SF6 .jemdoc

# jemdoc: menu{MENU}{index.html}, nofooter  
==Research Details

Catalogue:
. [Underwater.html Underwater Image Enhancement with a Total Generalized Variation Illumination Prior]
. [SF6.html The Design and Realization of Methane and Carbon Dioxide Infrared Gas Sensors (SRTP)]
. [Modeling.html Design of Thermal Protective Clothing Based on Heat Conduction Partial Differential Equation Model 09/2018 (China Undergraduate Mathematical Contest in Modeling)]


== The Design and Realization of Methane and Carbon Dioxide Infrared Gas Sensors (SRTP) \[[http://magazine.ijournals.net.cn/jgyhw/article/abstract/m500115 pdf]\]
	

=== Purpose
High-quality underwater images are of great significance for human exploration and excavation. There are two interfering factors:
. Low contrast and atomization -> underexposed and blurry
. Color deviation -> bluish or greenish \n

We designed a variational model based on Retinex to address these two issues.

=== Method

Three steps to promote the quality of underwater images:

~~~
{}{img_left}{underwater_figure1.jpg}{alt text}{500}{170}
~~~

. Color correction
. +*Retinex-based decomposition with a total generalized variation prior*+
. Post-processing

In step2, Retinex algorithm is applied to decompose the reflection and illumination. Following the nature of the illumination, the total generalized variation (TGV) prior is introduced to the illumination:

~~~
{}{img_left}{underwater_figure2.jpg}{alt text}{300}{70}
~~~

To solve the proposed model, a high-efficiency iterative optimization scheme is designed and implemented by Matlab. The scheme can be summed up in *Algorithm 1* and *Algorithm 2*.

~~~
{}{img_left}{underwater_figure3.jpg}{alt text}{300}{300}
~~~

=== Experiments

We carried out plenty of experiments are carried out and evaluate the enhanced images from two points: *visual effect* and *objective evaluation*.

~~~
{}{img_left}{underwater_figure4.jpg}{alt text}{600}{600}
~~~

~~~
{}{img_left}{underwater_figure5.jpg}{alt text}{550}{200}
~~~