景物边缘提取的边缘分段自增强算法

An Edge Extraction Algorithm Using the Accumulation of Sub-edge Self-reinforcement of Search Trajectories

景物边缘信息是进行图象分析和识别的重要属性,如何有效地从噪声图象中提取边缘是这些领域中的难点。该文提出了一种边缘分段自增强算法用于噪声图象的边缘提取。该算法首先对噪声图象进行小尺度高斯滤波,并使用该文设计的新型边缘检测算子获取引导信息,此边缘检测算子在定位精度、抑制噪声和虚假边缘方面具有很好的性能;然后对各搜索轨迹进行分段自增强,最后根据自增强累积的程度获取噪声图象中的边缘。实验结果表明:此算法能够有效地从噪声图象中提取物体的真实边缘,并能最大限度地保留细节信息。

Edge is an important attribute for image analysis and recognition,but it is difficult to effectively extract edge from noisy image in the domain.In this paper,a novel sub-edge self-reinforcement algorithm for edge extraction in noisy image is proposed.Firstly,the noisy image is filtered by a small scale Gaussian Filter,then a new Large Template Edge Detector is designed in order to get more accurate lead information for the subsequential sub-edge self-reinforces;Finally,the real edge of noisy image is extracted from the accumulated search trajectories.Experimental results show that the method can extract real edge in noisy image while keeping the image details.