复杂背景下红外图像目标的快速定位

摘要

Under the complicated background of infrared images ,it is impossible to use the image seg-mentation method to extract the target position quickly and accurately .For this reason ,a secondary classification method based on the clustered histogram is proposed .The method uses the statistical histogram characteristics of multi-dimensional image module as the basis for measurement . The change of the contrast between the images to the comparison of statistical characteristics between the histograms will help distinguish a target from the background information quickly .The result of real images testing show s that the method can be used to determine the position of the target effectively and that it can reduce the processing time when used together with other segmentation algorithms . T herefore ,the method is suitable for engineering application .%针对复杂红外背景下现有图像分割算法不能对目标位置进行快速提取这一问题，提出了一种基于聚类直方图的二次分类方法，即通过将滑动模块间多维统计特征作为度量依据，将图像对比转化为对直方图间分类统计特征的比较，从而实现对缺乏明确轮廓的目标的快速位置估计。对真实图像进行的试验表明：该方法能够有效获取目标位置，在结合其他分割算法时能够降低算法处理时间，适合工程应用。