CAJ | 학술논문

针对子块级背景建模方法无法保证所提取前景形状的精确性及像素级背景建模方法无法有效处理非平稳场景的问题,提出了一种背景建模分层模型,首先采用文中子块级建模算法得到较为粗糙的背景区域和前景区域,然后利用混合高斯模型对特定图像区域执行像素级的前景提纯或背景模型更新操作,2种不同层次的算法通过非对称前向反馈机制进行级联.实验结果表明,所提分层模型在能够有效处理非平稳场景的同时保证了所提取前景形状的精确性,且对光照突变不敏感,建模效果优于级联算法中任一独立算法,而处理时间小于2种独立算法处理时间之和,满足了实时处理要求.
침대자괴급배경건모방법무법보증소제취전경형상적정학성급상소급배경건모방법무법유효처리비평은장경적문제,제출료일충배경건모분층모형,수선채용문중자괴급건모산법득도교위조조적배경구역화전경구역,연후이용혼합고사모형대특정도상구역집행상소급적전경제순혹배경모형경신조작,2충불동층차적산법통과비대칭전향반궤궤제진행급련.실험결과표명,소제분층모형재능구유효처리비평은장경적동시보증료소제취전경형상적정학성,차대광조돌변불민감,건모효과우우급련산법중임일독립산법,이처리시간소우2충독립산법처리시간지화,만족료실시처리요구.
Block-based background modeling couldn’t obtain the exact shape of foreground, while pixel-based approach-es couldn’t handle non-stationary backgrounds effectively. To solve the problem, a hierarchical scheme for background modeling was presented. The hierarchical model used block-based method proposed to obtain coarse background and fo-reground regions firstly, and then the operations of pixel-level foreground refining and model updating based on Gaussian mixture model were performed on special regions of the input image. These two algorithms in different levels were com-bined by adopting an asymmetric feed-forward strategy. Experimental results show that the hierarchical method proposed can obtain the exact shape of foreground and process non-stationary scenes well, in addition, it is insensitive to illumina-tion change and can provide better results than any single approach in it, meanwhile, the integrated computation time is shorter than the sum of those of running the block and pixel-level methods, and satisfies real-time processing.