(For USM Staff/Student Only)

EngLib USM > Ω School of Aerospace Engineering >

Development of visual defect detection for aerospace composite pane

Development of visual defect detection for aerospace composite pane / Muhammad Firdaus Mohd Sharif
Pemeriksaan visual merupakan salah satu ujian bukan musnah yang memerlukan pemeriksaan dengan menggunakan penglihatan manusia atau penglihatan mesin. Kaedah ini kebiasaannya digunakan dalam industri terutamanya industri Aeroangkasa untuk mengesan kecacatan pada panel. Kecacatan yang biasa berlaku dalam panel komposit pesawat adalah calar, kemek, lubang jarum dan lubang yang terdedah ini adalah jenis kecacatan yang biasa berlaku pada bahan komposit dan kebiasaannya ia berlaku semasa proses fabrikasi. Dalam kajian ini, satu algoritma pengesanan kecacatan dengan kaedah pemprosesan imej digunakan telah dihasilkan untuk penyelidikan dan tujuan pendidikan. Algoritma ini boleh mengesan hingga 4 jenis kecacatan seperti calar permukaan, permukaan kemek, lubang jarum dan lubang penggerudian. Kajian ini dibahagikan kepada dua bahagian utama iaitu sistem perolehan imej dan sistem pemprosesan imej. Dalam sistem perolehan imej, imej sampel ujian telah ditangkap dengan menggunakan tiga kaedah yang berbeza yang Sistem Proses Kualiti Rendah, Sistem Perolehan Mengimbas Mesin Pengimbas Rata, dan Sistem Unjuran Imej Sebenar. Selepas itu, sistem pemprosesan imej akan memproses imej yang diperolehi dari tiga kaedah sistem perolehan imej di mana satu kaedah penyingkiran ditolak telah digunakan dalam kajian ini. Kecacatan diklasifikasikan kepada beberapa jenis iaitu simulasi kecacatan, kecacatan sebenar oleh Sistem Perolehan Mengimbas Mesin Pengimbas Rata dan kecacatan sebenar oleh Sistem Unjuran Imej Sebenar. Pengesanan kecacatan telah dilakukan oleh sistem pemprosesan imej yang mana kecacatan ke atas sampel ujian telah dikesan dan peratusan kecacatan telah dikenal pasti. Hasil penting imej yang telah diproses daripada semua kaedah telah dibincangkan. Kajian pengesanan kecacatan visual pada panel komposit aeroangkasa telah dilakukan. _______________________________________________________________________________________________________ Visual inspection is one of the Non-Destructive Testing which requires inspection by using human vision or machine vision. This method normally used in industries to detect the defect, especially in the Aerospace industry. The most common defects occur in Aerospace composite panel are scratched, dented, pinhole and nicked which these kind of defects commonly occurs on composite material and it commonly was happening during the fabrication process. In this research, an algorithm of defect detection by using an image processing method was developed for research and educational purpose. This algorithm can be detect up to 4 types of defect which are surface scratch, surface dented, pinhole and drilling holes. This research can be divided into two main parts which are image acquisition system and image processing system. In image acquisition system, the images of test samples were captured by using three different methods which are Low Processing Image System, Flatbed Acquisition Scanning System, and Projection Real Image System. After that, image processing system will process the acquired image from the three methods of image acquisition system where an elimination-subtracted method was used in this research. The defects were classified into several types which are simulated defect, actual defect by Flatbed Acquisition Scanning System and actual defect by Projection Real Image System. The defect detection was done by the image processing system which the defect on the test samples were detected and the percentage of defects was determined. The important result of processed image from all methods has been discussed. The study of visual defect detection on aerospace composite panel was done.
Contributor(s):
Muhammad Firdaus Mohd Sharif - Author
Primary Item Type:
Final Year Project
Identifiers:
Barcode : 00003105258
Accession Number : 875006751
Language:
English
Subject Keywords:
Visual inspection; Non-Destructive Testing; human vision
First presented to the public:
5/1/2016
Original Publication Date:
5/15/2018
Previously Published By:
Universiti Sains Malaysia
Place Of Publication:
School of Aerospace Engineering
Citation:
Extents:
Number of Pages - 117
License Grantor / Date Granted:
  / ( View License )
Date Deposited
2018-05-15 11:41:49.318
Date Last Updated
2020-05-05 14:16:42.548
Submitter:
Mohd Jasnizam Mohd Salleh

All Versions

Thumbnail Name Version Created Date
Development of visual defect detection for aerospace composite pane1 2018-05-15 11:41:49.318