船载雷达遥感图像的油膜快速识别方法

Fast oil film identification of shipborne radar remote sensing image

在溢油事故发生后,如何有效监测海上油膜是应急处置工作的重要保障。本文提出了一种船载导航雷达遥感图像的油膜快速识别方法:首先,采用垂直噪声检测算子与Otsu方法,提取原始图像中的同频干扰噪声;其次,应用线性内插平滑同频干扰噪声;再次,采用灰度调节矩阵与对比度受限自适应直方图均衡化方法,增强油膜区域内外的对比度;最后,对图像全局应用直方图双峰阈值法,快速提取海上油膜。实验表明,对比其他全局自适应阈值法,直方图双峰阈值法在油膜识别方面具有较强的适用性;对比局部自适应阈值法,本方法具有较高的效率。这种方法的成功应用,将为海上溢油的实时治理工作提供快速的数据支持服务。

After the oil spill accident,effective monitoring of offshore oil films is an important guarantee for emergency disposal.A fast oil film identification method of shipborne navigation radar remote sensing image is proposed here.Firstly,the vertical noise detection operator and Otsu method are used to extract the cofrequency interference noises in the original image.Next,linear interpolation is applied to smooth the cofrequency interference noises.Then,the gray adjustment matrix and Contrast Limited Adaptive Histogram Equalization method are used to enhance the contrast around the oil films.Finally,the histogram bimodal method is applied globally to extract offshore oil films quickly.Compared with other global adaptive threshold methods,histogram bimodal method has strong applicability in oil film identification.Compared with the local adaptive threshold method,the proposed method has faster efficiency.The successful application of the proposed method will provide rapid data support service for the real-time treatment of oil spills.