鉴于传统 Canny边缘检测算法在高斯滤波方差和高低阈值选取上需要人工干预, 不具备自适应能力, 以及其在梯度计算上的缺陷。提出了一种改进的 Canny边缘检测算法。改进算法使用自适应平滑滤波代替高斯滤波, 在平滑图像的同时锐化了边缘；使用水平、垂直、45.和 135.四个方向梯度模板计算图像梯度, 改善了传统 Canny算法在计算梯度时对噪声的敏感性；引进 Otsu算法自适应地根据图像灰度生成高低阈值, 避免了人为设定高低阈值的难题。实验结果表明, 改进算法在检测到更多边缘细节的同时, 也具备较强的自适应性。特别地, 在噪声环境中, 改进算法比传统 Canny算法检测效果更优。

Because the variance of gaussian filter and the high and low thresholds should be determined artificially, Canny algorithm has no adaptive capacity. What is more, it has defect in calculation of gradient amplitude. An improved edge detection algorithm was put forward based on Canny. Adaptive smooth filter was used to smooth image instead of gaussian filter, which could overcome noise influence and sharpen image edge effectively. The improved algorithm also used 4 gradient templates in x-axis direction, y-axis direction, 45° direction and 135° direction to calculate gradient amplitude. Finally, Otsu algorithm was used to get high and low thresholds adaptively based on the gray. The experimental results show that the improved algorithm can detect more edge details with strong adaptability. Particularly, it has much better effect of edge detection than traditional Canny algorithm in the noise environment.