【摘要】 在分析了传统的多尺度边缘融合方法存在的角点漏检和大尺度边缘信息丢失等问题的基础上,提出了一种新的基于相邻尺度间梯度方向的多尺度边缘融合方法.该方法充分利用相邻尺度间对应边缘点梯度方向相近这一特点,在没有增加其它辅助方法的前提下,只是通过改变多尺度边缘融合的依据和条件,即可改善边缘检测的最终效果,在保持算法简单性的同时,相比于传统方法有更好的弱边缘检出率、角点检出率、抗噪能力和更高的算法效率,仿真研究也证明了该融合方法的有效性.更多还原

【Abstract】 The method for multi-scale edge fusion may lose the corners and the weak edges on large scales. To solve the problem,a method for multi-scale edge fusion based on inter-scale gradient direction is proposed,which can improve final result of edge detection by changing the condition of multi-scale edge fusion but adding other assistant methods.The theoretical analysis and the experimental results show that the method has strong anti-noise capability and high efficiency,and it detects more week edges and corners.更多还原