在点目标探测模型中，目标的形状纹理信息丢失，可用特征少，背景杂波对目标检测造成严重干扰。传统分块方差统计法会引入背景低频波动，使杂波偏大，而且杂波度量值随分块单元尺度变化较大。本文提出一种基于匹配滤波的杂波量化改进方法，首先以点目标像斑分布为模板，对背景图像进行匹配滤波，得到互相关矩阵；然后将高相关性区域标记为杂波重点区域，计算方差作为局部杂波，取局部杂波的均方根作为全局背景杂波。该方法突出了对空间杂波区域的选择性，能有效降低背景低频波动和传统分块尺度对杂波的影响。对遥感图像样本进行杂波量化和点目标检出仿真，结果表明本文方法对杂波统计尺度有很好的稳定性，杂波量化与探测率的相关性高于传统方差统计方法。

In the point target detection model, the shape and texture information of the target is almost lost, and the number of characteristics available for the detection is small, so clutter caused by background may lead to serious interference in post-processing. The traditional block variance statistic method takes low-frequency fluctuation into account, which makes the clutter larger than its real value, and the clutter quantification result varies greatly with unit scale. In this paper, an improved clutter quantification method based on matched filtering is proposed. First, the background image is filtered with a matched mask obtained by a target speckle, and the result is the cross correlation matrix of the background to the target. Then, regions with high correlation are selected to calculate the variance as a local clutter. At last, the root mean square of the variances are taken as a global background clutter. The improved method filters the background to select a reasonable clutter region. It can effectively reduce low-frequency fluctuation of the background, and the result is stable to block the scale. Clutter quantification and the point target detection simulation are conducted. The results show that the method proposed in this paper is robust to unit scale. Compared to the traditional method, it has higher correlation between the clutter and the probability of detection.