一种基于数学形态学的齿轮边缘检测方法研究

Gear edge detection method based on mathematical morphology

针对传统的边缘检测算法在提取齿轮边缘过程中存在的抗噪性较差的问题,对图像边缘检测中算法的抗噪性、边缘检测连续性和细节保留能力进行了研究,在对数学形态学的边缘检测算法进行归纳的基础上,提出了一种基于数学形态学的齿轮边缘检测方法。针对采集到的图片,利用最基本形态学操作组合,对齿轮锯齿边缘和内部圆环部分,分别使用了形态学填充和形态学连通域去噪,然后使用形态学边缘检测算子提取了边缘,最后将两部分边缘边缘相加,得到了最终齿轮边缘检测图;再对采集的齿轮图片加入了椒盐噪声,重复进行了边缘检测试验。研究结果表明:该算法能够检测出较好的齿轮边缘,在抗噪性、边缘检测连续性、细节信息保留等方面都有较大的提高。

Aiming at the problem that the traditional edge detection algorithm has poor anti-noise ability in the process of extracting gear edge, the anti-noise ability, continuity of edge detection and detail preservation ability of the algorithm in image edge detection were studied. The edge detection algorithm of mathematical morphology was summarized. A method of gear edge detection based on mathematical morphology was proposed, and the images collected were analyzed. The combination of basic morphological operations were used, the sawtooth edge and inner ring of gear were separated and applied to morphological filling and morphological connected domain denoising. Then the edge was extracted by morphological edge detection operator. Finally, the two edges were added together to get the final gear edge detection map. Then salt and pepper was added noise to the collected gear image, and the edge detection test was repeated. The results indicate that the algorithm can detect better gear edges, and has a great improvement in noise resistance, continuity of edge detection and details information preservation.