采用 SURF算法对地面背景下红外目标识别。采用具有实时性的自适应中值滤波器和小波分频与直方图均衡的图像增强方法, 对图像进行预处理, 拉开目标与背景的灰度差异, 从而突出目标, 以便识别。对预处理完的图像采用 SURF特征提取匹配的方法进行红外目标识别。仿真实验中重点对 SURF特征匹配阈值和匹配特征点的数量进行了研究。实验表明, 文中方法对地面背景下红外目标识别效果较好。

SURF algorithm is used in the infrared target recognition in ground background. A real-time self-adaptive median filter and an image enhancing method with wavelet frequency division and histogram equalization are used to pre-process an image. It kicked a goal with gray background differences, and made a target more evidently to be identified. The SURF feature extraction and matching method can be used for preprocessed image to achieve infrared target recognition. The simulation experiments focus on the number of the SURF matching threshold and feature points especially. The experiments show that the proposed method in ground background infrared target recognition is good.