【正文】 ......................................................................12..................................................................................12...........................................................................................14...................................................................................15..................................................................................................16.......................................................................................17.......................................................................................17..................................................................................................18第3章圖像分割.............................................................................................18..........................................................................................................18......................................................................................18.......................................................................18......................................................................................19..................................................................................20..................................20...........................................................................................20...........................................................................................21..................................................................................................25第4章陰影檢測理論基礎(chǔ).............................................................................25..........................................................................................................25..............................................................................................25..............................................................................25..............................................................................26..................................................................................................26第5章陰影抑制算法.....................................................................................26..........................................................................................................26..........................................................................27..........................................................................................27...................................................................................29..................................................................................................34結(jié)論.................................................................................................................34參考文獻..........................................................................................................35致謝................................................................................................................36附錄一...........................................................................................................3748第1章緒論在現(xiàn)代的視頻監(jiān)控及多媒體應(yīng)用技術(shù)中，常常需要檢測出運動的人體或車體，并將其與背景分離。關(guān)鍵詞 圖像分割 陰影抑制AbstractIn the field of intelligent video surveillance, video technology, multimedia technology, often need to detect a human body or other objects, separate them with background, that is the context of solving realtime target segmentation. Video image object segmentation results, will target classification, tracking and behavior understanding such an important impact on subsequent processing. Image segmentation has been for many years in research attention, also raised thousands of algorithms. Goal of the current popular methods of segmentation, shadow detection, many are neglected, the goal is always to be detected, together with the shadow. The merger will cause the shadow of goals, objectives and some distortion of the shape of a serious problem, causing segmentation and tracking error. As the shadow directly affect target detection, a followup treatment effect affecting the key factors, the need for further research. The aim of this theory based on image processing based on some of the traditional edge detector is theoretically analyzed, using simulation experiments to test their effect on edge detection, contrast analysis of the effect of edge detection algorithm. Introduce some monly used color space and color space conversion algorithm, systematically expounded the various methods of image segmentation, analyzes and summarizes the advantages and disadvantages of several monly used segmentation. Use RGB color space, the background difference method using the initial segmentation of the image, then use region growing to remove the target of external noise, split the target image with a shadow. Then, the paper summarizes the basic assumptions shadow detection and the general framework of the current mainstream home and abroad shadow detection and suppression, that the goal of these methods for the removal of the existing problems in the shadow. Different images of the shadows and objectives of the body characteristics , be designed to remove the shadow of two algorithms. Based on Edge Information39。針對不同圖像的陰影和目標體的特點，擬設(shè)計一種去除陰影的算法。選用RGB彩色空間，利用背景差分法對圖像初步分割后，再利用區(qū)域生長法去除目標外部的噪聲，分割出帶影子的目標圖像。本課題擬根據(jù)圖像處理的理論基礎(chǔ)，對一些傳統(tǒng)的邊緣檢測算子進行了理論分析，用仿真實驗測試其邊緣檢測的效果，對比分析各邊緣檢測算法效果。陰影會引起目標的合并、目標形狀的失真等一些嚴重問題，引起分割和跟蹤錯誤。圖像分割多年里一直受到研究人員的重視，也提出了數(shù)以千計的算法。安徽建筑工業(yè)學院畢 業(yè) 設(shè) 計 (論 文)課 題 視頻序列圖像分割及陰影抑制 算法的研究 摘 要在智能視頻監(jiān)控領(lǐng)域、影視技術(shù)、多媒體應(yīng)用技術(shù)中，常常需要檢測出人體或其它物體，并將其與背景分離，即解決實時背景下目標的分割問題。視頻圖像的目標分割結(jié)果，將對目標分類、跟蹤及行為理解等后續(xù)處理產(chǎn)生重要影響?，F(xiàn)今比較流行的目標分割的方法，有不少是忽略陰影檢測的，目標總是與陰影一起被檢測出來。由于陰影直接影響目標的檢測，成為影響后續(xù)處理效果的關(guān)鍵因素，有必要進一步研究。介紹幾種常用的彩色空間以及彩色空間的轉(zhuǎn)換算法，系統(tǒng)地闡述了圖像分割的各種方法，分析總結(jié)了幾種常用分割方法的優(yōu)缺點。然后，分析總結(jié)了陰影檢測的基本假設(shè)和一般框架，及國內(nèi)外目前主流的陰影檢測與抑制算法，指出了這些方法用于去除目標陰影時存在的問題?；谶吘壭畔⒌年幱耙种扑惴ㄟm用于目標體