Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/secretshardul/golem-photogrammetry

Run Micmac photogrammetry workloads on Golem's decentralized cloud.
https://github.com/secretshardul/golem-photogrammetry

golem micmac node photogrammetry typescript

Last synced: about 2 months ago
JSON representation

Run Micmac photogrammetry workloads on Golem's decentralized cloud.

Awesome Lists containing this project

README 

        
# Photogrammetry (3D scanning) on Golem

An online IDE which lets you run Micmac photogrammetry workloads on Golem. Photogrammetry jobs are resource intensive and can run into hours. Golem is ideal for such batch workloads.



## Features

- Run Micmac builds online any local setup.

- Fully programmable. Look here for the entire list of commands.

- Run intense workloads on Golem cloud, without consuming local resources.

- Build results notified on email.



## Project demo

http://65.0.59.23/



Note: Ping @the.godfather on Golem Discord server or raise an issue if facing issues. Yagna daemon is unstable on EC2.



## Demo video

[![1 golem-youtube](https://user-images.githubusercontent.com/49580849/113905798-2a72a100-97f1-11eb-8365-1e080b9abbfb.png)](https://www.youtube.com/watch?v=A0-PHUgftMk)



### Photogrammetry for gaming

Chernobolyte is a sci-fi game based on 3D-Scanned Chernobyl. This game is also the inspiration for this project.



[![6-chernobolyte](https://user-images.githubusercontent.com/49580849/113909587-77587680-97f5-11eb-889d-b2048695b98c.png)](https://www.youtube.com/watch?v=vM6wGofkhxw)



https://80.lv/articles/chernobolyte-game-based-on-3d-scanned-chernobyl/



## Screenshots

### Web app

![2-app-top](https://user-images.githubusercontent.com/49580849/113908176-dfa65880-97f3-11eb-890b-f0dda6ac5988.png)



![3-app-bottom](https://user-images.githubusercontent.com/49580849/113908166-dd43fe80-97f3-11eb-983a-32ccd0febf9b.png)



### Email results

![4-success-email](https://user-images.githubusercontent.com/49580849/113908186-e1701c00-97f3-11eb-9ed4-291a15cbe4e4.png)



### Viewing generated `.ply` files in Meshlab

![5-results](https://user-images.githubusercontent.com/49580849/113908181-e03eef00-97f3-11eb-8590-6981d6f9fb30.png)



## Instructions to run locally

- Install node modules using `npm install`

- Create a `.env` file containing environment variables. This contains Sendgrid and Yagna provider keys. Copy from [`.env-template`](./.env-template) for reference.

- Start golem locally using `yagna service run`

- Open a new terminal and run `yagna payment init --sender`. Then start Node server using `npm run start:dev`

- Follow the instructions on `localhost:3000`



## Testing other photogrammetry workloads



- Download a sample dataset from https://micmac.ensg.eu/data/

- Extract the zip file and find the runner shell script

- The raw images and hyperparameters will be sent to a folder `/golem/resource/extracted/` on Golem. Accordingly modify the paths in the shell script.



![datasets](![image](https://user-images.githubusercontent.com/49580849/117320192-9c4e1100-aea9-11eb-8c49-16253bbf40e9.png)

)



## Tech stack

- Golem requester and Golem Node.js SDK

- Typescript

- Express.js

- Sendgrid email service



## Limitations

- Micmac is a low-level programming framework. There is no GUI, but rather developers need to supply a shell script program and hyperparameters.

- There's no generic script for multiple scanning workloads. Developers are expected to provide the orientation data and processing instructions.



## References

- Photogrammetry: https://en.wikipedia.org/wiki/Photogrammetry

- Micmac: https://github.com/micmacIGN/micmac

- Chernobolyte: https://80.lv/articles/chernobolyte-game-based-on-3d-scanned-chernobyl/