针对传统图像边缘检测速度慢和连续性差的缺点,通过构造图像信息测度特征属性,提出一种基于图像信息测度和ELM的图像边缘检测方法,采用度量F作为图像边缘检测的评价指标.研究结果表明,ELM图像边缘边缘检测效果优于LVQ、BP和Sobel算子,图像边缘更加清晰、纹理性较强、连续性好,并且具有较好地抗噪声性能.%In order to overcome the disadvantages of the traditional image edge detection,such as slow speed and poor continuity,a new image edge detection method based on image information measure and ELM is proposed.F measurement is used as the evaluation index of image edge detection,the results show that the edge of ELM image edge detection effect is better than those of LVQ,BP and Sobel operator,image edge is more clear and strong texture,good continuity,and has better anti noise performance.
展开▼