基于朗斯基函数的混合高斯模型运动目标检测
发布时间:2018-06-19 23:47
本文选题:混合高斯模型 + 运动目标检测 ; 参考:《计算机应用研究》2016年12期
【摘要】:针对传统的混合高斯模型在进行运动目标检测时存在拖影和性能差的缺点,提出了一种融合朗斯基函数和帧间差分法的混合高斯背景建模算法。该改进算法通过朗斯基矩阵行列式判断相邻像素间空间域相关性,以此增加模型参数更新条件,改进模型参数更新机制;并利用帧间差分法检测运动目标轮廓的灵敏性,将两种检测结果布尔或运算,完善目标轮廓。实验结果表明,该改进算法对拖影现象达到很好的抑制作用,并使算法检测性能得到提高。
[Abstract]:Aiming at the drawback and poor performance of the traditional hybrid Gao Si model in moving target detection, a hybrid Gao Si background modeling algorithm combining the Ransky function and the inter-frame difference method is proposed. The improved algorithm determines the spatial correlation between adjacent pixels by Ransky matrix determinant to increase the updating condition of model parameters, improve the updating mechanism of model parameters, and detect the sensitivity of moving target contour by inter-frame difference method. Two kinds of detection result Boolean or operation are used to perfect the target profile. The experimental results show that the improved algorithm can effectively suppress the drag phenomenon and improve the detection performance of the algorithm.
【作者单位】: 河北工业大学电子信息工程学院;
【基金】:河北省高等学校科学技术研究青年基金资助项目(Q2012012)
【分类号】:TP391.41
,
本文编号:2041886
本文链接:https://www.wllwen.com/kejilunwen/ruanjiangongchenglunwen/2041886.html