海上小目标检测的视知觉方法

针对海上小目标检测问题，提出采用基于视觉非线性的图像分割方法，同时结合最佳分割区域自动调整算法，获得海上小目标的优良分割结果，并利用目标背景间存在较强反差这一先验知识，从分割所得二值图像中确定真实目标．算法中同时考虑了对亮、暗目标的处理，因而对目标的灰度变化不敏感．大量海上小目标检测的实验结果证实了本方法的有效性．

A Visual Perceptual Method for the Detectionof Small Targets on Sea-Surface

To detect small targets on the sea,an approach which adopts the image segmentation technique based on visual nonlinearity is proposed. An algorithm for the automatic adjustment of the bestsegmentation region to achieve fairly good segmentation results of small targets is developed. Then apriori knowledge that there exists a fairly sharp contrast between the targets and the background isused to determine the true targets from the segmented binary images. As the algorithm can deal witheither bright or dark targets, it is insensitive to the variation of the grey level of the targets. The effectiveness of the algorithm proposed is proved by a large number of experimental results.