针对星图中空间点状目标占有像素数少、信噪比低的问题,提出一种基于生物侧抑制原理的星图点目标检测方法。首先通过多级滤波对图像进行预处理,对信噪比较低的星图进行杂波抑制,其次采用一种基于生物视觉侧抑制理论的背景抑制网络模型进行二次滤波,最后采用二维熵分割提取图像中点目标。实验结果表明,该方法能够在有效地抑制图像中的强杂波背景的同时较好地增强目标强度,并且具有较高的检测率和较好的抗噪能力。

To detect star targets in shortwave infrard images, a background inhibition network modeling is firstly built on the theory about the lateral inhibition of biology visions. Secondly, a novel region of interest extraction method which was built on the RENYIs entropy is proposed. Finally, a new viewpoint filtering method to reduce the false star target of the regions of interest is suggested, which was by the means of statistic filtering based on the pre-knowledge of star targets in the low resolution SAR images. The test results indicate that the lateral inhibition method can not only suppress the background and clutters well, but also can enhance the star target perfectly. The extracting method about regions of interest and false alarm reducing method also have good performance as well as adaptive characteristics in our experiment.