摘要
明确了子波变换、Ｍａｒｒ－Ｈｉｄｒｅｔｈ滤波器和神经网络侧抑制的相互关系，分析比较了优化法和自适应平滑法两种图像边缘检测方法的特性，就过零点检测给出了具体方案，为清除虚假过零点，设计了二阶导数过零与三阶导数小于零相结合的算法，取得了较好的边缘检测效果。
The mutual relation between the wavelet transformation , the Marr Hidreth filter and later- al inhibition of the nervous network is made clear. A comparison is made between the specific features of two methods of edge detection , the optimal detector method and the adaptive smoothing one. A concrete approach to determining the zero-crossing is given to eliminate the false edges . An algorithm based on a combination of the zero-crossing of the second derivative and a negative third derivative has been designed. The result of edge detection is found to be satisfactory.
出处
《华中理工大学学报》
CSCD
北大核心
1997年第4期29-32,共4页
Journal of Huazhong University of Science and Technology
