eTechnology Hub: Innovation through digital technology

Our Centre has brought together the best minds in Australia to use the latest innovation in digital technology to manage pest animals more effectively – from completely novel formats such as the Wild Dog Alert nodes through to automated gate mechanisms for feral pig traps or feral cat lures.

The project team is a collaboration between NSW Government and University of New England scientists and professional staff, along with contracted professionals who are engaged for their specific expertise.

This unique mix of experts ensures that the aspirations of on-ground practitioners (the technology users) and the specialist technical expertise (the technology specialists) are unified so that practical but robust tools are designed and tested.

There are four innovative projects within the ‘e-technology hub’ which are solely focused on trying to integrate technology and automation into new and existing tools for both livestock and biodiversity enhancement via effective pest management.

The outcome is to create practical and accessible tools for practical people.

Current projects include:

  1. Wild Dog Alert
  2. Feral Cat Tactile Lure
  3. Autonomous Gate Trap and Remote Recognition Response Device
  4. Sentinel Bait Station

Wild Dog Alert (WDA)

The original objective of the WDA project was to deliver a device that could detect and transmit a message of a wild dog detection from a remote location and initiate an alert in real-time. During the evolution of the research into camera traps and this device, it was recognised by the team that using off-the-shelf camera traps to build a wild dog alert was not going to achieve the objective to a high level.

As a result the project team has built a world-first device called Wild Dog Alert Node (WDAN), a tri-sensor, 360 degree detection and photographic zone, stand-alone power source, in-built computing power and satellite transmitting device. To communicate with this device we have also had to build a Cloud System to process data and send alerts, software to automate image processing and analysis and a WDA phone application to receive alerts. Further, the team have also integrated the WDA system into off-the-shelf camera traps using the telecommunication network. This includes an SMS camera trap system and a VHF-telecommunication camera trap system.

  1. Wild Dog Alert Node (WDAN): The WDAN is both a technological test-bed and our prototype solution to monitoring of dogs in remote tableland and outback areas. The WDAN incorporates technology otherwise unavailable on commercial camera traps (such as Tri-Tech sensors). The WDAN is solar-powered and contains a small-computer which processes the images captured by the 360-degree camera and sensor system. On-board artificial intelligence software detects the presence of a dog and sends a text alert message via satellite to the WDA Cloud service and on the user’s app (or email). The WDAN is designed to operate independently in remote areas and via integration with the WDA Cloud.
  2. Wild Dog Alert Buckeye Cam PC Base: The Buckeye Cam PC Base is a commercially available camera trap system which transmits images over its own radio telemetry network using VHF. Transmitted images are received via a base station computer which is connected to the internet. The base station computer is powered by mains power or a battery bank with an optional uninterruptable power supply. The base station computer then transmits the images to the WDA cloud for further processing and alert. The WDA PC Base system is intended for properties on the order of several square kilometres based around a homestead with internet connectivity (even via satellite internet) and regular power supply.
  3. WDA 3G/4G cellular system: This WDA SMS system is designed to operate in areas of cellular coverage using commercially available camera trap systems. Such camera traps can transmit image data directly over the cellular network to the WDA Cloud service and hence provide image data for alerts. Each 3G/4G camera can operate and transmit independently within cellular coverage. The limitation with the WDA 3G/4G cellular system is of course the limited coverage in rural areas.
  4. ClassifyMe: The team have also developed species recognition software which is downloaded from a web-site https://classifymeapp.com and installed on the end-users personal computer. The software hosts a range of species identification models and importantly a model designed to detect the presence of dogs. The ClassifyMe software is designed to scan SD cards extracted from standard commercially available camera traps. It provides an option to end-users with low budget and little to no telecommunications infrastructure. ClassifyMe only requires a laptop computer after it is initially downloaded and installed. The software doesn’t provide near-real-time alerts but regular retrieval of camera trap SD cards by the landholder allows monitoring for the presence of dogs and as such provides another option to stakeholders in wild dog control.

Feral Cat Tactile lure

The Feral Cat Tactile lure taps into the cats behavioral Achilles-heal of being attracted to visual lures. This device is aimed at testing whether a light based lure and a movement lure cause feral cats to engage with the device and as such provide proof-of-concept that automated visual lures are a weakness that warrants further investigation. Prototype trials have been focused largely on whether lights and movement can reliably attract feral cats to the device. Early results show promise and warrant further assessments.

A Feral cat is attracted to the lights of the Feral Cat Lure device

Foxes on the other hand are often warned off by lights and noise as evidenced by this short video

Autonomous Gate Trap and Remote Recognition Response Device

The AGT/RRR integrates automated species detection and identification technology, developed in WDAN, to initiate two different gate mechanisms. It could be integrated into a Boar Buster feral pig trap removing the need for manual gate triggering. Further, the same system could be integrated into water traps for feral goats in the western rangeland. The system uses a PIR to connect with the central computer and is used as an event trigger when an animal presents at the device. The inbuilt computer acquires images, then makes decisions through statistical analysis for when to trigger the gate.

The RRR device is constructed similarly but is specifically designed for integration into dam exclusion networks in the WA rangelands to exclude wild dog access to water. This gate will be fitted into exclusion fence lines and will open when livestock approach the fence line but will remain closed when stock are absent or when wild dogs approach.

Video 1: The RRR device detects sheep moving past and the gate automatically opens so they can  access dam water.

Video 2: The RRR device detects a dog moving past and the gate automatically shuts to prevent the dogs access to dam water.

Video 3:  Proof-of-concept trials testing the ‘machine learning gate closing pig trap’. Safety structures in place to protect pigs from injury.

Sentinel bait station.

The Sentinel Bait station is a new device that was designed and built to overcome the short-comings of ground baiting methods in regards to bait availability. A critical element of efficient baiting is to ensure baits are palatable, toxic, attractive and available. However under some situations these requirements can’t be met continuously throughout the program. The sentinel bait station aims to store baits on site to maintain their attractiveness, palatability and toxicity with in a secure chamber, and deliver a bait when one is removed by a dog, thus ensuring that a bait is always available to another visiting dog. The SBS is designed to detect the presence of a dog and removal of a bait and present another toxic bait within a pre-programmed time.

Video 1 : The sentinel bait station is connected to the wild dog alert node which send a message to the dispenser when the dog is detected

Video 2: The dispenser responds by opening and presenting the bait to the dog.

Sentinel-Bait-Station

The proto-type sentinel bait station

Sentinel-Bait-Station

The sentinel bait station uses a WIFI connection to detect a wild dog using the wild dog alert node.

Project collaborators include:

  • Australian Government
    • Department of Agriculture
  • NSW Government
    • Department of Primary Industries
    • Local Land Services
    • National Parks and Wildlife Service
  • WA Government
    • Department of Primary Industries and Regional Development
  • University of New England
  • Meat and Livestock Australia
  • Australian Wool Innovation
  • Hunter Land Management
  • Pest Lures

Ecological staff include:

  • Dr Paul Meek (Project Leader)
  • Dr Guy Ballard
  • Dr Peter Fleming
  • Mr Heath Milne
  • Dr Karl Vernes
  • Dr Malcolm Kennedy

Technology staff include:

  • Dr Greg Falzon (Principal Computational Scientist)
  • Dr Ehsan K Oshtorjani
  • Dr Saleh Shahinfar
  • Dr Edmund J Sadgrove
  • Dr Robert Farrell
  • Dr Beau E Johnston
  • Mr Derek Schneider
  • Mr James C. Bishop
  • Mr Andrew Shepley
  • Mr Glen Charlton
  • Mr Joshua Stover
  • Mr Amos M Munezero
  • Mr Elrond Ka-Wai Cheung
  • Mr David Luckey
  • Mr Christopher K Lawson
  • Mr Jaimen Williamson
  • Ms Atalya T Mather-Zardain
  • Mr Ryan Layland