摘要
超宽带雷达成像结果中,目标通常占据多个像素点,且目标形状多变、图像动态范围大,为目标检测带来了显著挑战。针对超宽带雷达扩展目标检测问题,本文提出了一种基于隐马尔可夫随机场的检测方法。该方法通过马尔可夫随机场表征目标图像的空间特征,结合高斯混合模型描述目标图像的幅度特征,并构建似然比检测器实现目标检测。与现有基于幅度的检测算法相比,本方法同时融合了幅度与形状特征,在背景与杂波未知的情况下,实现了更为稳健的多目标检测。
Target detection in ultra-wideband radar systems encounters significant challenges,including target extension,shape variability,and wide dynamic range.To address these issues,this paper proposes an extended target detection algorithm based on the hidden markov random field.The proposed algorithm employs a likelihood ratio test framework,wherein the geometrical features of the image are modeled using the HMRF,while the amplitude features are characterized by Gaussian mixture model.In contrast to existing methods,the proposed algorithm effectively integrates both intensity and spatial features of the image,enabling robust detection performance without requiring prior knowledge of clutter or noise characteristics.
作者
李虎泉
Li Huquan(Nanjing Research Institute of Electronics Technology,Nanjing,China)
出处
《科学技术创新》
2025年第20期100-103,共4页
Scientific and Technological Innovation
关键词
超宽带雷达
扩展目标检测
隐马尔可夫随机场
高斯混合模型
ultra-wide band radar
extended target detection
hidden markov random field
gaussian mixture model
