A study of lighting techniques and an automatic vision system

• Main Author: Prahladan Ganesan.
• Other Authors: Zhong Zhaowei
• Format: Final Year Project
• Language: English
• Published: 2010
• Subjects: DRNTU::Engineering::Computer science and engineering::Computing methodologies::Image processing and computer vision
• Online Access: http://hdl.handle.net/10356/20794
• Institution: Nanyang Technological University

More and more manufacturing companies are using machine vision technology to improve their productivity and reduce costs. At the core of this growth in machine vision are manufacturer’s ever-increasing demands over the quality of manufactured parts. Therefore in a machine vision system, the image processing and illumination techniques is the key to successful inspection. The quality of the image depends on various important factors such as illumination techniques, camera, lens, defect characteristics. This report discusses about some of these factors using the actual data from the industry. This report also introduces the equipment being used for this project. At first, the manual inspection system is being introduced. From there, the introduction of Automated Vision system is illustrated. Then, it goes into the details of the methodology applied in the system. It also explains the setup of this system. In addition, the images used for this project are all captured from this machine. Two different lighting sources such as lamp and white LED are chosen to produce golden images. These images are analyzed in terms of clarity and quality of the image. It is observed that using white LED gives clearer and sharp images. The actual images are also represented in this report with the two different types of light sources. Next two different type of camera is being used, with low & high resolution camera and the difference in terms of cycle time and productivity is observed. Finally, various images captured during inspection are also analyzed which reflects the capability and room for improvement of this system.