【摘要】 提出一种新颖的图像边缘检测算法,包括边缘检测和边缘增强两个阶段.在边缘检测阶段,新的检测算子不仅可以克服传统算子对边缘拐点、终点的漏检现象,还可以有效地去除噪声,从而更加精确地定位边缘.在边缘增强阶段,引入Hopfield神经网络,通过迭代计算网络优化的能量函数,逐步地弥补缺失边缘、消除假边缘,达到边缘增强的目的.最后针对不同类型图片进行边缘检测,得到较好的结果,证明了该算法的可行性.更多还原

【Abstract】 A novel algorithm is presented for image edge detection and enhancement.During the edge detection phase,a new edge detection operator is employed to label exactly the edge because it can get rid of the conventional operators’ miss-detection including the inflection and end points of edges and eliminate efficiently noise.During the edge enhancement phase,the Hopfield neural network is introduced to make up for missed edges and clear out false edges step by step via computing the energy function iteratively for network optimization.In this way the edge detection was done for the images of different types and the results are proved preferable to conventional detection procedure,thus verifying the feasibility of the algorithm.更多还原