研究了地面背景下的红外目标识别技术。首先，从提高中值滤波实时性的角度考虑，提出了改进的自适应中值滤波算法，用改进算法对图像滤波；然后，通过空域图像增强方法和基于数学形态学的图像增强方法相结合的方式，提出一种针对地面目标的图像增强算法，拉开目标与背景的灰度差异、突出目标；最后，用基于SIFT特征提取的图像配准方法对增强后的地面背景下的红外目标进行识别（采用MATLAB进行编程仿真试验），匹配过程中重点讨论了匹配阈值的选择问题。实验结果表明，应用该方法对地面目标进行识别的效果比较好，具有一定的实用性和可靠性。

The paper mainly focused on infrared target recognition on the ground. Firstly, an improved adaptive median filter algorithm was proposed to enable the real-time processing and was used to filer the infrared image of the target. Secondly, an approach to enhance image of infrared target was presented, which combined mathematical morphology with image enhancement in spatial domain. It widened the gray level and highlighted the target. Finally, a image matching method based on SIFT (Scale Invariant Feature Transform) was used to recognize the ground target (simulated by MATLAB). In the matching process, we focused on the selection of match threshold. The experimental results show that this method is practical and reliable.