Thermal Camera for 100% detection

Problems with normal tracking cameras

Prototype Cacophonometer with Raspberry Pi and multiple cameras.

The goal of this project is to develop tools that eliminate 100% of predators. To do this the device must therefore be able to detect 100% of predators. Chew cards and tracking tunnels can miss over 60% of predators. Standard camera traps are thought to miss as little as 5% of predators due to not starting fast enough or the light/sound scaring animals away. There are also issues with false positives making it difficult and time consuming to filter the videos

Our goal was to test the different camera set ups to see how well the different technologies worked. One camera is an off-the-shelf tracking camera. The other two are Raspberry Pis - one with a IR camera, light and motion sensor the other with a thermal camera that is taking a picture every 2 seconds.

Advantages of thermal cameras

  • 100% detection possible
  • get full video of animal behaviour - the camera does not turn on and off is animal is still (not triggering motion sensor)
  • no false positives to trawl through
  • no delay is start up
  • it may be easier for Machine Learning method to detect differences in heat patterns rather than light patterns (even though it may be harder for humans to tell the difference)
  • no shutter noise

Video 1: Possum climbing a tree with Thermal Camera

Where to next?

The next set of projects aim to

  • run some code on the camera to work out when it is activated
  • send the files to the cloud
  • convert into viewable video
  • run the machine learning to auto detect animals
  • have the device activate audio and video lures to rapid experiments can take place
  • ultimately link this to a device that can activate a kill method