Making a light-up 3D LED Tree

A project for lighting LEDs, calculating their 3d locations, and then lighting up animations and images.

Test Setup

Run sudo python3 ./step0/test.py and verify that the script outputs "OK" and an image is captured in ./step0/out.

Capture

This step captures images of each LED lit up individually. It does take some time and requires you to rotate the tree on the ground.

Images are stored as jpegs as captures\{angle}\{led_num}.jpg where angle is the angle of rotation (in degrees) clockwise from the initial position and led_num is the led number on the strand.

1. Ensure the camera is set up in a fixed location and can see the entire tree. If needed, run the test script in 1.
2. Ensure the tree legs are marked so it's easy to turn the tree accurately. I put blue tape on a leg and directly in front of one leg so I can see 45 and 90 degree turns easily.
3. Start the script using

   sudo python3 ./step1/capture.py -s 0 -e 499
   

   When the script finishes capturing one angle. It will prompt you to turn the tree and press a button to continue.

The capture script supports starting and ending at any contiguous range. If a LED is present in one rotation, it will be in all of them. In some cases, the captured image may be completely black. This likely indicates that the LED is obscured by whatever it is hanging on. Don't worry too much about this. The processing scripts try to fix these issues.

Captured Images

Images should not be included in the repo since they are quite large. If desired, drop them to a file sharing and include a link below.

https://drive.google.com/drive/folders/1jmmNhNl4d18Hpue5HqjxG8EZfJDkkLYr

Converting images to coordinates

The process images consists of determining where the LED is on the image and outputting that to CSV. The x/y coordinates are then processed by the coordinate_processing.py script to output 3D coordinates using 2d to 3d projection and error correction.

These scripts use PIL, OpenCV, NumPy, and MatPlotLib as dependencies.

LED detection in images.

Let's convert images to x/y coordinates.

python3 ./step2/image_processing.py -s 0 -e 500 -i "/home/nomoor/Pictures/tree_captures" -o "/home/nomoor/Pictures/processed_tree_captures"

2D to 3D coordinates

python3 ./step2/coordinate_processing.py "

This step converts the x/y coordinates in the previous step to 3D coordinates that can be used for animation.

Create an animation for playback

Then you can read in the coordinates and generate an animation playback csv.

python3 test_animations.py -i hat_tree_coords_2021_v2.csv -o valentines_animation.csv

Playback animation

``