| Summary: | 碩士 === 國立臺灣科技大學 === 機械工程系 === 104 === This thesis proposes an object-on-conveyer tracking system, integrating industrial robot operation and 2D image processing techniques, to perform moving object tracking and line-following functions. The proposed system uses 2D image features to localize the position of the object on conveyer, and calculate the 3D assembly positions by the use of object CAD model information. The 3D assembly positions are then used to calculate robot joint angle sets through inverse kinematics and the robot can start to perform the 3D assembly tasks on the moving object. For the line or contour following function, the image processing technique is used to detect the 2D object on conveyer and thus obtain the contour point positions. The encoder information of the conveyer will be fed to the robot so as to perform the contour following action. If a 3D object is to be followed, the image processing technique will localize the object first. Then, through the CAD model, all positions of the contour will be calculated. These positions are finally transformed to contour positions of robot following by using again the conveyer encoder information. This proposed system requires only 2D image processing techniques so that it is flexible, low cost and with a high commercial value.
|
|---|
