基于图像处理的移动目标测量方法的研究-检测技术与自动化装置专业论文.docxVIP

优秀毕业论文 精品参考文献资料 黑龙江大学硕士学位论文Abstract 黑龙江大学硕士学位论文 Abstract ln recent yearS， human-computer interaction problem has been jmportant research direction in the field of machine vision．At present it has extensiVe application and potential development space in many fields，such traffic contml，intelligent monitoring，air navigation，Somatic game for body building，etc．As a hot topic in the neld of human-computer interaction，This paper mainly researches the moVing target detection and target tracking in the measurement of moVing target． For the defects of three—frame difference algorithnl in moVing target detection，an improVed modeling method based on Gaussian miXture model combined with three-fr锄e difference algorithm has been proposed． The main idea is to establish multidimensional Gaussian mixture model for each of the background pixels．Then the three．frame diⅡ-erence method was blended in the model to detemline the background area and sports in realtime．So that the holes caused by three．frame di骶rence can be removed and Extemal rectangle which can follow the ta唱et-s moVement was added in，generated Gaussian model．Thus reducing the problem of large amount of calculation as a result of the inter、，ention of Gaussian model，SaVe operatjon time and achieVe the ideal e能ct of noise cancellation．The eXperimental results show that：compared with three-厅ame difrerence method，this method has better denoising e旋ct and reduce more conlputation，applicable to reaI—time target detection． Aiming at C锄shift algoritllm interferes between ta唱et color and background color around and the problem of invalidation in the complex and changefuI background． this paper presents an improved Camshift algorithm based on SURF f．eature point matching，For example，When the hand and facial oVerlapping phenomenon occurs，do the SURF matching according to the current f}ame moVing targets and the original image．Thus sustained good tracking be achieVed under the color conf’usion．Based the tracking algorithm in