摘要
为了降低图像序列运动目标检测中背景重构的时间复杂度和空间复杂度,提出一种基于子图像块归类的背景重构算法。在假设背景图像块以最大概率出现在图像序列中的前提下,选择编号频率最高的图像块作为背景图像块进行背景重构。在运动目标有较长的暂时停顿情况下,即观测长度较长时利用该算法进行实时背景重构具有明显的优点。仿真结果表明,该算法能够准确地重构背景,并有效地避免混合现象,从而实现对运动目标的完整提取,以便进一步识别或跟踪。
In order to reduce the time and space complexity of the background reconstruction in the image sequence moving target detection, we proposed a background reconstruction algorithm on the basis of classified sub-image blocks. In the premise that background image blocks appear with the greatest probability in image sequence, the highest frequency code block is selected as a background image to conduct the background reconstruction. When moving target is in a longer suspension of the circumstances, that is, the length of observation time becomes longer, using the real--time background reconstruction using the algorithm has obvious advantages. Simulation results show that the algorithm can accurately reconstruct background, and effectively avoid the phenomenon of mixing, thereby to achieve a complete extraction of moving targets so as to further identify or track.
出处
《太原理工大学学报》
CAS
北大核心
2009年第6期586-588,共3页
Journal of Taiyuan University of Technology
基金
国家自然科学基金资助项目(60843006)
山西省自然科学基金资助项目(20051037)
关键词
背景图像提取
分块图像
运动检测
background image extraction
block images
motion detection