【摘要】 通过对比传统边缘检测算法、小波变换多尺度边缘检测算法和最优滤波二阶差分算法,本文认为:传统边缘检测算法没有考虑多尺度问题,但由于算法以样点为单位,所以在检测小级别边缘特征方面仍有优势,且可较容易获得边缘方向的描述;应用小波变换可以实现多尺度边缘检测;通过最优滤波二阶差分算法也可以实现多尺度边缘检测。以上各种方法可作为相干体技术的重要补充。更多还原

【Abstract】 By comparing traditional edge detection algorithm,the wavelet transform multi-scale edge detection algorithm and the optimal filtering second-order difference algorithm,it is believed in this paper that the traditional edge detection algorithm does not give the consideration for the multi-scale problem,however,because the algorithm takes the sample point as the unit,it still has the advantage in detecting small edge characteristics and can be used to easily obtain the description for edge direction.Furthermore,the multi-scale edge detection can also be implemented by applying either optimal filtering second-order difference algorithm or the wavelet transform.Each algorithm mentioned above can be regarded as the important supplement for the coherent volume technique.更多还原