Research On CCD Based On Image Processing And Pattern Recognition Technology

This thesis designs the system of image acquisition and researches pattern recognition algorithm based on area array CCD image sensor in the basis of reading lots of related documents, And then, simulate the two algorithms of collecting images.The work of thesis as follows:(1)Read the papers and documents on the CCD sensor development and application. Have a certain understanding in the development of the CCD image sensor process, sensor structure, sensor working principle, performance and development methods. For future study and work laid a good foundation.(2)Design and innovate the drive circuit and control circuit of the image acquisition system, experiments carried out the two circuits. The resules show that the optimum circults are in a good condition. It can basically achieve the goal of stabilizing the image acquisition targets, and proved that the improvements are feasible.(3)Introduce the basic theory of pattern recognition from the aspects of statistical pattern recognition, structure pattern recognition and intelligent pattern recognition. In particular, analysis and research deeply on the application of pattern recognition theory and methods in the fields of medical image processing.(4)Using three-story-driven top-down strategy for reducing the time and space complexity in the process of image recognition.(5) Especially from the perspective of cluster analysis of medical image pattern recognition research. The main contents include climbing algorithm and Gaussian mixture based on the realization, simulation, and data analysis of the results. With other methods of contrast, this paper uses algorithm has good effect to restore the image advantages, in Medical Image Segmentation and recognition in the field is of practical value.(6) The simulation and experimental results can be shown climbing algorithm and Gaussian mixture based on the improved image processing algorithms for medical image pattern recognition with effective and practical for Medical Image Segmentation and identification of a better technology platform. Overall, to achieve a relatively good recognition.With the recognition of medical image pattern recognition aspects of the deepening, The clinical application are more and more to the Image Classification and automatic detection of disease.In the basis of current study and work, We are prepared in the following areas for further development:(1)Continue in the present study, Focus on the practical application of the proposed new method and performance of the algorithm for further testing, evaluation, analysis and improvement, achieve accurate judgments on the clinical cases and automatic identification.(2)Evaluation of the introduction of cluster density model, On the proposed model based on kernel density clustering for medical image climbing algorithm and Gaussian mixture model of medical image data weighted clustering algorithm evaluation of the the next step effect, to achieve the basic algorithm of the innovative.(3)optimize the application program for further design, so that the speed can obviously improved...