摘要
在数字图像处理领域,形心跟踪的目标捕获有着广泛的应用,尤其在对于天空等相对简单均匀的背景。形心跟踪算法通过目标与背景在灰度上的差别确定是否是目标,其具有较高的跟踪准确性和稳定性,容易实现目标自动捕获,但当图像背景不均匀而且目标与背景灰度相近时,会严重影响对目标的捕获,而不均匀背景在日常是不可避免的。本文分析了有关形心跟踪算法的特点和不均匀背景图像的特点,提出了基于微分视场的不均匀背景图像目标捕获算法,并应用在实际的工程应用系统中,该算法把整个视场图像分为多个更小的图像块分别进行处理,然后对处理结果进行综合分析,获得真正的目标;通过实验和在工程中应用证明,该算法能较好地克服不均匀背景对弱小目标的捕获跟踪影响,改善弱小目标的捕获效果。
The centroid target tracing algorithm is widely used in digital image process area, especially when the background is simple and even like sky. The centroid trace algorithm determines if there are targets in the image by the gray difference of target and background. It has higher trace precision and stability and can realize automatic target capture. But when the image has uneven background and the target and background is closing in gray - level, which is unavoidable in some cases, the target capture capacity will be affected severely. The characteristic of centroid algorithm and uneven background image is analyzed, and a new algorithm based on mini - field to capture target in uneven background is proposed. The new algorithm has been applied in real project. The algorithm divides the whole view field image into sev- eral small blocks and process these image blocks respectively, then analyze the process results to obtain real target. The experiments and project applications are proved that the algorithm can overcome the influence of uneven background on weak target capture and improve the capture effect of weak target.
出处
《长春理工大学学报（自然科学版）》
2007年第2期26-29,共4页
Journal of Changchun University of Science and Technology(Natural Science Edition)
基金
中科院二期创新项目(C04708Z)
关键词
形心跟踪
目标捕获
不均匀背景
实时跟踪
centroid trace
target capture
uneven background
real -time tracing
