Software Beesbook

Beesbook detection and tracking software

Example - Getting Started

To test the video processing pipeline, use the example Jupyter notebook on Colab. You can track an example video or upload your own.

• Beesbook tracking example:

To install locally, run the install script (tested on Ubuntu 20+ and Mac), and get started with the example jupyter notebook (run in the installed 'beesbook' environment).

• Install script. Note that for GPU support on Linux, Nvidia and CUDA drivers are needed also; follow the Nvidia driver installation instructions at bb_imgacquisition - Basler Support
• Example notebook

Packages

The following tools are used to process to video files and extract individual bee detections and tracks, and perform related functions. The core packages include bb_binary, bb_pipeline, bb_pipeline_models, bb_tracking, bb_behavior, and bb_utils. Other useful tools include bb_analysis, beesbook_backend, and beetag_app.

Localizing and Detection Pipeline

The toolchain to localize bees and decode tags in images and videos can be found here, along with the trained models to run detection and tracking

• bb_pipeline
• bb_pipeline_models

The resulting data will be stored in the bb_binary format:

• bb_binary

Bee Tracking

The code to apply camera homographies, link detections (marked and unmarked) over time, assign unique track identifiers, and correct tag decoding errors can be found here:

• bb_tracking

A graphical user interface to annotate own detections to create training data for a tracking classifier is here:

• bb_analysis

Productivity

Code to run tracking and interact with a database based on the tracking results can be found here:

• bb_behavior

Helper tools to convert and draw BeesBook tags, detect fiducial markers, and more can be found here:

• bb_utils
• PDF with fiducial (placement) markers already generated

Python scripts to generate templates with the BeesBook binary code

• bb_tags

A browser-based GUI interface for recording videos and running the Beesbook detection and tracking pipeline:

• bb_gui

Code to read and save temperature from ds18b20 temperature sensors (which can be placed in a hive comb) with Arduino and Python:

• bb_temperatureloggers

Software for a server that can be used in an API to retrieve image data from frame IDs can be found here:

• beesbook_backend

Android-App to decode markers with your phone (outdated):

• beetag_app

Monitoring

Code to periodically send images and calculated quantities from current recordings to a Telegram messenger bot:

• bb_monitor