The Study Of Motion Target Intelligent Recognition And Track Technology Based On Infrared Image

Target detection and tracking system based on infrared detection technology, belongs toPassive detection system, not emit any energy to the outside world in the detection process, Ithas a strong anti-detection capabilities, long working hours, and to adapt to the complexenvironmental advantages, been widely used in the military field of air defense early warningand missile guidance. Therefore, research complex background infrared image targetintelligent tracking and detection algorithm, breakthrough algorithm can not only at thetheoretical level, become a vanguard weapon can also be applied to the actual battlefield.This thesis based on the profound study of the research status and development ofdomestic and overseas technology of tracking and detection of infrared image target offers aset of algorithm scheme for effectively tracking and detecting the small dim target in complexbackground. The research work includes the following aspects:Firstly, through the analysis of the infrared spectrum, to understand the concept of theinfrared light. Based on this study the composition and working principle of infrared imagingsystems, summed up infrared imaging system characteristics. With visible light imagecontrast, summarizes the characteristics of the infrared image by histogram. Analysis of theradiation characteristics of the various types of infrared small target and background.Summarize the noise that can impact the infrared image quality in the imaging system.In addition, under the premise of the characteristics of the target and background infraredimage analysis, put forward the idea of image preprocessing, and the establishment of amathematical model of the infrared image. On this basis, studied the infrared imagepreprocessing algorithm in the time domain, transform domain and space domain threetheoretical system. and make simulation to more application of the algorithm, comparativeanalysis of simulation results filtering performance of each algorithm, and came to thepreprocessing algorithm in this paper.Furthermore, study the two categories in infrared target detection, comparative analysisof the advantages and disadvantages of the two types of algorithms. Combining targetcharacteristic in herein, using the detect before track method of threshold segmentationdetection algorithm as an object of study. Combined the advantage of genetic algorithmoptimization, proposed the improved Otsu algorithm, the proposed algorithm uses a geneticalgorithm for inter-class variance decision function optimal threshold, the solution of theoriginal algorithm computationally intensive problems in the process of seeking the threshold. Experimental results show that the proposed algorithm to ensure the accuracy of the detectionpremise, optimize the complexity of threshold value calculation.At last, establish maneuvering target tracking framework. Research Bayesian theory,linear filtering techniques and nonlinear filtering techniques. Proposed an improved particlefiltering algorithm, proposed algorithm under the framework of the interactive model,establish three target motion model to cover all of the possible target state of motion, differentmodels to match the corresponding filter to simplified algorithm structure. The experimentalresults prove that the proposed algorithm reduces the root mean square error in the trackingprocess, enhance the accuracy of target tracking, and has better real-time.