对于复杂云层背景下的红外弱小目标检测跟踪一直是图像处理的研究热点。受复杂云层和强光等因素的干扰,目标很容易淹没在背景中。针对传统方法无法兼顾较高的目标检测概率和较低的目标检测虚警率的问题,提出了一种目标检测跟踪方法。通过梯度预滤波的方法获取初始目标,二次比对的方法获取真实目标;通过目标航迹建立与删除的方法降低虚警率;通过抗云层遮挡卡尔曼滤波跟踪的方法提升目标检测能力。实验结果表明,该技术不仅很好地抑制了复杂云层背景和强光干扰背景,而且通过目标建航和抗遮挡跟踪等方法极大提升了目标检测概率,降低了虚警率。

The infrared dim small target detecting and tracking problem under complex cloud layer background has always been a hot issue in the study of image processing. By the interference of complex cloud and light, target is drowned in the background easily. In the view of the traditional methods which could not balance the high target detection probability and low false alarm rate, a series of target detecting and tracking strategies is proposed in this paper. The initial target positions are achieved by gradient pre-filter, and then they are confirmed by twice comparing. The false alarm rate is suppressed by the method of target path creating and deleting, and the target detecting ability is enhanced by the Kalman filtering tracking of cloud cover resistance as well. It is showed that, the proposed method not only inhibits the complex cloud background and strong interference background, but also greatly improves the target detection probability and reduces the false alarm rate by target navigation building and block tracking resistance in the experiments.