| Summary: | 碩士 === 南臺科技大學 === 機械工程系 === 104 === This paper applies visual servoing control technology to the path tracking of object contour, and uses PTZ camera and CNC X-Y table as two individual hardware platforms to verify the proposed concept.
The paper introduces the image processing technique of Hough transform to calculate the image coordinates of contour path which is then converted into CNC processing command by the neural network learning scheme of Self-Organizing Maps (SOM). The SOM used in this study is to construct an input-output visual servoing map by collecting a sufficient amount of experimental data. Relevant experimental result demonstrates the center of image frame of PTZ can accurately track along the contour path of the image.
In addition, this study extends the concept of image contour tracking to tackle a related problem raised in the X-Y table machining control. When equipped with a laser pointer that projects a light spot along the object contour in order to simulate the edge cutting, the proposed visual servoing system is capable of converting image coordinates of the contour to the coordinates of the movement of X-Y table, and able to make the laser spot track along the desired contour. Experimental results verify feasibility of the image contour tracking control scheme.
Keywords：Visual servoing control, Hough Transform,Self-Organization Maps
|
|---|
