分层像素级图像融合中线性轮廓信息识别仿真  

Simulation of linear contour information recognition in hierarchical Pixel-Level image fusion

作　　者：石静泊 Shi Jingbo(School of Transportation Information,Shaanxi Communications Vocational and Technical College,Xi'an 710016,China)

机构地区：[1]陕西交通职业技术学院交通信息学院,陕西西安710016

出　　处：《现代科学仪器》2019年第4期54-58,共5页Modern Scientific Instruments

基　　金：陕西高等教育教学改革研究项目:基于现代学徒制的信息安全与管理专业校企协同人才模式研究项目编号:17GY017。

摘　　要：为提升分层像素级图像融合的效率和效果,文章提出了一种基于图像分割的线性轮廓信息识别方法,首先对模糊块状的源图像进行层次划分,通过灰度值的差值计算获得清晰的像素区域,以获得分层像素级图像的最佳融合区域;而后,根据传统Canny算子阈值设置不合理造成的轮廓边缘不连续或伪边缘的问题,引入最大信息熵进行阈值自适应选择,以改进Canny算子,去除伪边缘的问题,并融合自适应生长区域算法、Otsu阈值分割识别轮廓边缘的断点信息,解决轮廓边缘不连续的问题。最后,利用Matlab仿真分析,验证了改进Otsu阈值法和自适应阈值Canny算子的线性轮廓信息识别算法的准确度、完整性更高。To improve the efficiency and effect of hierarchical pixel-level image fusion,this paper proposes a linear contour information recognition method based on image segmentation.Firstly,the source image of blurred block is divided hierarchically,and clear pixel regions are obtained by calculating the difference of gray values to obtain the best fusion region of hierarchical pixel-level image.Then,according to the threshold of traditional Canny operator.The problem of contour edge discontinuity or pseudo-edge caused by unreasonable settings is solved by introducing maximum information entropy for threshold self-adaptive selection to improve Canny operator and remove the problem of pseudo-edge.The problem of contour edge discontinuity is solved by combining adaptive growth region algorithm and Otsu threshold segmentation to identify the breakpoint information of contour edge.Finally,the simulation analysis using MATLAB verifies that the improved Otsu threshold method and the adaptive threshold Canny operator are more accurate and complete.

关 键 词：像素级图像融合 线性轮廓 CANNY算法 改进的Otsu阈值法 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]