基于计算机视觉测量技术的图像轮廓提取方法研究

A Study on Contour Extraction Method in Computer Vision Measurement Technology

图像分割与轮廓提取是计算机视觉测量技术的关键环节。针对传统边缘检测方法中存在的问题,结合计算机视觉测量技术的特点,提出了一种实用的轮廓提取方法。该方法采用灰度阈值法进行图像分割,并用数学形态学方法对二值图像进行缺陷修补,通过链码跟踪存储轮廓信息,实现了具有单像素边缘的图像轮廓提取。文中给出了关键技术的原理及实现方法。实验表明,与经典的边缘检测方法相比,此方法具有抗干扰性强、精度高等特点,能满足工程测量的实际需要。

Image segmentation and border extracting technology play an important role in the computer vision measurement system. Aiming at the shortcomings of traditional edge detection methods, considering the features of computer vision measurement, a practical contour extraction method is introduced. In the method, image segmentation is based on the gray threshold method, the mathematical morphology method is adopted to remedy the defects of binary image, the contour of image is stored into chain - code through contour tracking algorithm. Using this method, the one - pixel - wide border of image can be easily extracted. The principles and algorithms of key technologies of the method are described. The experiments show that the features of the method such as denoise and precision are better than that of the traditional edge detection methods. It can be applied to practical engineering measurement system.