High-Level Project Summary
So basically our task coincides of performing earth analysis on climatic change . So we chose the topic of Global Warming and how it affects us in our daily lives. We were able to showcase a regression model which we able to train in such a way that using custom ROI we were able to come up with some astonishing results. Our datasets includes glaciers, flood affected areas , canals, oceans etc. Thus using google earths pro timeline feature we extracted satellite images with 5 years of range between each image. Therefore our model was able to give us predictions of will happen to glaciers in the upcoming years. Our project will help humanity and authorities a lot to alert what's coming.
Link to Final Project
Link to Project "Demo"
Detailed Project Description
1) Satellite images using google earth pro of glaciers, canals, streams, oceans.
2) Range = 5 years of timeline years
3) Preprocessing using pythons language's library open cv.
4) Separating each color pixel values using HSV ( Hue Saturation Value) which tells us the darkness of each color. The darker the color the more quantity it is present in and vice versa and were able to come to this outcome of donut chart.
5) ROI (Region of Interest) was able to give us some accurate predictions by rectangulating an area where the most change in water level was observed.
6) Authorities can take advantage of this as using these prediction our model makes future predictions so we can know about any disaster that would strike us in the upcoming years and warn the citizens about it and make earth an habitable place.
Space Agency Data
Earth Analysis on climatic change.
I use Image processing for this project because I thought it I can easily analyze the data using pixel count . And got the predictions for next coming years. I got this idea from Xbox. Why not use image processing to solve this problem.
Hackathon Journey
Really commendable experience!
Learnt a lot, Really good interaction with peers.
Global Warming was already our project so we gave it a slight touch combined with NASA development.
References
Dataset = Kaggle / google earth pro
Interface made using react Js
image processing using python language
Tags
#advance#software#globalwarming#AI#modeltraining#modelpredicition