摘要
提出一种防抖动区域自动搜索方法.对于防抖动区域的每个可能位置,计算匹配区域在该位置附近移动时与防抖动区域之间的相似性测度函数值,可知该函数在多大范围内具有单峰特性.若单峰范围满足工程要求则将该防抖动区域作为候选区域,由操作人员根据它是否有可能受到运动物体的影响来决定它是否可用.防抖动区域的尺寸从一个比较小的值逐渐增大,直到搜索到可用的防抖动区域为止.实验结果表明,这样做既可以减少陷入局部最优又可以提高抖动消除速度.
An automatic method was proposed for searching anti-shake regions.For each possible position of an anti-shake region,the matching region was supposed to move around the position,and the value of similarity measure function between these two regions was calculated for each movement.With these values,the range within which the function with only one extreme value was determined.If the range was wide enough for practical utility,the anti-shake region was presented as a candidate to the operator.To adopt it or not according to whether it was possible to be affected by moving objects.The dimension of the anti-shake region was increased from a small value,until a usable anti-shake region was found.Experimental results showed that this method could decrease the chance of falling into local optimum and increase the speed of video stabilization.
出处
《郑州大学学报（理学版）》
CAS
北大核心
2014年第2期59-63,共5页
Journal of Zhengzhou University:Natural Science Edition
关键词
图像处理
相似性测度函数
单峰范围
图像匹配
视频抖动消除
image processing
similarity measure function
range of single extreme value
image matching
video stabilization
