基于光照不变性模型的运动目标检测新算法

运动目标检测有两大难点，即光照变化的影响，阴影对运动目标准确提取的影响。高志伟等人提出了基于彩色边缘的运动车辆检测，实现了复杂背景下的运动目标检测，但该方法无法消除阴影的影响。为了克服光照及阴影的影响，提出了基于一维不变性图像的背景模型，更新了最小熵投影角度，设计了运动目标检测的新算法。以交通运输领域为例，将本文算法和多层前景算法、边缘检测算法做了对比，通过实验验证了该方法在检测运动目标时能够克服光照变化影响,并有效抑制阴影。

A new algorithm for moving target detection based on illumination invariant

Moving object detection has two difficulties: illumination variant and shadows. Moving vehicles detection algorithm based on color edge was presented by Gao zhiwei et, which actualizes the detection of moving object under complex background. However, this method cannot eliminate the influence of shadow. In order to overcome the influences from illumination and shadows, the background model based on illumination invariant is proposed, the minimum entropy projection angle is updated, and a new detection algorithm for the moving target is devised. Take the traffic transportation as an example. Results from the experiments verify that this method can eradicate the influence of lighting variances, restrain shadows effectively .