Running the Trap#


Prerequisites for this page are, that your Raspberry Pi

  1. Has Raspberry Pi OS Buster,

    Note that this is not the latest Raspberry Pi OS. Check /etc/os-release to see what VERSION_CODENAME you have. It must be buster.

  2. Has the camera module enabled, with a camera module that is working and in focus,

  3. Has an expanded filesystem,

  4. you are comfortable logging in to your Raspberry Pi remotely over ssh,

  5. and DynAIkonTrap is installed.

If not, please refer to the sections Preparing the Raspberry Pi and Installing DynAikonTrap.

Modes of Operation#

We currently support two modes of operation

  1. Live-mode: DynAIkonTrap can operate a camera module and work as an in-field Camera Trap.

  2. Emulated-mode: DynAIkonTrap can operate on previously captured video files to detect animals.

Live mode#

  1. Connect to your Raspberry Pi by using ssh from the terminal.

  2. Start the DynAIkonTrap with


    After a few moments you should see the output appearing. If you wave your hand in front of the camera you should begin to see messages about detected movement appearing.

  3. To stop DynAIkonTrap hit Ctrl+C.

If you have reached this point, then well done! You have successfully set up the camera trap.

Changing your path

To save typing the long ~/.local/bin/dynaikontrap everytime, you can add the line

export PATH=$PATH:$HOME/.local/bin/dynaikontrap

to the file ~/.bashrc. Now dynaikontrap will launch with


The remaining sections on this page have some useful recommendations on how to best use this camera trap. We recommend that you have a look at the Tuning the Trap guide before deploying the camera trap properly to make sure your system is fully optimised for your use-case.

Long-term deployment using screen#

If you start the Trap using the dynaikontrap command, the program will stop as soon as you log out of the Raspberry Pi. This is not very useful as you will likely not want to keep the terminal connection open for days or weeks on end. A simple solution is to use the screen command.

Install screen on the Raspberry Pi with

sudo apt install screen

then issue the following commands:

# Start a new screen session called "dynaikontrap"
screen -S dynaikontrap

# Start the camera trap within the screen session

You can now exit the screen session without stopping the camera trap by typing Ctrl+A, and then the D key to “detach” from the session.

Now when you log out from the Raspberry Pi, the camera trap will continue to run.

Checking progress#

You can check progress easily using our DynAIkonTrap web-viewer! This is a server hosted on the deployed device. For more information on how to use the web-viewer, check out our Web Viewer.

One can also use screen to monitor progress directly over ssh. This is easily done by starting an ssh session to the RPi. You can then reattach to the screen session using:

screen -r dynaikontrap

You will be able to see any logs produced by the DynAIkonTrap.

Stopping the long-term deployment#

Reattach to the screen session as mentioned above for Checking progress. Once in the dynaikontrap session use Ctrl+C to quit the DynAIkonTrap code.

It is also safe to simply shutdown the Raspberry Pi by running:

# "-h 0" means to "halt" in 0 seconds i.e. now
sudo shutdown -h 0

The camera trap code will not automatically start again when the Raspberry Pi is powered on. Remember to unplug the Raspberry Pi once it is shut down as it will continue to draw a very small amount of power if left plugged in.

Retrieving Observations from the Camera Trap#

The absolute simplest option for a novice Raspberry Pi user may be to shutdown the Raspberry Pi and to plug the SD card into their computer to access the files directly.

Reading files directly from the SD Card

To read the files directly from the SD card you will need to be using Linux on your main computer. Windows cannot read files from an ext4 filesystem.

One can also retreive observations over the internet, using Secure Copy (SCP) over SSH

scp -a <username>@<hostname>.local:~/dynaikontrap/output/ ./


scp -a ecologist@raspberrytrap.local:~/dynaikontrap/output/ ./

copies all files from the default video output directory onto the current directory on your computer.


A slightly more complicated solution that allows automatic saving of files to a separate device is as follows. If you have a second Raspberry Pi you could use this as a server. Let’s state some assumptions:

  • The camera trap is called dynaikontrap

  • The output directory has been set to ~/videos

  • The second computer (could be a second Raspberry Pi) is called server

On dynaikontrap you could then run:

sshfs ~/videos ecologist@raspberrytrap.local:~

to automatically save all files from dynaikontrap’s output to the server’s home directory. Note that sshfs may not be installed, but you can install this with sudo apt install sshfs on Ubuntu/Debian systems. In this configuration the files are actually saved physically to server, so you could have a more reliable hard disk drive on this device and serve the files to other devices connected on the local network.


The camera trap does have a RESTful server API, but code for the server is not released. This is left as an exercise for the reader. Using frameworks like Django can make this a fairly simple process. We do not have the resources to write and maintain the necessary code for this, but we would be happy to answer questions you may have and hopefully help you set something up.


DynAIkonTrap integrates with DynAIkon’s web API, FASTCAT-Cloud. This may be used to upload detections automatically to your account through our API endpoints. You can configure the camera trap to do this with your account details following instructions on the Tuning the Trap page.

Emulated mode#

DynAIkonTrap may also be run on pre-recorded video, emulating an in-field live camera trap. These pre-recorded clips will then be filtered using our AI video pipeline.

As a dependency, you will need to install Vid2Frames with

pip3 install
dynaikontrap --filename pre-recorded-observation.mp4

This will run the camera trap on the video input, watch the output log to see if animals are detected! When the video is processed, exit the program with Ctrl+C.