High-Level Project Summary
We have developed an android application that allows users to convert grayscale Juno cam images into RGB. At first, we developed a Deep Learning model which will import the images from the Juno cam android application. After retrieving the images and performing pre-processing techniques which include conversion of images to the array, splitting, reshaping, resizing, and feature scaling, The pre-processed images are then converted to RGB by the use of Open cv (image processing library).Furthermore, the RGB image then can be edited in the image editor as per the user requirements and visualization can be done .Integrated AR/VR [ Instagram & Facebook AR/VR Effects], accessing the METAVERSE.
Link to Final Project
Link to Project "Demo"
Detailed Project Description
WHAT IT DOES?
Our android application converts the Junocam grayscale images into colored images (RGB).
It removes the noisy data and gives us clear and processed images.
TOOLS, CODING LANGUAGE, AND SOFTWARE USED ARE:
1. To train machine learning models - We have used various python libraries like Numpy (numerical data), Open CV(image processing), OS (sourcing path to directory).
2. To deploy and colorize - We have used two deep learning models: deploy.prototxt and colorize.caffemodel .
3. To develop an image editor - We have used intermediate python.
4. For android development - We have used Android studio.
5. For visualization - We have used Augmented reality(AR).
HOW WORK?
We have used different python libraries and deep learning algorithms in our project.
The Deep Learning model (deploy.prototxt, colorize.caffemodel) which we have used will import the images from the Junocam android application.
Our model will work in four different phases including :
Phase 1 - In Phase 1 we will Train a Deep Learning Model for the conversion of grayscale junocam images where it will retrieve and preprocess the image by splitting, reshaping, resizing, and feature scaling the images. The preprocessed images are then converted to RGB by the use of Open cv (image processing library).
Phase 2 - In phase 2 we will Develop an Image Processing Editor for Junocam images where RGB images will be edited in the image editor as per the user requirements.
Phase 3 - In this phase, we will Develop an Andriod Application using android studio.
Phase 4 - In our last stage we will use AR/VR Effect [ Instagram & Facebook AR/VR Effects ] for visualization and Application Guide.
BENEFITS:
Our application removes the noisy data and gives us the clear and processed image.
It provides us with modified images.
It provides scalibility and reliability.
It is user-friendly, and faster to use.
HOPE TO ACHIEVE:
We hope to achieve a smart but simpler way to extract data from junocam images and visualize it in a broad spectrum by using AR (Augmented reality).
Space Agency Data
SPACE AGENCY DATA:
We have used the resources (images) provided by Nasa Space App and the Junocam website.
The Images provided by them were grayscale. we have converted the grayscale images into coloured images (RGB).
The images provided by NASA were very fascinating.
While processing the images we got to learn about RGB (Red, Green, Blue) factors in the minutest details.
We also got to know about new computer science technologies like Computer vision, Python libraries, and Deep learning.
References:
https://www.nasa.gov/solve/feature/junocam
Hackathon Journey
The journey during these two days was extremely wonderful.
Our team enjoyed it to the fullest.
We were very fortunate to interact with some great personalities on the very first day of the inauguration ceremony which provided us immense motivation.
The approach to develop this project was, we got to know about Different python libraries, Computer vision techniques, and Deep learning models.
As students of computer science and AIML, we had prior knowledge about Python, and image processing using Computer vision.
We were very glad to be a part of this hackathon. which has helped us to build our confidence, and work on our communication skills.
In future also we would be glad to participate in more such hackathons.
References
REFERENCES:
TOOLS WE HAVE USED:
1. To train machine learning models - We have used various python libraries like Numpy (numerical data), Open CV(image processing), and OS (sourcing path to directory).
2. To deploy and colorize - We have used two deep learning models: deploy.prototxt and colorize.caffemodel .
3. To develop an image editor - We have used intermediate python.
4. For android development - We have used Android studio.
5. For visualization - We have used Augmented reality(AR).
FURTHER REFERENCES ARE:
https://www.missionjuno.swri.edu/junocam/processing
https://www.gimp.org/
https://www.nasa.gov/solve/feature/junocam
https://www.nasa.gov/mission_pages/juno/overview/index.html
Tags
#DeepLearning #Imageprocessing #computervision #python #AR/VR #Machinelearning #JunoCam #Joveian #jupiter #missionjuno