| With the development of telecommunication ,computer ,micro-electronic techniques and the appearance of the high-speedprocessing chip and mass storage, the video image processing has been acommon and hot research field of academic areas and industrial groups. Thetechnologies of video segmentation, object recognition, object tracking,vision adjustment of video sequence and filtering of sensitive image areasare mostly concerned in video image processing. In this dissertation, by thefeature that the video image can be divided into background object andothers, the layer model of video sequence is presented. Based on this layermodel and features of different objects of the scene video sequence, aimedat the background object, the segmentation technology with integrateddetection of distributing matrix is studied. The experiment model, keytechnical modules and performance parameters of this technology areanalyzed, and the PC simulation results are presented, comparing ourmethod with other methods.After getting the background layer image, the segmentation methods ofmoving objects are discussed. Firstly the common segmentation methodsare described, then, the segmentation method based on the differencebetween the background and video sequence is studied.In addition, the detail visibility of video sequence and its adaptiveadjustment are studied. Based on the visibility determined by mean andvariance, its detection and the way to change it are discussed. Using thechanging way, the clear image can be turned into blurry one, and vice versa.Then, based on the visibility affected by the contents, its adaptiveadjustment is studied. The technology discussed can be used to filter thesensitive areas or detect the abnormity of video image.In summary, the dissertation is mainly concentrated on the objectslayering and visibility analysis. This dissertation has established thetheoretical and technical basis in the related area for the following research... |
PDF Full Text Request