提出了一种用于运动阴影检测的Boosting判别模型.这种方法先利用Boosting在不同的特征空间来区分前景和阴影,然后在判别随机场(DRFs)中结合前景和阴影的时空一致性,实现对前景和阴影的分割.首先,差分前图像与背景图像得到颜色不变子空间和纹理不变子空间 然后在这两个子空间上应用Boosting来区分前景和阴影 最后利用前景和阴影的时空

一致性,在判别随机场中通过图分割的方法准确地分割前景和阴影

-A new method for the detection of movement of the shadow of Boosting discriminant model. This approach to the use of Boosting in different feature space to distinguish between prospects and the shadow, and then identify with the airport (DRFs) in light of the prospects and the shadow of the temporal and spatial coherence, the realization of shadow on the prospects and the partition. First of all, the difference with the background image before the image to be the same color and texture subspace invariant subspaces　and then the application of these two sub-space to distinguish between Boosting prospects and shadow　Finally, the prospects and the shadow of the space-time consistency in the identification with the airport through the partition plan divided the method is accurate and the shadow of the prospects