摘要
通过图像边缘融合提高图像的成像质量和识别能力。传统的图像边缘融合方法采用小波包加权滤波方法,当图像出现丰富结构信息或者干扰较强时,融合效果不好。提出一种基于灰度窗口相关系数匹配的图像边缘融合算法。实现图像的边缘融合的基础是进行图像归一化分割,设计边缘特征提取方法,得到的图像边缘进行Hough变换直线检测,提取出直线段,通过超像素网格的形成的方法绘制图像的边缘融合边界寻优路径,得到超像素网格,实现了基于灰度窗口相关系数匹配的图像边缘融合处理。仿真结果表明,该算法的对图像的边缘融合效果较好,度量了区域间的差异和区域内的相似性,提高图像的分割和边缘融合质量。
To improve the imaging quality and the ability to identify image by image edge fusion. By using the wavelet pack?age weighted filtering method for image edge fusion of traditional method, when the image appears abundant structure infor?mation or interference, the fusion effect is not good. Put forward a kind of image fusion technology, gray edge correlation co?efficient based on windows. The foundation of realizing fusion image edge is the image normalization, edge feature extrac?tion method of design, Hough transform to detect straight edge image obtained, extracting the straight line segment, formed by the method of super pixel grid rendering image edge fusion boundary path optimization, get the super pixel grid, realize image edge matching correlation coefficient based on the gray window fusion. The simulation results show that, the algo?rithm of image edge fusion effect is better, to measure the similarity of regional differences and within the region, improve the image segmentation and edge fusion quality.
出处
《科技通报》
北大核心
2015年第6期148-150,共3页
Bulletin of Science and Technology
关键词
灰度窗口
相关系数
图像融合
grey window
correlation coefficient
image fusion
