm has good robustness and security.