Please use this identifier to cite or link to this item:
metadata.theses.dc.title: Enhanced Non Linear Image filtering technique for noise removal in digital image
metadata.theses.dc.contributor.*: Muhammad Syafiq Alza
metadata.theses.dc.subject: Image filtering technique
Multi-type Noise
Digital Image 12-Mar-2019
metadata.theses.dc.description.abstract: One of the main challenges in image processing is to remove noise in digital image. Noise removal is vital because it caused problem such as degradation of the image quality and loss crucial information in image. Most common noise in vision system image is salt and pepper noise which cause by illumination and sensor temperature. Therefore, image filtering technique has been introduce to overcome these problem. Nonlinear image filtering technique has proven to have better efficiency than the linear technique which tend to blur the image. Median filter technique is one of the well-known nonlinear technique which has been widely used to remove salt and pepper noise. However, this technique only work well with low density salt and pepper noise. Therefore this research is aimed to improve image filtering technique based on median filter technique and decision based on algorithm for salt and pepper noise removal. The improvement of image filtering technique to remove multi-type noise is also addressed by using an enhanced image filtering technique that is develop based on hybrid median filter. The proposed filter are tested on three different images of lead frame, fruit, and Lena images. Their performance are then compared to median filter, Hybrid median filter, Hybrid sigma filter, Adaptive median filter and ben filter technique based on mean square error, structural similarity index and computational time taken. The result show that improved decision based algorithm can efficiently remove salt and pepper noise while enhanced median filter has the lowest computational time. Finding for multi-type noise also indicate that enhanced hybrid median filter is able to provide a good result in structural similarity index analysis.
metadata.theses.dc.theses.semester: 2018
metadata.theses.dc.theses.course: Master in Engineering Technology (Industrial Automation)
Appears in Collections:Master Theses

Files in This Item:
File Description SizeFormat 
Muhd Syafiq_UBis.pdf7.29 MBAdobe PDFView/Open    Request a copy

Items in UniKL IR are protected by copyright, with all rights reserved, unless otherwise indicated.