机械零件边缘检测方法的研究

Study on Edge Detection Method of Mechanical Parts

为了实现机械零件的有效检测,针对零件具有比较明显的轮廓特点,提出了一种机械零件图像的边缘检测算法.首先对零件图像进行去噪和灰度化处理,利用图像灰度直方图分布信息,构造求解最佳分割阈值的数学模型,在此基础上设计了边缘检测的最佳阈值分割算法.最后通过实例给出了用经典算法Sobel算子分割的效果图和采用本算法取得的效果图,通过对比充分说明了本算法的可行性和有效性。

For effective detection of mechanical parts and aiming at the obvious outline feature of mechanical parts, a novel image edge detection algorithm for mechanical parts is proposed. Firstly the denoising and graying of the image is done. Then with the image gray-scale histogram distribution information, the mathematical model with optical segmentation threshold is constructed. On this basis, the edge detection optimal threshold segmentation algorithm is proposed and designed. Finally, the comparison of the effect diagrams by classic Sobel operator segmentation algorithm and the proposed novel algorithm clearly indicates the feasibility and effectiveness of the proposed novel algorithm.