论文部分内容阅读
首次将局部统计活动轮廓模型引入SAR图像海岸线检测问题中,提出了一种基于局部统计活动轮廓模型的SAR图像海岸线检测方法。首先利用C-V模型进行粗分割,消除局部统计活动轮廓模型对初始轮廓线设置要求严格的限制,然后提出了一种基于G0分布的局部统计活动轮廓模型,进行精细分割。该模型采用G0分布对轮廓线上每一点的邻域进行统计建模,增强了模型数据拟合能力,提高了海岸线检测精度,加入水平集函数惩罚项,消除了重新初始化过程。实测SAR图像实验表明,本文方法可用于精确海岸线检测。
For the first time, the local statistical activity profile model is introduced into the SAR image coastline detection problem, and a SAR image coastline detection method based on local statistical active contour model is proposed. Firstly, the C-V model was used for rough segmentation to eliminate the restriction of the local statistical active contour model to the initial contour. Then, a local statistical active contour model based on G0 distribution was proposed for fine segmentation. The model uses G0 distribution to statistically model the neighborhood of each point on the contour line, which enhances the model data fitting ability, improves the coastline detection precision, adds the level set function penalty item, and eliminates the reinitialization process. Experimental SAR image experiments show that this method can be used for accurate coastline detection.