摘要
为了有效地进行高分辨率图像中线性目标的检测,提出一种基于Freeman链码的改进型Hough变换算法。首先,对图像进行增强和滤波处理,采用基于灰度一致化的方法对图像进行区域分割;然后,利用Freeman编码提取目标区域的边界;最后,对链码数据进行Hough变换,检测出平行线性结构。实验结果证明:该算法能有效地提取图像中平行线性目标,将其应用于资源三号卫星影像道路网目标的识别中,准确率高且实时处理性好。
This paper proposes a method for extracting linear object based on Freeman chain code and Hough transform with the purpose of extracting linear object effectively. After the original image is enhanced and filtered, a method based on the gray-level uniformization is used for region segmentation of image. Then the approach of Freeman chain code is carried out. Finally, the parallel linear structure is detected when Hough transform is used for the data of the chain code. The experiment results show that the proposed algorithm can extract the parallel linear structure of the image effectively, as evidenced by the fact that it showed high efficiency and high accuracy when it was applied to network target recognition in the ZY-3 satellite images.
出处
《国土资源遥感》
CSCD
北大核心
2014年第2期33-37,共5页
Remote Sensing for Land & Resources
基金
国家863计划项目"国家高技术研究发展计划"(编号:2011AA060203)资助
