Computational Xmas Tree lights!

GSD6338

Last update: Dec 23, 2022

Related tags

Image Processing XmasTree

Overview

Computational Xmas Tree

This repo contains the code for the computational illumination of a Christmas Tree!

It is based on the work by Matt Parker from @standupmaths, and his video "I wired my tree with 500 LED lights and calculated their 3D coordinates".

This version contains an original LED 3D calibration routine developed by @range-et based on Matt's explanations on the video, Grasshopper + C# scripts to generate animated light sequences developed by @garciadelcastillo, and python scripts for the Raspberry Pi to load and run them on the tree.

The main difference is that Matt's approach uses custom Python scripts with sequences ran directly on the RPi, whereas this workflow "bakes" the animations to CSV files that can then be ran from the RPi. The advantage of this approach is that the CSV sequences can be parsed on any graphics environment and displayed/simulated without needing to compile Python code. The downside is that the coordinates of the LEDs need to be known a priori before generating the CSV file (CSV files become tree-specific).

Jose Luis did a video series on ParametricCamp explaining (most of) this workflow: https://youtu.be/ma4JF6b4Dqk

Usage

Calibration

If you set a physical Christmas tree with lights, you can get an estimate of their 3D position using the following:

Run imageCapture.py from a Raspberry Pi with a webcam and connected to the tree. The script will take one image per LED for angle 0: $ sudo python3 imageCapture.py
Repeat this process by physically rotating the tree 45, 90, 135, 180, 225, 270 & 315 degrees, and correspondingly updating the MASTERDIR variable in the script. A sample dataset is provided.
Run mapper.ipynb to process all the images and generate the coordinates of the tree. This may take some time!
Run adjuster.ipynb to correct some of the outliers.

A version of the final outcome from our test tree is provided in file coords_adjusted.txt. If you don't have a tree yourself, feel free to use these as the starting point.

Questions, suggestions and PRs can be directed to @range-et!

Sequencing

Once 3D coordinates are available, the following C# Scripts can be used as inspiration to generate sequences:

(optional) The Grasshopper files use a version of the tree coordinates where X and Y are normalized between -1 and 1, and Z is proportionally scaled from 0 to whatever. Additionally, the coordinates are in CSV format. Use the xmas_tree_coord_normalizer.gh script to translate between the two syntaxes.
The xmas_tree_sequences_3d.gh and xmas_tree_sequences_no_3d.gh contain scripts to generate a few animated sequences, both spatially-aware (tree-specific) and not respectively.
(optional) If you only want to simulate already baked sequences, use xmas_tree_simulation.gh

Execution

A few scripts, based on Matt's original code, are provided to load the CSV sequences and run them on the tree.

Copy the content of the execution folder to the RPi, as well as a few of the CSV sequences.
To run a sequence: $ sudo python3 run.py light-sequence.csv
To run all the sequences in a folder: $ sudo python3 run-folder.py seq-folder/ 3 30. The optional number 3 indicates that each sequence will be looped 3 times before moving on to the next, and the optional 30 indicates that sequences will be blended over 30 frames.
If you stop a sequence and want to turn off the tree, run $ sudo python3 flush.py
If you need to fine calibrate the tree, you can turn on specific lights by ID running $ sudo python3 turnon.py 0 99 199 299 399 499.

Contributions

This repo is just a snapshot of the work we did in the Fall 2021 edition of the GSD-6338 course. Unfortunately, we do not have the resources to maintain, manage or extend it beyond what is available. If you want to ask questions, discuss standards, contribute new code, features and add new sexy goodness, please refer to Matt's repo or start your own fork. Crediting is always welcome, thank you!

Credits

Thanks to the superstar roster of TAs Indrajeet, Amelia and Jessica for putting together the tree and writing the foundational calibration code.

Big 💯 to all the students who gave it a try to write sequences for the tree!

Project led by Jose Luis Garcia del Castillo y Lopez as part of our work at GSD-6338: Introduction to Computational Design at the Harvard University Graduate School of Design.

Original idea by Matt Parker from @standupmaths.

Comments

Use RGB color ordering

The LEDs never were taking the values in GRB order. However, the library thought they were (because of a tricky default), so it was changing the order before sending it to the LEDs, which made it look like you had to set the colors in GRB order. You never have to if you initialize the library with the correct color ordering.

This initializes the library with the right color order setting, and removes the work-around.

opened by JasperHorn 3
Live trees with WASM

You can't currently watch a livestream on this christmas tree, but I think you could allow people to use WASM binaries that use imports of HTTPS Requests (and import a request executor, and have support for websockets), and then timing functions (like requestAnimationFrame, and some utility functions/datas (setLight(s), lightPositions) to then compute on-the-fly while not needing to worry about dependencies (because they can all be compiled in). This could also let you program a 3D Clock, event countdown, etc! And because it's WASM, it'll work on any tree!

(And also, for security reasons, display to the tree-owner the urls the script needs to access)

opened by AlexBigCheese 0
Player
I made a player for the CSV files:

https://github.com/Diggsey/xmas_tree

You just need Rust.

cargo run --bin xmas_tree_player --release -- sequences/barber_pole.csv
opened by Diggsey 10
2 scans and done method

Were I to scan a tree, I would do one front facing as normal and a second from above looking down the axis to capture the radius and angle of each light, from which the remaining 2 dimensions are obtained in one stroke, so all one needs from the first scan is the height. Faster scanning, more accurate estimates and surely, faster compiling, will be the result. A routine to compile the location estimates based on this type of data structure would be needed (I am not a programmer and new to GitHub) Richard

opened by feralmathematician 3
Rewritten the python runners to be more robust and add FRAME_TIME column

This rewrites the python run.py and run-folder.py quite dramatically.

Previously the CSV parsing code assumed that the columns were positional which meant adding more columns was impossible and would break if the colums were not in the "correct" order. This version reads the headers to find where they are. It will also populate any missing column with default values.

It also adds an optional FRAME_TIME column so that the user can specify how long each frame should remain for. If this is not defined it will run the frames as fast as the hardware will run them.

Fixes #2

opened by gentlegiantJGC 33

Computational Xmas Tree lights!

Related tags

Overview

Computational Xmas Tree

Usage

Calibration

Sequencing

Execution

Contributions

Credits

Comments

Use RGB color ordering

Live trees with WASM

Player

2 scans and done method

Rewritten the python runners to be more robust and add FRAME_TIME column

Owner

GSD6338

This app finds duplicate to near duplicate images by generating a hash value for each image stored with a specialized data structure called VP-Tree which makes searching an image on a dataset of 100Ks almost instantanious

Xmas-Tree-GIF-Tool - Convert any given animated gif file into an animation in GIFT CSV format

Build Xmas cards with user inputs

Christmas face app for Decathlon xmas coding party!

Use your Philips Hue lights as Racing Flags. Works with Assetto Corsa, Assetto Corsa Competizione and iRacing.

Custom component to calculate estimated power consumption of lights and other appliances

A hand tracking demo made with mediapipe where you can control lights with pinching your fingers and moving your hand up/down.

An OSINT tool that searches for devices directly connected to the internet (IoT) with a user specified query. It returns results for Webcams, Traffic lights, Refridgerators, Smart TVs etc.

Addon to give a keybind to automatically enable contact shadows on all lights in a scene

A simple Python script for toggling Philips Hue Lights by clapping

An open source bias lighting program which syncs up colored lights to the contents of your screen.

Python api wrapper for JellyFish Lights

Unsupervised text tokenizer focused on computational efficiency

Unsupervised text tokenizer focused on computational efficiency

ReproZip is a tool that simplifies the process of creating reproducible experiments from command-line executions, a frequently-used common denominator in computational science.

House-GAN++: Generative Adversarial Layout Refinement Network towards Intelligent Computational Agent for Professional Architects

JittorVis is a deep neural network computational graph visualization library based on Jittor.

This repository contains the code, data, and models of the paper titled "XL-Sum: Large-Scale Multilingual Abstractive Summarization for 44 Languages" published in Findings of the Association for Computational Linguistics: ACL 2021.

This repository contains the code, data, and models of the paper titled "XL-Sum: Large-Scale Multilingual Abstractive Summarization for 44 Languages" published in Findings of the Association for Computational Linguistics: ACL 2021.

Natural language computational chemistry command line interface.