提出了一种基于多特征距离图的红外弱小目标检测方法。弱小目标的许多特征,如局部熵、平均梯度强度等,不但刻画了弱小目标的特点而且易于提取。通过特征融合技术,可以将弱小目标检测问题转化成在一个多特征空间的极值求取问题。该方法利用已经提取的多个特征,采用特征融合技术构造一个距离图像,再对该图像进行二值化处理,达到目标检测的目的。通过对实际的红外图像序列进行小目标检测,验证了所提方法的可行性和有效性。 A weak infrared target detection method based on multi-feature distance map is proposed. Many features of weak targets, such as local entropy, average gradient strength, etc., not only portrayed the characteristics of weak targets and easy to extract. Through the feature fusion technique, we can transform the weak and small target detection problem into the extreme value problem in a multi-feature space. This method uses a number of features that have been extracted and uses a feature fusion technique to construct a distance image, and then binarizes the image to achieve the goal of target detection. Through the small target detection of the actual infrared image sequence, the feasibility and effectiveness of the proposed method are verified.