An Effective Tool for Enhancement of Low Quality Underwater Images

Capturing the high-resolution images in the water is a complicated task. Mostly these images suffered from low contrast, low visibility, and blurriness. There are different  factors which cause these problems in the underwater images. These factors include light attenuation, the p

Project Title

An Effective Tool for Enhancement of Low Quality Underwater Images

Project Area of Specialization

Artificial Intelligence

Project Summary

Capturing the high-resolution images in the water is a complicated task. Mostly these images suffered from low contrast, low visibility, and blurriness. There are different 
factors which cause these problems in the underwater images. These factors include light attenuation, the presence of different particles in water and wavelength absorption 
of colors. At present, there are well known underwater image enhancement techniques which include Histogram Equalization, Adaptive Histogram Equalization, Contrast Limited Adaptive Histogram Equalization. But there are some drawbacks in these techniques like over enhancement of the image, high computational cost and increase of the noise. So here comes the need for a new framework for the enhancement of underwater images

An Effective Tool for Enhancement of Low Quality Underwater Images (ETELQUI) is a desktop application that will enhance and preserve the information in the underwater images. We have decided to target only underwater images, but this application can be suitable to other images as well. This product will help the security forces, marine researchers, scuba divers and T.V broadcasters for their intended purposes.  In comparison to some other enhancement techniques, like Histogram Equalization, Adaptive Histogram Equalization and Contrast Limited Adaptive Histogram Equalization our techniques gave better results. ETELQUI will enhance the underwater images in such a way that it will increase the number of visible details of the image.

Project Objectives

The main objective of the product is the enhancement of the low contrast underwater images and preserve the natural appearance of the objects present in the image. The ultimate goal is to reveal the hidden information of the underwater images. 
 

Project Implementation Method

We have studied more than 18 research papers and by these research paper, we carried out multiple techniques and by these multiple techniques we designed a new framework for the enhancement of underwater images.  In this framework, 
Discrete Wavelet Transform (DWT), Gamma correction by using Particle Swarm Optimization (PSO) will apply on the input image. DWT is the algorithm by which one image is divided into four sub-images by applying the low pass and high pass filters on the input image. Low pass filter is denoted by L and high pass filter is denoted by H. These images are named according to the filters which are applied to the image. Name of sub images are LL, HL, LH, and HH. First alphabet of the sub image name, shows the filter which applies on the horizontal direction of the image and the second alphabet shows the filter which applies on the vertical direction of the image. The LL band or image is considered as a reduced version of the original image because it is the most detailed image. For further enhancement, we will use Gamma Correction on the LL image. In the Gamma Correction, there is a gamma correction factor or alpha factor which is used for establishing the more natural appearance of the image. For optimizing the alpha factor of Gamma correction, we will use Particle Swarm Optimization (PSO) technique. PSO is a stochastic optimization technique which makes the population of random solutions then search the optimal one from these random solutions. In the end, by using inverse DWT all the images are concatenated with each other and create the output enhanced image. 

Benefits of the Project

Water covers almost 71% of the earth’s surface. Nearly 60% population of living organisms live under the surface of the water. Today, at least 226408 marine species were discovered but according to the scientist, there are 25 million marine species which are not discovered yet. There are a lot of geographical changes occur under the surface of the water. Underwater images have an integral role in the study of the marine world because these images carry the critical information which is helpful in mine detection, diver visibility 
and different kind of searches under the water. Underwater cameras are the main tool for examining the different  activities in the water. Mostly these cameras are mounted on different underwater vehicles like autonomous underwater vehicles (AUVs), unmanned underwater vehicles (UUVs) and on ocean sensor network. Underwater images are also captured by the sonar. Quality of sonar image depends upon the wave strength of the sound. Capturing the high-resolution images in the water is a complicated task. Underwater  images with low contrast, low visibility, and poor natural appearance are captured due to the degradation effects of light spectrum and impurities which exist in water. So ETELQUI enhance your underwater images. ETELQUI ables you to see that important information which is hidden due to all the above problems. ETELQUI eridicate the blurriness from your images and enhance the underwater images.

Technical Details of Final Deliverable

We use C Sharp language as a core development language for this project. We use object oriented paradigm. In this project, we use EmguCV for the image processing capabilities. We also use Bunifu framework for the beautification of user interface of ETELQUI. The IDE use for this project is Microsoft Visual Studio 2017.

Final Deliverable of the Project

Software System

Type of Industry

IT

Technologies

Artificial Intelligence(AI)

Sustainable Development Goals

Life Below Water

Required Resources

If you need this project, please contact me on contact@adikhanofficial.com