基于DM642的Sobel算法改进研究

Improved research of Sobel algorithm based on DM642

边缘是图像的重要特征,是机器视觉的基础。在经典三阶Sobel算法模板的基础上,增加了除X轴向以及Y轴向算法模板之外的6个模板,使得三阶Sobel算法模板达到8个,经过实验,证实可以有效的提高边缘检出精度。更进一步的,创造性的将经典三阶Sobel算法模板提升到五阶Sobel算子模板,并在DM642平台上验证其算法的边缘提取精度较三阶算法模板有较大提升。

The edge is an important features of image, and also the foundation of computer vision. In this paper, it add another six templates besides the x-axis and y-axis ones,on the basis of the classic three-order Sobel algorithm, which increase the number of three-order Sobel algorithm from two to eight ,and also proved by experiment that it improved the precision of edge detection effectively. Further more, it change the sobel algorithm template from three orders into five orders in a creative way, and proved on the DM642 platforms that the edge extracting precision of its algorithm has a larger upgrade than the three order algorithm.