Research On Algorithm Design And Simulation Of Vision Recognition And Localization For UAV Autonomous Landing

With the demands of application and development of aeronautic technology, more and more attentions are paid to the research on unmanned aerial vehicle (UAV) all over the world in recent years. A lot of research on computer vision-based for UAV autonomous landing has been done. In this paper, image fusion,runway recognition and Getparm Methed in the recognition and location system for UAV autonomous landing and 3D simulation is researched.Firstly, image fusion for vision system is researched. At first, image registration algorithm based on mutual information is researched: this paper introduces the principle of image registration algorithm based on mutual information, and three new image registration algorithms are introduced and then analyzed and comprised in experiments. Furthermore, image fusion based on wavelet transform is discussed: The principle of image fusion based on wavelet transform is also introduced, and three different wavelets are introduced and the effects of fusion algorithms based on these wavelets are analyzed and comparised.Secondly, runway recognition based on support vector machine (SVM) is discussed mainly. Target recognition and applications of SVM in this technology are introduced. Furthermore, the principle of SVM is presented. The procedure of runway recognition is designed and the results are comparised and analyzed in experiments.Thirdly, the getparms mathed for autonomous landing is researched. Image processing has been done in order to detect edges of runway and horizon, and at last, flight parameters are got. A corresponding simulation experiment has been done.Finally, the dissertation uses OpenGL programming in VC++ to design visual system aspect in three-dimensional simulation system of vision-based for UAV autonomous landing.