基于局部熵滤波和梯度累积的海天线检测方法

Sea-sky line detection method based on local entropy filtering and gradient accumulation

针对海天线检测中海天背景红外图像对比度低、图像模糊及云层和海杂波亮带等干扰问题,提出一种基于局部熵滤波和梯度累积的海天线检测方法.首先对海天背景红外图像进行局部熵滤波从而进行图像增强;然后对滤波后的图像求取梯度图像,并将梯度图像区域累加结果的极大值位置作为海天线候选位置;最后在候选海天线位置附近求原始图像梯度极大值点,并通过随机抽样一致(RANSAC)算法拟合直线,将RANSAC算法输出内点多的直线作为最终的海天线.实验结果表明:所提出的算法能够在保证时效性的同时提升复杂干扰背景下海天线检测的准确率.

Aiming at the interferences in infrared sea-sky background images,such as low contrast,image blurring,cloud edge,sea clutter and bright bands,a sea-sky line detection method based on local entropy filtering and gradient accumulation was proposed.First,the local entropy filtering was performed on the sea-sky background image for image enhancement.Then,the gradient image was obtained from the filtered image,and the local maximum value position of the gradient image accumulation result was considered as the candidate position of the sea-sky line.Finally,the local maximum gradient points of original image near the candidate sea-sky lines were used to fit lines by random sample consensus(RANSAC) algorithm,and the line with the most interior points was determined as the sea-sky line.Experimental results show that the proposed algorithm can improve the accuracy of seasky line detection under complex interference background while ensuring the timeliness.