https://github.com/aysark/sky-insight

Hack the Map 2018 Winning Entry: Easily train machine learning models using tile maps & ArcGIS.
https://github.com/aysark/sky-insight

arcgis esri hackathon image-classification machine-learning satellite sky-insight

Last synced: 3 months ago
JSON representation

Hack the Map 2018 Winning Entry: Easily train machine learning models using tile maps & ArcGIS.

• Host: GitHub
• URL: https://github.com/aysark/sky-insight
• Owner: aysark
• Created: 2018-01-27T20:11:46.000Z (about 7 years ago)
• Default Branch: master
• Last Pushed: 2018-02-02T05:24:57.000Z (almost 7 years ago)
• Last Synced: 2024-08-01T00:52:43.439Z (6 months ago)
• Topics: arcgis, esri, hackathon, image-classification, machine-learning, satellite, sky-insight
• Language: JavaScript
• Homepage:
• Size: 31.6 MB
• Stars: 8
• Watchers: 5
• Forks: 6
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# sky-insight

![Demo Video](src/assets/sky-insight-preview-small.gif)

sky-insight allows you to:

- Easily train image classification machine learning models using satellite/aerial tile imagery & ArcGIS.

- Enrich your existing GIS data with real world planetary data.

- Monitor the physical world with ease using automation & image classification.

It features two modus operandi:

### Search

![Preview](src/assets/preview.png)

Search mode allows you to use your trained models or any existing models to automatically infer/predict on an area of interest of tiles.  You can then export your predictions as feature sets or geojson.  Above image is a ship tile classifier. 

### Train

![Preview](src/assets/preview2.png)

Train mode allows you to easily train your own image classification models using a simple tile annotation tool.  So you can build ship, airplane, oil tank, pivot irrigation farm classifiers- or whatever you fancy.  Above image is training of a pivot irrigation tile classifier. 

## Build & Run

``` bash

# install dependencies

npm install

# rename sample.env.js to env.js and fill in tokens

# serve with hot reload at localhost:8080

npm run dev

# build for production with minification

npm run build

```

For detailed explanation on how things work, consult the [docs for vue-loader](http://vuejs.github.io/vue-loader).