基于融合及形态学的自适应阈值图像边缘检测  

Morphological and Adaptive Threshold Algorithm of Image Edge Based on Data Fusion

作　　者：祁佳[1] 刘紫燕[1] 

机构地区：[1]贵州大学电子信息学院,贵州贵阳550025

出　　处：《电视技术》2014年第13期36-38,共3页Video Engineering

基　　金：贵州省科学技术基金项目([2011]2193)

摘　　要：传统的Sobel和Roberts边缘检测算子阈值固定,适应性和抗噪能力较差。通过将最大类间方差法和双峰法引入边缘自适应阈值的确定,利用自适应阈值的Sobel和Roberts算子检测出图像边缘后进行融合,再进行形态学的细化。实验结果表明,该算法检测出的边缘清晰,适应性好,对噪声抑制能力强。The traditional Roberts and Sobel edge detection operators have fixed threshold,poor adaptability and ability to resist noise. After the local threshold is calculated by OTSU and Two-peak method, the edges detected by Sobel operator and Roberts operator are fused and then the fused image edge is optimized by mathematical morphology. The results demonstrate that the proposed method has better performance ,which the image edge is clearer and the background noise is filtered as well.

关 键 词：最大类间方差法 双峰法 数据融合 形态学 边缘检测 

分 类 号：TP317.4[自动化与计算机技术—计算机软件与理论]