Research On Machine Vision Ghost Imaging Theory And Experiment

After years of developments,machine vision has been widely used in industry,agriculture,military,scientific research and so on.In the machine vision system,the imaging part is extremely important,which is the guarantee of image acquisition.The image quality good or bad,and the image features obvious or not,directly determine the success or failure of machine vision applications.With the science and technology developing to the nanoscale,the classic microscopic imaging system can not obtain high-resolution images due to the diffraction limit constraints,hindering the applications of the machine vision in the field of nanotechnologies.To address this problem,in this paper,we propose a method to improve the resolution of the imaging system by applying the ghost imaging(GI)technique in the machine vision system.Although the GI technique has the ability to break the diffraction limit,its poor imaging quality,slow imaging speed,and high hardware requirements,put many limits on its practical applications.Therefore,to solve the problems of GI in practical applications,the focus of this paper is to improve the imaging quality and speed of GI.Another focus of this paper is to reduce the computational cost of GI.The main contents of this paper are shown as follows:(1)Both the high precision micro-vision displacement measurement method and GI machine vision system(GIMVS)are studied.Firstly,taking the displacement measurement of the three-degree-of-freedom flexible positioning stage as an example,a high precision displacement measurement method is proposed based on the micro vision,and the proposed method is verified by simulations and experiments.Then,the shortcomings of the micro-vision system in practical applications are analyzed,and to address the shortcomings,the GIMVS scheme is proposed.In additon,the feasibility and superiority of the GIMVS scheme are analyzed.(2)GIMVS design and experimental platform construction.An imaging system is designed according to the principle of computational GI(CGI),and an experimental platform is built up correspondingly.The experimental platform construction process includes hardware and software two parts.In the hardware platform construction process,the hardware parameters and the relative instructions are introduced.In the software platform construction process,the basic software development process and the software's main function development process are illustrated.Finally,both the hardware and software platforms are tested.(3)A high quality CGI scheme is studied by using the optimal distance search method.To solve the inaccurate problem of the measurement distance between the spatial light modulator and the target object in the GIMVS,the relationship between the measurement distance and CGI is first analyzed,and then an optimal distance search algorithm is proposed,which is used to find the best distance.During the optimal distance search process,a compressive sensing(CS)algorithm is used for the image reconstruction.It is found that the CS-based GI is sensitive to the distance variation,and by combining with the CS algorithm,the proposed algorithm can not only be used to improve the imaging quality of CGI,but also to obtain the optimal distance.(4)The problems of reducing the amount of the measurement data that is required for imaging and improving the imaging quality are studied.To address the poor imaging quality and high computational cost problems in the GIMVS,we propose an adaptive differential correspondence imaging(CI)algorithm.The basis of the proposed algorithm is CI.In the proposed algorithm,the measurement data in the bucket detector is first processed by a differential technology,and then the processed data is sorted in an ascending or descending order.Finally,an adaptive algorithm is used to create the positive and negative subsets(PBS),and the object iamge is reconstructed from the PBS.Through simulation and experimental analysis,we find that the proposed algorithm can not only reduce the amount of the measurement data,but also can improve the imaging quality.(5)Research on high quality CI algorithm.To further improve the image quality and reduce the computational cost in the GIMVS,we propose a high quality CI algorithm based on CS technique.In the proposed algorithm,the measurement data from the bucket detector is first sorted in an ascending or descending order,and then according to the sorted data,the PBS is created by selecting the corresponding data from the reference detector.Finally,the object image is recovered from the PBS using the CS technique.The simulation and experimental results demonstrate that the proposed algorithm can dramatically improve the imaging quality and reduce the computational cost in the GIMVS.The results also show that the proposed algorithm enjoys the best comprehensive performance.Finally,the future research direction and contents are proposed based on the summarization of the whole paper.