Algoritma-Algoritma Pembetulan Warna Baharu Untuk Imej Digital Berwarna

Noordin, Mohd Naim Mohd Jain (2016) Algoritma-Algoritma Pembetulan Warna Baharu Untuk Imej Digital Berwarna. PhD thesis, Universiti Sains Malaysia.

[img]
Preview
PDF - Submitted Version
Download (121kB) | Preview

Abstract

Algoritma pembetulan warna imej digital penting untuk mengurangkan kesan warna sumber pencahayaan yang tidak sekata dalam sesuatu imej digital berwarna. Algoritma pembetulan warna terbahagi kepada dua proses iaitu proses penganggaran warna sumber pencahayaan sekeliling dan proses pembetulan. Walaupun terdapat pelbagai algoritma pembetulan warna yang diperkenalkan, namun imej keluaran yang dihasilkan masih mempunyai kesan fenomena penepuan dan terlebih pembetulan. Bagi mengatasi kekurangan algoritma konvensional ini, tiga algoritma pembetulan warna baharu diperkenalkan iaitu algoritma Pembetulan Warna Anjakan Taburan Piksel (PWATP), Pembetulan Warna Penyekatan Penepuan (PWPP), dan Pembetulan Warna Melalui Penyesuaian Imej (PWPI). Algoritma yang diperkenalkan mengaplikasikan pembetulan warna berdasarkan kepada proses penganjakan piksel pada satah 2D dua saluran warna. Daripada keputusan analisa kualitatif yang dijalankan, didapati algoritma PWATP berjaya menghasilkan imej keluaran terbaik bagi imej dalaman. Warna kulit manusia yang dihasilkan kelihatan lebih semula jadi dan menarik. Algoritma PWPP yang mengaplikasikan algoritma penyekatan penepuan didapati sesuai digunakan bagi memperbetulkan imej yang mempunyai kawasan terlebih pendedahan. Algoritma PWPI pula didapati paling sesuai digunakan bagi imej bawah laut dengan kesan fenomena terlebih pembetulan yang minimum berbanding algoritma pembetulan warna yang lain. Hasil keputusan analisa kuantitatif menunjukkan ketiga-tiga algoritma pembetulan warna yang dihasilkan bukan sahaja dapat memperbetulkan warna imej yang digunakan, malah mengekalkan kecerahan imej dan meningkatkan maklumat yang terkandung pada imej, meningkatkan indeks semula jadi, dan meningkatkan beza jelas imej. Satu analisa kuantitatif baharu yang dinamakan Analisa Penentuan Penepuan Imej (APPI) juga diperkenalkan di dalam penyelidikan ini. Analisa APPI berjaya mengesan piksel-piksel yang mengalami fenomena penepuan dan terlebih pembetulan dengan lebih tepat dan adil. Keputusan yang diperolehi juga dapat dihubungkaitkan dan selari dengan analisa yang dijalankan secara kualitatif. ________________________________________________________________________________________________________________________ Color correction algorithm for digital color image is important to reduce the impact of inhomogeneous surrounding illumination in digital color image. Color correction algorithm is divided into two main processes, namely the surrounding illumination color estimation process and color correction process. Although there are various color correction algorithms introduced, the resultant images produced are still prone to the effect of saturated and over-corrected phenomenon. To overcome the drawback of these conventional algorithms, three new color correction algorithms were introduced namely Pixel Distribution Color Correction (PDCC), Saturation Avoidance Color Correction (SACC) and Adaptive Color Correction (ACC). The proposed algorithms apply the color correction processes by shifting the pixel distribution based on the 2D two color channel plane. From the qualitative analysis results, the PDCC algorithm was able to produce the best resultant image for indoor images. The produced human skin color looks more natural and more pleasantly in viewing. The SACC algorithm which employs the saturation limitation algorithm was found to be suitable to correct the over-exposed area on the image. The ACC algorithm on the other hand is the most suitable algorithm to be used to correct the color of underwater images with effect of minimal saturation phenomenon as compared to other conventional color correction algorithm. The results of quantitative analysis showed that all the proposed three color correction algorithms not only have the ability to correct the colors of the images, but also are able to preserve the brightness of the image, enhance the image details, increase naturalness index, and increase the contrast of the image. A new quantitative analysis namely Saturation Pixel Detection Analysis (SPDA) is also introduced in this research. The SPDA analysis successfully detects the saturated and over-corrected pixel more accurately and unbiased. The result produced by SPDA can be correlated and coherent with the qualitative analysis carried out.

Item Type: Thesis (PhD)
Additional Information: Full text is available at http://irplus.eng.usm.my:8080/ir_plus/institutionalPublicationPublicView.action?institutionalItemId=3131
Subjects: T Technology
T Technology > TK Electrical Engineering. Electronics. Nuclear Engineering > TK7800-8360 Electronics
Divisions: Kampus Kejuruteraan (Engineering Campus) > Pusat Pengajian Kejuruteraaan Elektrik & Elektronik (School of Electrical & Electronic Engineering) > Thesis
Depositing User: Mr Mohd Jasnizam Mohd Salleh
Date Deposited: 03 Sep 2018 08:19
Last Modified: 03 Sep 2018 08:19
URI: http://eprints.usm.my/id/eprint/41656

Actions (login required)

View Item View Item
Share