一种快速的图像边缘精确提取算法

图像边缘检测是图像处理与分析中最基础的内容之一。现有的边缘检测算法存在检测精度低、抗噪性能差、处理速度慢等缺点。针对这些问题,提出了一种快速图像边缘检测算法。通过基于图像边缘信息的阈值分割,能快速地提取出图像的边缘轮廓,通过对其进行中心细化,可达到对物体边缘精确定位的目的。仿真实验证明,该算法在滤除图像噪声、保留细节边缘、细化边缘宽度和保持边缘连通性方面都获得了不错的效果,并且运算量小,既适用于对视频图像的实时处理,也适用于对图像中的隐蔽目标特征的提取和分析。

A method for quickly and exactly extracting image detail edge

The image edge extraction is one of the basic contents in image manipulation and analysis.However,at present,there are some problems existed such as low extraction precision,weak resistance of noises,low process speed and so on.To resolve these problems,a new method of image detail edge extraction is proposed.In this method,the image to be manipulated was firstly segmented based on the information of edge characters with great speed,which could accentuate the edge areas and remove part background and noises,then a thinner arithmetic is adopted to thin the edge areas from both sides so that the central line of edge could be exactly located.Simulations show that this method does well in resisting image noises,keeping detail edges,thinning edge width and keeping the connectivity of edges.In addition,this method needs no complex operation so that it can process the images with great speed,therefore,it is not only effective on character extractions for inconspicuous target in the images,but also fit for real-time image manipulation.