摘要
基于机器视觉技术的产品自动检测、分类和测量系统,由于非接触性和快速、精度高等优点在实际生产中得到广泛关注。论文研究并建立了目标体几何形状和尺寸检测机器视觉系统。原始图像经过预处理、获取图像轮廓特征后,采用一个基于不变矩、相对矩和提取角点特征的新方法自动识别目标体的形状。根据工业应用实际情况,线性模型标定获取摄像机内外参数,计算目标体形状特征的真实尺寸。实验证明该方法计算复杂度相对于常用的Hough变换大大降低,适合实时处理,实验结果较为理想。
Machine vision systems for automated inspection, sorting and parameters measurement have been focused in the manufacturing processes owing to the advantages of non-contact, high-speed and high-precision. A geometric shape recognition and geometry parameters measurement system is developed based on machine visionr. A new method of geometric shape recognition is suggested based on the integrated technologies of invariant moments, relative moments and corner detection. For geometric parameters measurement, a linear model calibration is realized to get the inner and outer parameters of camera for the actual geometric parameters. It is obtained from the experiments that the results are satisfied and the complexity of the method is greatly less than the common method which is called Hough Transform.
出处
《微计算机信息》
2009年第16期222-223,280,共3页
Control & Automation
基金
广东省科技计划项目
项目名称:医护用健康监护分析仪的研究与开发(2007B010400049)
关键词
机器视觉
不变矩和相对矩
角点检测
摄像机标定
Machine vision
Invariant moments and relative moments
Corner detection
camera calibration
