近海岸目标快速提取

针对海面运动载体对复杂海岸背景下的水上目标检测问题,提出了基于海岸线信息的近海目标自动检测方法.通过量化子图像的区域复杂度以及单元区域上下邻域的灰度差异,提取海岸线区域的主要轮廓.采用哈夫变换进行投票加权处理,确定海岸线的精确位置,使用主频灰度对对海岸线以下的海面部分进行滤波,消除杂波效应,引入聚类方法,剔除伪目标,实现目标的有效区分.试验证明所提方法能够检测出不同倾斜状态下的海岸线,并实现精确的目标定位,单帧处理在0.2s以内,具有准确性和快速性.

Fast extraction of offshore targets based on sea line detection

A feasible method based on sea line extraction was proposed to automatically detect targets above water in the context of complex offshore situations for surface vehicles.The method was carried out as follows.The regional complexity of sub-images and the average gray difference of their up and down neighborhood were quantified to predict the sea-sky region.The important contour of this region was obtained.A weighted vote in the Hough transform was introduced to pick the exact line.The sea part below the sea line was filtered by the main gray value to remove clutter effect.The smoothed image was clustered and pseudo targets were excluded effectively.The experimental results prove that the present method can detect the sea line under different tilted conditions and locate the targets exactly.The time cost per frame is within 0.2s.It is robust and fast.