欢迎访问《图学学报》 分享到：
图学学报

图学学报

• 专论：第30届计算机技术与应用学术会议 (CACIS 2019 雅安) • 上一篇    下一篇

基于卡尔曼滤波的 SAR 图像边缘检测方法

  

  1. (西安科技大学计算机科学与技术学院，陕西 西安 710054)
  • 出版日期:2019-10-31 发布日期:2019-11-06
  • 基金资助:
    陕西省教育厅科研计划项目(17JK0513)；陕西省自然科学基础研究计划项目(2019JM-162)；西安科技大学博士启动金项目(2019QDJ007)

Edge Detection for SAR Images Based on Kalman Filter

  1. (College of Computer Application Technology, Xi’an University of Science and Technology, Xi’an Shaanxi 710054, China)
  • Online:2019-10-31 Published:2019-11-06

PDF (PC)

300

摘要： 针对传统 Canny 边缘检测算法对合成孔径雷达(SAR)图像的相干斑噪声抑制程度 太高，导致大量边缘的真实信息丢失问题，提出一种新型 Canny 算子边缘检测算法。首先建立 合适的非对称半平面区域(NSHP)图像模型，将空间模型转换成卡尔曼滤波可适用的系统状态方 程；然后用“预测+反馈”的方式对图像去噪；最后通过双阈值算法提取图像的边缘。仿真实验表 明，该方法可以有效地抑制 SAR 图像中的相干斑噪声，同时能较好地保留图像的边缘信息，相 对于传统的 Canny 算法有较好的检测效果。

关键词: Canny 算子, 边缘检测, NSHP, 卡尔曼滤波, SAR 图像

Abstract: Traditional Canny edge detection algorithm suppresses the speckle noise of synthetic aperture radar (SAR) images too much, causing much loss of real edge information loss. To tackle this problem, this paper proposed a new Canny operator edge detection algorithm. Firstly, the method established a suitable non-symmetric half plane (NSHP) image model, then converted the spatial model into a system state equation applicable to Kalman filter; after that, then we adopted the method of prediction and feedback to denoise the image. Finally, the edge of the image was extracted by dual threshold algorithm. Experimental results show that the proposed method can effectively suppress the speckle noise of the SAR image and preserve the edge information well, and provide better detection effects than traditional Canny algorithm.

Key words: Canny operator, edge detection, NSHP, Kalman filter, SAR image