相对边界矩在模式识别中的应用

The Application of Relative Edge Invariant Moment on Pattern Recogniton

针对采用图像二维不变矩识别目标形状时,计算量大,耗费时间长的缺点,本文提出了相对边界矩算法,这种算法是利用目标物体的边界进行识别的,所以不仅计算速度快而且在离散状态下也具有平移、旋转和比例因子不变性,实验证明,此方法是一种快速的行之有效的方法。

To improve the speed of target recognition with two-dimension invariant moment of image, a relative edge invariant moment is put forward in this paper. Because this algorithm takes advantage of target edge, it is not only fast but also invariant as shift,scaling and rotation. Finally through tests, the results show this algorithm is fast and feasible.