针对传统边缘检测算子具有对噪声敏感、提取出的边缘细节特征没有得到很好的保持等不足,文中在Canny算法启发下提出了一种基于多结构元的数学形态学边缘检测算法,并经过数学形态学细化算法提取图像边缘,并与Sobel边缘检测算法和Canny边缘检测算法得到的图像边缘仿真结果进行对比,分析比较其优点不足,得出结论:文中算法得到的图像边缘其定位准确且较平滑,轮廓清晰,很好地保留了边缘细节特征,比较容易并行实现,而且由于此种边缘检测算法抗噪能力较强,能较好解决性噪比和单边缘响应之间的矛盾.%According to the traditional edge detection operators to noise sensitive, the extracted features edge details are not well maintained, it presents a multi-structuring elements based on morphological edge detection algorithm inspired by the Canny algorithm , and the edge got through morphological thinning algorithm, and compare the simulation results of the Sobel edge detection algorithm and Canny edge detection algorithm, analyse the advantages and disadvantages to a drawn conclusion; the edge of this paper's algorithm extracted has a more accurate orientation and relatively smoother contour, more clear outline, easy to parallel implementation, well maintained features edge detail, and because this kind of edge detection algorithm has a strong antinoise ability, it can well solved the conflict between performance/noise ratio and single edge response.
展开▼
