基于YUV空间球状码元码本背景模型的运动物体检测算法研究

在光照变化强烈的场景中,传统的码本背景模型对运动目标检测效果较差,该文融入了梯度信息来改变传统码本背景模型中的亮度范围,对传统度量颜色距离的柱状单词结构做了改变,用YUV颜色模型建立码本背景模型,降低了传统的码本背景模型在光照变化比较频繁的场景中的敏感性。通过改进的算法对一组典型的复杂背景视频图像对进行验证,证明了改进的球状码本背景模型的算法确实能够取得更好的检测效果。

Based on YUV Space Spherical Symbol Codebook Background Model Detection of Moving Objects

When the shining light changes greatly,codebook background model is less effective for the strong scene of moving target detection.The integration of gradient information is introduced to the traditional codebook background model to change the traditional range of brightness.And the traditional measure of the color word from columnar structure is also changed with the YUV color model to create a codebook background model,which reduces the sensitivity of frequent changes in light shining that holding in the traditional code of the background model.And through the application of a typical set of background video to the improved algorithms,the validation gives a proof of the improved algorithm which can indeed achieve better target detection results.