Research On Edge Detection Based On Single Pixel Imaging

With the rapid development of computer technology,images have become one of the important ways for human beings to obtain information.Single-pixel imaging,as an emerging optical imaging technology,has received widespread attention since it was proposed.Edge detection technique is used to extract the edge information of the image,which can effectively reduce the amount of data in the image processing process.The application of edge detection to single-pixel imaging systems has become a new research hotspot,as it enables edge imaging of unknown objects and has good interference immunity.Existing techniques still suffer from excessive data volume,too many measurements and poor image quality.In this paper,the following three algorithms are proposed to address the shortcomings of the existing techniques,combining single-pixel imaging technology,edge detection theory and compressed sensing technology:1.A computational ghost imaging technique based on the NESTA algorithm is proposed.The combination of sparse sampling technology and computational ghost imaging technology solves the problem of difficult data transmission due to the large amount of speckle pattern data in the imaging process.First,a spatial light modulator is loaded with a series of phase-only masks to generate speckle patterns to illuminate the object image,and a bucket detector that has no spatial resolution is used to collect the reflected or transmitted total light intensities.The speckle patterns are compressed and transmitted to the receiver,this process that significantly reduces the amount of data in the transmission process.The receiver performs a sparse reconstruction of the compressed speckle pattern using the NESTA reconstruction algorithm,and then performs a second-order correlation between the reconstructed speckle pattern and the intensity values to obtain the object.Finally,the feasibility and effectiveness of this algorithm are verified through specific experiments,and the numerical analysis shows that this algorithm has good compression performance.2.A sub-pixel edge detection algorithm based on computational ghost imaging is proposed.Firstly,the phase patterns required in the process of computational ghost imaging are half-pixel shifted,and two sets of phase patterns are designed to illuminate the scene sequentially in conjunction with the edge detection theory.the light intensity values recorded by a bucket detector are subjected to a second-order correlation operation with the un-shifted phase patterns to obtain the edge images.This scheme improves the imaging quality by selecting the Hadamard matrix as the phase patterns,and improves the edge detection accuracy by sub-pixel shifting of the patterns.Numerical simulation experiments to analyse the feasibility and effectiveness of the algorithm from several aspects.3.A sub-pixel edge detection algorithm based on four-step phase shift technology is proposed.Firstly,the sinusoidal structured light patterns required in the four-step phase shift imaging technology are half-pixel shifted,and two sets of phase patterns are designed to illuminate the scene sequentially in conjunction with the edge detection theory,and the Fourier spectrum of the image edge information is obtained by calculating the light intensity values collected by the bucket detector,and finally the edge image to be detected is obtained by inverse Fourier transformation of the obtained spectrum.This scheme improves the detection accuracy by sub-pixel-level shifting of the patterns,and the dithering technique is invoked to overcome the limitations of the traditional four-step phase shift technique.Simulation experiments and numerical analyses verify the feasibility and effectiveness of the algorithm.