Kinematic analysis of a slider crank mechanism via a pre–calibrated vision system developed by using two commercial cameras


Thesis Type: Postgraduate

Institution Of The Thesis: Middle East Technical University, Faculty of Engineering, Department of Mechanical Engineering, Turkey

Approval Date: 2014

Thesis Language: English

Student: Mehmet Hilmi Eralp

Supervisor: REŞİT SOYLU

Abstract:

There are two main objectives of this study. The first objective is to develop a vision system consisting of 2 inexpensive commercial cameras. In general, by self – calibration methods reconstruction of a scene by using uncalibrated images is performed up to a scale only. However, in this thesis reconstruction of a scene is to be performed such that one obtains the actual values of the distances in the scene. For this purpose, it is assumed that the extrinsic parameters of the cameras are known. Therefore, one needs to determine the intrinsic parameters of the cameras only. In order to calculate the intrinsic parameters, two methods, that take advantage of the simplified Kruppa equations and the equal eigenvalue theorem, are used. The results obtained via the two methods are compared with the results obtained by using a calibration pattern. A triangulation process is then performed to calculate several known distances in the scene by using the method that gives better results for the intrinsic parameters. The actual and estimated distances obtained via the vision system are then presented and compared. The second objective of this study is to perform kinematic analysis of a slider crank mechanism by using the developed vision system. The position, velocity and acceleration analyses of the slider crank mechanism are realized by using several markers that are attached on the moving links of the mechanism. The positions of the markers are calculated by using the vision system. This data is then utilized to determine the joint variables, joint velocities and joint accelerations of the slider crank. The results thus obtained via an encoder attached to the input link of the mechanism are compared with the results obtained via the developed vision system. The effects of the locations of the markers and the effects of the number of markers used on the accuracy of the results are also investigated.