基于灰度投影和块匹配相结合的稳像算法.pdfVIP

基于灰度投影和块匹配相结合的稳像算法 TheCombinationof andBlock for Gray——ScaleProjection MatchingAlgorithmImage 尹宾。段会川 YlN BinDUANHui—chuan doi：10．3969／j．issn．1672—9528．2009．04．018 嚣饕暑荤囊在视频稳像过程中为兼顾算法的精度与效率，本文通过分析稳像的原理及目前常用的几种算法．提出了 一种基于灰度投影和块匹配相结合的匹配策略。首先运用灰度投影算法来缩小块匹配的待匹配区域， 然后运用块匹配算法进行运动估计与补偿。在此过程中，充分地体现了灰度投影算法运算量小以及块 匹配算法精度高的特点。实验结果表明在保持块匹配算法优势的同时降低了算法的运算复杂度。 罄安键舞道视频稳像灰度投影块匹配精度运算复杂度 AbstractFor theaffectof and inthe ofthe stabilization．In considering accuracyefficiencyprocessimage this the stabilizationofthe and several used paper，byanalyzingimage principlepresent commonlyalgorithms，A block with whichcombined and tobe matchingstrategy gray—-scMeprojectionblock——matchingalgorithmbrought using tonarrowthe tomatchtheblock up．Firstgray—scaleprojectionalgorithm question matchingregional，then formotionestimationand Inthis of characteristics usingblock—matchingalgorithm compensation．process，the thesmallcalculatedamountandthecharacteristicsof were reflected．Themanifest hilghaccuracyfully experiment thatthenew maintainthe ofthe atthesanletimereducethe algorithm advantagesalgorithm computingcomplexity． Video stabilization Block Precision Keywordsimage Gray—scaleprojectionmatching Computing complexity 进行逆向移动，补偿