Author: Huo, Shengzeng
Title: A robotic line scan system with adaptive ROI for inspection of defects over convex free-form specular surfaces
Degree: M.Sc.
Year: 2021
Subject: Robotics
Manufacturing processes -- Automation
Hong Kong Polytechnic University -- Dissertations
Department: Department of Mechanical Engineering
Pages: A-C, xii, 122 pages : color illustrations
Language: English
Abstract: With the development of China's manufacturing industry, the need for an automatic defect inspection system in the field of quality control becomes necessary and urgent. In this paper, we present a novel robotic system to perform automatic defect inspection tasks over free-form specular surfaces. The autonomous procedure is achieved by a six-DOF robotic manipulator, equipped with an image acquisition subsystem, consisting of a line scan camera and high-intensity line lighting. The whole pipeline of the system includes scanning path planning, projection registration and image processing. Given the CAD mesh model of the specular workpiece, a point cloud that samples from the model will be used to guide our scanning path planning. The first step is about point cloud preprocessing, filtering out the irrelevant points for the appearance inspection. Next, a K-means based unsupervised classification algorithm is implemented to segment the surface into regions with similar curvature. Then, the scanning path is computed by using an adaptive algorithm that adjusts the position and the orientation of the camera to observe regions with irregular shapes properly, ensuring that all the points of the point cloud are in the view during the process of scanning. Finally, the concept of nearest neighbor search is used to optimize the complete scanning path for the whole object. On one hand, the scanning sequence in terms of different regions is arranged with an iteration search method. On the other hand, the orientation of the line scan camera is also optimized to minimize the rotation adjustment.
To deal with the high reflection issue of the specular surface, a novel iterative closest point-based projection registration method that robustly localizes the object in the robot's coordinate frame system is proposed. Then, the scanning pose in the object's frame can be transformed into the robot's frame. The position and the orientation of the pose is converted to the command for the robot manipulator. A graphical user interface is provided to simplify the operation in the real production line. The mathematical model of the line scan camera is analyzed and computed, in which the scanning range ofthe linear motion at a discrete moment is described as a rectangel, consisting of field of view and depth of view. With a linear motion, the scanning rectangle integrates as a cube. The inspection completeness can be proved if all points on the surface are in the cube. According to the characteristics of the line scan camera, the motion of the robotic manipulator should match with the sampling rate of the line scan camera, ensuring that the captured images are complete without distortion. An image processing pipeline through contour search and approximation is proposed to automatically detect defects in the captured high-resolution images. The defects in the individual 2D images can be mapped to the 3D object space according to the corresponding scanning pose. A detailed experimental study for our proposed vision-guided robotic scanning system is reported to validate the performance. The results show that our proposed system is capable of finishing the automatic defect inspection task with high accuracy and being applied to the real production line.
Rights: All rights reserved
Access: restricted access

Files in This Item:
File Description SizeFormat 
5930.pdfFor All Users (off-campus access for PolyU Staff & Students only)3.98 MBAdobe PDFView/Open

Copyright Undertaking

As a bona fide Library user, I declare that:

  1. I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
  2. I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
  3. I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.

By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.

Show full item record

Please use this identifier to cite or link to this item: