Using machine learning, deep learning, mini batch stochastic gradiant decent, and some AI to compute the surface area of ice from a 100 X 100 pixel .png image and return the projected inflow/outflow. This method uses eigen vectors taken from a training model and projects them onto a test model to predict an independent variable (y_hat). This y_hat can be surface area occlusion, outflow/inflow, water height. The idea is that this system would use the existing infrustructer to reduce current computation and memory requirements to a small fraction. Additionally, the project aims to deploy ultra light hardware that can run the regression function on site at < 80 milliamps.
Requires the standard data science libraries for python3.
import pandas as pd
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import MinMaxScaler
#Import the right package
from scipy.io import loadmat
from PIL import Image
from numpy import linalg as LA
import matplotlib
import matplotlib.pyplot as plt
import numpy as np
import imageio
import sys
import gc
import math
import random
These instructions will get you a copy of the project up and running on your local machine for development and testing purposes. See deployment for notes on how to deploy the project on a live system.
A step by step series of examples that tell you have to get a development env running
Say what the step will be
Give the example
And repeat
until finished
Explain how to run the debugging 'watch' script for this repo, if applicable
Explain how to run unit tests, if applicable
Add additional notes about how to deploy this on a live system. Do not include any credentials, IP addresses, or other sensitive information
So far the system just requires python3. In the future, This system will need boto3 and aws components. I may rewrite this program in RUST.
- Python 3 - The main web framework used
Please read CONTRIBUTING.md for details on the process for submitting pull requests to us. Please read CODE_OF_CONDUCT.md for details on adhering by the USGS Code of Scientific Conduct.
We use SemVer for versioning. For the versions available, see the tags on this repository.
Advance the version when adding features, fixing bugs or making minor enhancement. Follow semver principles. To add tag in git, type git tag v{major}.{minor}.{patch}. Example: git tag v2.0.5
To push tags to remote origin: git push origin --tags
*Note that your alias for the remote origin may differ.
- Daniel Beckman - Lead Developer - USGS Web Informatics & Mapping
See also the list of contributors who participated in this project.
This project is licensed under the Creative Commons CC0 1.0 Universal License - see the LICENSE.md file for details
In the spirit of open source, please cite any re-use of the source code stored in this repository. Below is the suggested citation:
Code is scratch code so far. This will change. Lots of other people's work in math and alogothims needs to be added.
This project contains code produced by the Web Informatics and Mapping (WIM) team at the United States Geological Survey (USGS). As a work of the United States Government, this project is in the public domain within the United States. https://wim.usgs.gov
- Hat tip to anyone who's code was used
- Inspiration Note
- This project authored by the USGS WIM team
- WIM is a team of developers and technologists who build and manage tools, software, web services, and databases to support USGS science and other federal government cooperators.
- WIM is a part of the Upper Midwest Water Science Center.