Led Cosmetic Flaw Vision Inspection System

dc.contributor.authorChu, Jenn Weng
dc.date.accessioned2016-10-17T06:14:46Z
dc.date.available2016-10-17T06:14:46Z
dc.date.issued1997-01
dc.description.abstractCurrently, in the Light Emitting Diode (LED) component industry, one of the challenges faced by most of the manufacturers is how to effectively eliminate cosmetic rejects from the finish products. Since there has yet a solution to totally eliminate the cosmetic rejects from the product by improving their manufacturing processes, various visual inspection check points need to be established to prevent the reject units from reaching their customers. This cosmetic flaw inspection process is currently performed visually by human, however, it is not effective. Hence it was resolved that if the process can be fully automated, it can reduce manufacturing cost drastically, increase productivity and improve out-going product quality. Finally the 6 sigma goal and world class manufacturing can be achieved. A research and development work in automated vision inspection system to resolve the above problem is presented in this thesis. The developed system is capable to detect diffuse LED's defects mimely scratches, bubbles, contamination, fuzzy dome and off centre in less than 200ms using a 200MHz Pentium PC and a Pulnix high speed camera. Various image processing techniques are utilised to accomplished the inspection task. A machine vision approach which comprises pre-processing, image segmentation, clean up and feature extraction operations, was implemented to perform the automated cosmetic flaw inspection. In the laboratory tests conducted, the system was found to be 100% accurate in detecting LED dome defects on LEDs of different colour and intensity. The system can also classify defects into different categories and was found to be 90% accurate.en_US
dc.identifier.urihttp://hdl.handle.net/123456789/2752
dc.subjectEliminate cosmetic rejectsen_US
dc.subjectfrom the finish productsen_US
dc.titleLed Cosmetic Flaw Vision Inspection Systemen_US
dc.typeThesisen_US
Files
License bundle
Now showing 1 - 1 of 1
Loading...
Thumbnail Image
Name:
license.txt
Size:
1.71 KB
Format:
Item-specific license agreed upon to submission
Description: