Abstract: Segmentation plays a key role in many image processing applications, including SAR image analysis. The role of image analysis is to provide an optimum representation of features. To achieve this, we propose a three-step algorithm for detection of water target contour in synthetic aperture radar (SAR) images. (1) Target regions are found by applying a point-based filter or an averaging-based filter. The filter is handled by partitioning an image into windows, realizing the filter in each window. Pixels are grouped into different target regions of similar characteristics. (2) A weighted statistical operator and algorithm of removing single point are used to reduce noise. They are done in the horizontal and vertical direction to eliminate false edges. (3) We use a statistical template to detect closed target contour.

摘要： 设计了基于图像分析理论，根据SAR图像特点形成水上目标轮廓的检测算法。该算法的实现分为3步：(1)基于数理统计的目标检测方法，形成目标区域；(2)用加权统计滤波算子和去除孤立点算法消除噪声，规范需要检测的水上目标；(3)提取目标轮廓。系统在不同的SAR图像上，取得了良好的检测结果。

关键词: SAR图像, 目标检测, 图像处理

CLC Number: 

• TN957.52