https://github.com/jsten07/countyourbirds

An application to automatically recognize birds, specify the species, count and save pictures of them, only by use of one camera and a Raspberry Pi. The images and recorded values are made available online on https://opensensemap.org/.
https://github.com/jsten07/countyourbirds

birdfeeder birds birdwatching citizen-science classification detection image-recognition motion opencv raspberry-pi raspberry-pi-camera tensorflow tensorflow-lite

Last synced: about 1 month ago
JSON representation

An application to automatically recognize birds, specify the species, count and save pictures of them, only by use of one camera and a Raspberry Pi. The images and recorded values are made available online on https://opensensemap.org/.

• Host: GitHub
• URL: https://github.com/jsten07/countyourbirds
• Owner: jsten07
• License: gpl-3.0
• Created: 2020-12-10T13:20:50.000Z (almost 4 years ago)
• Default Branch: main
• Last Pushed: 2024-02-07T14:18:39.000Z (9 months ago)
• Last Synced: 2024-10-11T16:41:17.492Z (about 1 month ago)
• Topics: birdfeeder, birds, birdwatching, citizen-science, classification, detection, image-recognition, motion, opencv, raspberry-pi, raspberry-pi-camera, tensorflow, tensorflow-lite
• Language: Python
• Homepage:
• Size: 17.4 MB
• Stars: 16
• Watchers: 6
• Forks: 4
• Open Issues: 1
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# countYourBirds

This is an application to automatically recognize birds, specify the species, count and save pictures of them, only by use of one camera and a Raspberry Pi. The images and recorded values are made available online on https://opensensemap.org/. 

By using [Motion](https://pypi.org/project/motion/), movements are detected and corresponding images are captured. Afterwards, an object detection by [TensorFlow Lite](https://www.tensorflow.org/lite) is used to check if birds are present on the image. If birds are present in the images, the images are examined with another tensorflow lite model and an attempt is made to identify the species to which the bird belongs. Accordingly, the images on which birds were detected are stored and the detected birds are counted, also separated by species. The images and values are then automatically made available on [openSenseMap](https://opensensemap.org/) and can be retrieved at any time. 

The software is provided in relation to Citizien Science in such a way that citizens with a camera, a Raspberry Pi and the instructions available here can also use the application. It is convenient to install it together with a bird feeder. 

Instructions and all other information can be found in the [wiki](https://github.com/jsten07/countYourBirds/wiki).  

This software was developed as part of the study project "Image Recognition for Citizen Science" in the winterterm 2020/21 at [ifgi](https://www.uni-muenster.de/Geoinformatics/en/index.html) at the [University of Münster](https://www.uni-muenster.de/en/). 

The application was advanced in the [Birdiary](https://github.com/Birdiary) project, that shows collected data on the website [www.wiediversistmeingarten.org](https://wiediversistmeingarten.org/view/).