基于双目视觉的齿轮倒角检测方法的研究

发布时间：2018-12-10 21:34

【摘要】：双目立体视觉测量系统是通过模拟人类视觉系统功能来获取场景中被测物体的深度信息,恢复场景中被测物体的三维信息。双目立体摄像机通过采集齿轮倒角的图像信息,来实现非接触式测量。非接触式测量相对于机械接触式测量有着高效率,对被测物体无损伤等优点。因此,基于双目视觉齿轮倒角测量方法的研究具有重要的理论意义和实用价值。本文以双目立体视觉的原理为基础,通过利用机械装置移动单摄像机实现对齿轮倒角三维测量方法进行研究。论文的主要研究内容包括以下几个方面:(1)利用两个不同的摄像机,建立摄像机成像模型,并考虑成像过程中相机畸变的影响。采用张正友标定方法对两个摄像机进行标定,得到两个摄像机的内部参数和两个摄像机之间的外部参数。利用两个摄像机参数的标定结果进行极线校正,将双目立体视觉模型变为理想的平行双目视觉。(2)基于移动单目相机的双目立体视觉测量方法是利用移动平台承载一个单摄像机实现了两个摄像机双目立体视觉的测量功能。该测量方法通过用机械装置代替软件计算,实现了平行双目立体视觉的测量。该测量方法的实验包括实验装置的稳定性,摄像机的移动中心距实验和测量实验等。通过实验数据分析,验证了该测量方法的可行性。(3)对基于特征点的立体匹配模型,本文在研究出多种边缘检测算法的基础上,结合实际情况提出了相对应的边缘提取算法,拟合边缘,求解出特征点。通过特征点匹配得到了被测物体的深度信息以及特征点的空间坐标,进而达到测量的要求。
[Abstract]:Binocular stereo vision measurement system is to simulate the functions of human visual system to obtain the depth information of the measured object in the scene and restore the three-dimensional information of the measured object in the scene. Binocular stereo camera can realize non-contact measurement by collecting image information of gear chamfering angle. Compared with mechanical contact measurement, non-contact measurement has the advantages of high efficiency and no damage to the object under test. Therefore, the research of bevel angle measurement based on binocular vision gear has important theoretical significance and practical value. Based on the principle of binocular stereo vision, this paper studies the three-dimensional measurement method of gear chamfering by moving a single camera with a mechanical device. The main contents of this paper are as follows: (1) using two different cameras, the camera imaging model is established, and the influence of camera distortion in the imaging process is considered. The internal parameters of two cameras and the external parameters between two cameras are obtained by using the calibration method of Zhang Zhengyou. The results of calibration of two camera parameters are used for polar line correction. The binocular stereo vision model is transformed into the ideal parallel binocular vision. (2) the binocular stereo vision measurement method based on the mobile monocular camera is to carry a single camera with a mobile platform to realize the binocular stereo of two cameras. Visual measurement function The measurement of parallel binocular stereo vision is realized by using mechanical device instead of software calculation. The experiments of this method include the stability of the experimental device, the moving center distance of the camera and the measuring experiment. Through the analysis of experimental data, the feasibility of the measurement method is verified. (3) for the stereo matching model based on feature points, this paper studies several edge detection algorithms. According to the actual situation, the corresponding edge extraction algorithm is proposed to fit the edges and solve the feature points. The depth information of the measured object and the spatial coordinates of the feature point are obtained by matching the feature points, and then the measurement requirements are achieved.
【学位授予单位】：天津科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TG86

【参考文献】