人工智能
计算机视觉
计算机科学
特征提取
轮廓波
图像分割
图像分辨率
斑点图案
散斑噪声
合成孔径雷达
滤波器(信号处理)
模式识别(心理学)
分割
遥感
地理
小波变换
小波
作者
Wenhao Shu,Yuxuan Bie,Ruigang Zhang,Ying Sun,Zhaoyang Zeng,Hui Wang
标识
DOI:10.1109/ciss60136.2023.10380023
摘要
As an important ground object, remote sensing image road extraction is a very popular research subject and widely used. With the continuous development of SAR imaging technology, its spatial resolution has become higher and higher, and the road in the image is no longer displayed as a linear feature in low resolution, but a narrow and long planar feature, which leads to the previous low resolution road extraction method cannot be effectively used in high resolution, and a new high resolution road extraction method is needed. In this paper, a high-resolution SAR image road extraction method based on MRF model is proposed. The Contourlet threshold filter is used to reduce the impact of speckle noise, image segmentation is performed using MRF model, and road edge information is extracted based on Canny operator. Finally, geometric feature constraints and edge connections are performed. The results verify the accuracy and stability of the road extraction method proposed in this paper.
科研通智能强力驱动
Strongly Powered by AbleSci AI