红外小目标检测的滞后阈值分割法

为提高红外小目标检测的精度,针对小目标的阈值分割,提出滞后阈值分割法。通过快速非最大值抑制预提取红外小目标的中心像素,最大程度缩减前景像素数。利用邻域梯度方差判别小目标是否为杂波,自适应地调整灰度阈值,完成对小目标的低虚警分割。分析一般场景下分割所得前景、背景的组成,给出了滞后阈值分割的阈值取值方法。实验表明滞后阈值分割法将虚警概率限制在很小的范围内,并取得了优于其他分割法的检测概率。在小目标检测中,该算法具有较强的实用性。

Hysteresis threshold segmentation in infrared small target detection

In order to improve the accuracy of infrared small target detection,the hysteresis threshold segmentation is proposed for threshold segmentation of small targets.It pre-extracts the central pixel of the infrared small target by fast non-maximum suppression,minimizes the foreground pixel number.The neighborhood gradient variance is used to decide whether the small target is clutter.The gray threshold is then adjusted adaptively,and the small target is segmented with lower false alarm probability.Based on the analysis of the components of foreground and background in the general scene,the threshold value for the small target segmentation is given.Experiments show that the hysteresis threshold segmentation method limits the false alarm probability to a small range and achieves a detection probability superior to other segmentation methods.On the small target detection,the algorithm has strong practicability.