| The capability of Pattern recognition is an important part of human intelligence. Realizing the automatic pattern recognition by computer is a key breakthrough of intelligence machine developing. Pattern recognition is not only a research field having great scientific value, but also a great key technique which we need to break through in many applications of the digital network times.Despite the classification algorithm for pattern recognition research has been engaged in a long time and has achieved many results, but so far, the machine identification of skills not yet in comparison with the human cognitive abilities, this is still a difficult open issue, so they and other subjects and theories for pattern recognition is still a strong research value and significance. At the same time, pattern recognition classification algorithm research to further enhance the complex environmental conditions of the text recognition, speech recognition, fingerprint recognition, remote sensing, medical diagnostics, industrial inspection, weather forecasts, satellite images and other areas of aviation technology and application issues are an important reference and instructive.Integrating the research work of some application direction, such as classification and recognition of electromagnetic leakage of computer video text image and blood volume pulse sampled-data classification, this paper systematic describes the composition of the statistics pattern recognition system, traces the key techniques which are widely used in many hot application, focuses on information classification and recognition algorithm for a major technology and applications. Meanwhile, the realization of a key technology in the new instance of the application, the design of electromagnetic leakage of computer video image recognition system and blood volume pulse sampled-data classification system for the corresponding realization of the original technology has been improved and achieved a very good effect.Preprocessing, feature extraction and selection, classifier design are three key techniques of pattern recognition. In this paper, a study conducted around the following work:The first, the iterative threshold method was used in computer electromagnetic leakage of the text image binarization to obtain a better de-noising effect; The second, the use of blind signal processing techniques (such as PCA, FastICA, Infomax, etc.) to extract the blood volume pulse characteristics of sampled data, studies show that features extracted by the FastICA make the classification results more accurate; The third, Bayesian probability and statistics-based classification algorithm and support vector machine classifiers was respectively applied to the computer electromagnetic leakage of video image recognition system and blood volume pulse sampled-data classification system, and have achieved very good results, which laid a good foundation for the further research of these two areas; Finally, studied on the design of a novel telemetry capsule microstrip antennas.The main contribution and innovation of this paper include:Aiming at the preprocessing problems of electromagnetic leakage text image, we purposed an iterative threshold method as a pretreatment means, and the traditional Bayesian probability and statistics-based classification algorithm was applied to an innovative image recognition in the text of electromagnetic leakage and get a good recognition results; basing on the research of blood volume pulse sampled-data classification, we purposed some blind signal process methods to select features, combined with support vector machine classification algorithm, which achieved a successful classification result. In addition, we also designed a novel defected ground structure and applicated it on telemetry capsule microstrip antennas. The simulated results show that the high order harmonics of microstrip based on DGS could be restrained successfully and radiation character is improved. |
PDF Full Text Request