基于阴影抑制的运动目标有效分割方法

针对背景差分法常将运动阴影误检测为前景目标的问题,提出了智能交通系统中一种基于阴影抑制的运动目标分割方法。在RGB空间通过可变阈值背景差分提取了运动目标和阴影;在HSI（Hue,Saturation,Intensity）空间,分别对黑色和非黑色运动目标采用不同的阴影检测条件进行了阴影抑制;结合形态学滤波和面积多尺度阈值滤波去除了噪声影响。实验结果表明本文提出的方法能有效去除阴影及噪声影响,解决目标断裂及黑色运动目标丢失问题,提高检测质量。

A Valid Segmentation Method for Moving Objects Based on Shadow Suppression

Aiming at the problem that shadows cast by moving objects in ITS （Intelligent Transportation System） are often detected as foreground target by background subtraction algorithm, a method of mobile objects segmentation is proposed based on shadow suppression. In tkGB color space, the moving objects and shadows are detected by background subtraction with alterable threshold. In HSI （Hue, Saturation, Intensity） color space, different shadow detection conditions are selected for shadow suppression though distinguishing black mobile objects and others. Moreover, morphological filter and multi-scale area filtering are combined to dispose of noise. The experimental results show that the method proposed in this paper can effectively get rid of shadow and noise, solve objects rupturing and black objects missing and improve detection quality.