论文部分内容阅读
提出了基于背景预测的单帧检测弱小目标改进算法.该算法采用改进的最大值背景预测算法.为了能同时检测点目标和小目标,预测背景的四个预测窗的位置与目标点之间保持一定的距离,保证预测窗中的像素能准确反映当前像素灰度中的背景灰度信息.为了保证单帧的较低的虚警率,同时结合多方向多极梯度算法,对背景预测初步检测出来的可能目标点进一步检测,以排除绝大多数的虚假目标.经实验证明,该算法在单帧图像上能很好的检测出弱小目标来,并能满足实时性的要求.
This paper proposes an improved single-frame detection weak target algorithm based on background prediction, which uses an improved maximum background prediction algorithm.In order to be able to detect both point targets and small targets simultaneously, the position of the four prediction windows of the prediction background and the target point are maintained A certain distance to ensure that the pixels in the prediction window can accurately reflect the background gray information of the current pixel gray.In order to ensure the low false alarm rate of single frame and combined with the multi-directional multi-gradient gradient algorithm, the initial detection of background prediction The possible target points are further detected to eliminate most of the false targets.Experimental results show that this algorithm can detect weak targets in single-frame images and meet the requirements of real-time performance.