In order to meet the aviation’s and machinery manufacturing’s pose measurement need of high precision, fast speed and wide measurement range, and to resolve the contradiction between measurement range and resolution of vision sensor, this paper proposes an orthogonally splitting imaging pose measurement method. This paper designs and realizes an orthogonally splitting imaging vision sensor and establishes a pose measurement system. The vision sensor consists of one imaging lens, a beam splitter prism, cylindrical lenses and dual linear CCD. Dual linear CCD respectively acquire one dimensional image coordinate data of the target point, and two data can restore the two dimensional image coordinates of the target point. According to the characteristics of imaging system, this paper establishes the nonlinear distortion model to correct distortion. Based on cross ratio invariability, polynomial equation is established and solved by the least square fitting method. After completing distortion correction, this paper establishes the measurement mathematical model of vision sensor, and determines intrinsic parameters to calibrate. An array of feature points for calibration is built by placing a planar target in any different positions for a few times. An terative optimization method is presented to solve the parameters of model. The experimental results show that the field angle is 52 °, the focus distance is 27.40 mm, image resolution is 5185×5117 pixels, displacement measurement error is less than 0.1mm, and rotation angle measurement error is less than 0.15°. The method of orthogonally splitting imaging pose measurement can satisfy the pose measurement requirement of high precision, fast speed and wide measurement range.
The online measurement of the metal surfaces’ parameters plays an important role in many industrial fields. Because the surfaces of the machined metal pieces have the characteristics of strong reflection and high possibilities of scattered disturbing irradiation points, this paper designs an online measurement system based on the measurement principles of linear structured light to detect whether the parameters of the machined metal surfaces’ height difference and inclination fulfill the compliance requirements, in which the grayscale gravity algorithm is applied to extract the sub-pixel coordinates of the center of laser, the least squares method is employed to fit the data and the Pauta criterion is utilized to remove the spurious points. The repeat accuracy of this system has been tested. The experimental results prove that the precision of inclination is 0.046° RMS under the speed of 40mm/sec, and the precision of height difference is 0.072mm RMS, which meets the design expectations. Hence, this system can be applied to online industrial detection of high speed and high precision.