Abstract: The visible image background clutter is correlative in space and time domain. To detect small moving targets in complex visible image background clutter, the spatial-temporal difference method is presented. Based on the visible image sequences model, the characteristic and difference of targets, background and noise was analysised qualitatively. With these, the spatial-temporal difference method (STDM) for background suppression was presented and also it’s realization via assembly line and pipeline scheme was given. Many experiments prove the availability in practice.

Key words: Visible image, Background clutter, Target detection, Spatial-temporal difference method, Pipeline scheme

摘要： 可见光图像背景常常表现出时域和空域的双重相关特性。针对可见光图像背景下的运动小目标检测，提出了一种空时域联合差分的检测方法。以可见光图像序列模型为基础定性分析了目标、背景和噪声的基本特性和差异；在此基础上提出了空时域联合差分的背景抑制方法，结合流水线结构给出了一种实时性好的具体实现方案。实验结果证实了该方法的有效性。

关键词: 可见光, 背景起伏, 目标检测, 空时域差分, 流水线结构