论文部分内容阅读
在图像处理与模式识别领域经常根据物体的边界来识别物体。对物体边界特征描述的三种主要方法,即基于极半径函数的傅里叶描述子、边界序列矩和改进的不变矩进行了比较研究,给出了实验结果和结论。
In the field of image processing and pattern recognition, objects are often identified by their boundaries. The three main methods of object boundary feature description, that is, the Fourier descriptor based on the polar radius function, the boundary sequence moment and the improved invariant moments are compared and the experimental results and conclusions are given.