摘要
通过图像的边缘融合,挖掘图像中亮度变化明显的点,提高对远程图像的视觉特征分辨能力。传统的图像边缘融合算法采用灰阶量化边缘分解技术,由于图像边缘编码向量在码书中的排列是无序,导致边缘融合效果不好。提出一种改进的基于向量量化谱分解的图像边缘融合算法。采用向量量化谱分解技术,对信号与图像数据进行压缩,生成融合图像的灰度直方图,构建图像的向量量化边缘融合算子,实现算法改进。仿真结果表明,采用该算法能有效检测出图像的边缘亮点结构,保留了图像重要的结构属性,实现对图像边缘的准确检测,提高了峰值信噪比20 d B,展示了较高的边缘融合质量。
Through the edge of the image fusion, image mining in brightness changes obviously, improve the visual features of remote image resolution. Traditional image edge fusion algorithm using gray level quantization edge decomposition tech-nique, due to the arrangement of image edge encoding vectors in the codebook is in disorder, lead to the edge of fusion ef-fect is not good. A vector quantization of spectral decomposition of image edge fusion algorithm is proposed based on im-proved. Vector quantization of spectral decomposition technique, to compress the signal and image data, gray histogram gen-eration fusion image, construct the vector quantization of image edge fusion operator, improved algorithm. The simulation re-sults show that, by using this algorithm can effectively detect the edge of spot image structure, retains the structure proper-ties of important image, to achieve accurate detection of image edge, improves the peak signal-to-noise ratio of 20 dB, it has the high edge fusion quality.
出处
《科技通报》
北大核心
2015年第10期142-144,共3页
Bulletin of Science and Technology
基金
基于大功率LED宽幅数码冲印系统的研制(GJJ13571)
江西省教育厅科学技术研究项目(0123ss-0127ly)
关键词
向量量化
图像处理
边缘融合
vector quantization
image processing
edge fusion