Research On Quantum Image Filtering Method

Quantum image processing is an interdisciplinary subject of image processing and quantum computing.It aims to accelerate image processing by using the characteristics of quantum computing.Quantum image filtering as an important research direction of quantum image processing has received more attention from many scholars.This thesis conducts research on the quantum image filtering,and the main research content includes the following three aspects.A quantum image filtering based on the median of decentralized neighborhood is proposed.Firstly,the existing neighborhood preparation module is improved,so that the quantum entanglement state of the neighborhood pixel and the central pixel can be correctly prepared.Then,a new median of decentralized neighborhood calculation algorithm is proposed,and corresponding quantum circuits are designed to improve the noise reduction effect.Then,a more concise quantum circuits of double-threshold noise detection is designed,and a quantum comparator with lower quantum cost is introduced to reduce the complexity of quantum circuits.Finally,the proposed method is validated experimentally with the quantum computing simulation platform.The results show that the proposed quantum image median filtering method has lower complexity and better noise suppression performance than the existing methods from two aspects of quantum circuit complexity and image noise reduction.A quantum K-nearest neighbor mean filtering for images is proposed.Firstly,a quantum basic module for absolute value calculation of positive integer difference is designed,and it is applied to the quantum circuit of pixel grayscale distance calculation.Then,a new sorting module is designed with pixel grayscale distance as the sorting condition,and the quantum circuit design of K-nearest pixel selection and mean calculation is realized by combining the divide-two module.Finally,the proposed method is simulated and compared with the classical method in terms of quantum circuit complexity and image noise reduction.The results show that the proposed method proposed has lower complexity while maintaining the same filtering effect as the classical method.A quantum vector median filtering for color image is proposed.Firstly,a color image neighborhood preparation module which conforms to NCQI color image representation model is designed by improving the quantum circuit of gray image neighborhood preparation.Then a quantum module for calculating three-dimensional vector distance is designed,from which a quantum circuit for calculating the aggregation distance of neighborhood pixels is constructed.Then,based on the principle of comparative sorting,a quantum circuit for extracting vector median pixels is designed.Finally,the proposed method is simulated and compared with classical and existing quantum color image spatial filtering methods in terms of quantum circuit complexity and image noise reduction.The results show that the proposed method has better filtering effect,does not cause image color bias,and has lower complexity.