摘要
边缘检测是图像识别的中心环节,快速、精确地提取图像物体的边缘信息一直是国内外研究的热点。本文介绍了Roberts算子、Prewitt算子、Sobel算子、LOG算子和Canny算子等经典边缘检测算子,对其性能和算法特点进行分析。运用Matlab进行算法的仿真,并对其检测结果进行分析和比较。得出Sobel算子、Prewitt算子检测斜向阶跃边缘效果较好,Roberts算子检测水平和垂直边缘效果较好,LOG算子和canny算子提取的边缘比较完整,位置比较准确,并能够检测出图像较细的边缘部分。为进一步研究图像边缘检测理论奠定了一定的基础。
The edge detection is the center link of image recognition. Fast, accurately to extract image edge information is a research hotspot. This paper introduces the Roberts operator, Prewitt operator, Sobel operator, LOG operator and Canny operator and other classical edge detection operators, and the lelatire performance and the character of algorithm a nalysis. The simulation using the Matlab algorithm, is carries on and the analysis and comparison of the test results are acqmred. The Sobel operator, Prewitt operator to detect oblique step edge effect is good, and Roberts operator is fir for detecting the horizontal and vertical edge effect. LOG operator and Canny operator edge is more complete, the position is much more accurate, and can detect the image edge portion of a fine. This useach Layies a foundation for the further study of image edge detection theory.
出处
《北京测绘》
2014年第3期65-68,146,共5页
Beijing Surveying and Mapping
