Mathematical model for underwater color constancy based on polynomial equation

Research output: Contribution to journalArticlepeer-review

4 Citations (Scopus)

Abstract

Nowadays, the underwater researches become one of the distinguish researches since the ocean are considered as the lungs of this world. Therefore, underwater image processing researches should be encouraged especially in the differences of color analysis between underwater and on the air objects. The colors of objects are changed when the objects are put underwater. This is due to discoloration from light scattering, refractive index, the wave, the distance of camera and environmental underwater effect. Discoloration in water is highly dependent on the intensity level and wavelength of each color image. In this paper propose polynomial equation approach to enhancement underwater color constancy. Polynomial approach is conducted through two steps: first, determining the relation between the color intensity of an image on the water surface and the color intensity of an image in a certain depth, second, determining the coefficient of constant function relation between the color intensity of an image on the water surface and the color intensity of an image in a certain depth by using least square. The result of polynomial approach is measured by using Peak Signal to Noise Ratio, yielding an average value of 19.64 and visually the result of the image color approximates its original color. It can be concluded that polynomial approach can determine the color constancy level which in turns can enhance the underwater image just as its original color.

Original languageEnglish
Pages (from-to)159-166
Number of pages8
JournalJournal of Theoretical and Applied Information Technology
Volume87
Issue number1
Publication statusPublished - 10 May 2016

Keywords

  • Mathematical model
  • Peak signal to noise ratio
  • Polynomial equation
  • Underwater color constancy

Fingerprint

Dive into the research topics of 'Mathematical model for underwater color constancy based on polynomial equation'. Together they form a unique fingerprint.

Cite this