图像序列中目标关键帧快速搜索算法

在目标测量时所获得的图像序列中,如何定位目标关键帧(最有利于目标测量的图像)的位置,对目标识别的效率和测量设备的性能有着显著的影响.针对具有复杂特性的目标图像序列,提出了一种基于帧间像素灰度差值来定位目标关键帧的快速搜索算法.该算法仅仅利用像素灰度值这一最基本的特征,将图像序列中相邻两张图像的同一像素的灰度差值与给定阈值相比较,统计高于阈值的像素个数,再与另一给定阈值相比较,进而确定目标关键帧的位置.实验结果表明,该算法对目标大小不同、形状不同,环境不同,信噪比较高的图像序列都具有快速、稳定的搜索效果.

A Fast Search Algorithm for Target Key Frames in Image Sequence

In the image sequence, how to determine the key frames of moving targets that are most convenient to target detection is of much importance to both the effectiveness of target identification and the performance of measuring equipments. According to the image sequence of complex targets, a new fast search algorithm has been proposed based on the pixel gray difference of contiguous frames. After diverse experiments on various platforms under different circumstances, it’s showed that this search algorithm for target key frames in image sequence is simple, fast, stable and compatible for target identification application.