pyFishSenseDev

This repository contains a Python implementation of the Fishsense lite pipeline. It contains tools for camera calibration, laser calibration, and fish length calculation. This pipeline assumes that you already have access to all the data necessary to run this pipeline. As this code is meant to be a temporary measure, the user may need to edit some file paths, so apologies in advance.

After cloning

1. Install the poetry packages. poetry install.
2. Install git annex. On Ubuntu, execute sudo apt install git-annex.
3. Install git-annex-remote-synology from git-annex-remote-synology.
4. Run git-annex-remote-synology setup --hostname e4e-nas.ucsd.edu
5. Run git annex enableremote synology.
6. Download data git annex copy --from=synology.

Development Dependencies

We provide a docker container which has the dependencies pre-installed. In order to use this, please ensure you have docker installed on your system. When running Visual Studio Code, ensure that you use the option to reopen in the container. This step will time for the intial setup.

Execute Unit Tests

poetry run pytest

Old

Steps to use

1. Run poetry install. This will install all required dependencies.
2. Run poetry shell to execute all following commands within the poetry environment.
3. Create a data folder in this directory. Add a raw subdirectory and add your raw data inside. This can take whatever folder structure you need.
4. Process all raws by running python camera_imaging_pipeline/process.py. This will place all images inside data/png, copying whatever folder structure is inside data/raw.
5. Calibrate the lens using python e4e_camera_calibration. Documentation on this tool can be found here.
6. Undistort all laser and fish images using python undistort_images.py. This will place all images inside data/png_rectified, copying whatever folder structure is inside data/png.
7. Label laser calibration images using python label_laser_calibration.py.
   • This script uses the following command line arguments:
     • -i or --input specifies the input path which contains all rectified laser images.
     • -o or --output specifies the location to save the generated csv file.
   • Use the following key bindings when using this tool:
     • ESC will skip the current image, use this when the laser dot is difficult to find, the calibration board is not completely in frame, or things are out of focus.
     • r will remove the current label, in case you make a mistake
     • e will save the added label and load the next image
8. Calibrate the laser using python laser_calibration.py. This script uses the following command line arguments:
   • -c or --calib specifies the location of the lens calibration file generated from step 3.
   • -i or --input specifies the location of the csv file from part 5.
   • -o or --output specifies the location to save the generated laser calibration file.
9. Label the fish images using python label_laser_fish.py
   • This script uses the following command line arguments:
     • -i or --input specifies the input path which contains all rectified fish with laser images.
     • -o or --output specifies the location to save the generated csv file.
   • Click to label the laser dot, snout tip, and tail fork, in that order.
   • Use the following key bindings when using this tool:
     • ESC will skip the current image, use this when the laser dot is difficult to find, or things are out of focus.
     • r will remove all current labels, in case you make a mistake
     • e will save the added label and load the next image
10. Run the fish analysis using python fish_mass.py. This script uses the following command line arguments:
    • -c or --cameracalib specifies the location of the lens calibration file generated from step 3.
    • -l or --lasercalib specifies the location of the laser calibration file generated from step 6.
    • -i or --input specifies the location of the csv file generated from step 7.
    • -o or --output specifies the location of the generated csv file containing fish lengths and masses.