We apply technology to solve environmental issues for a wide range of stakeholders, with a focus on drones and Artificial Intelligence (AI).
Ethical
We strive for positive environmental impact in the way we operate and what we deliver for our clients.
Evidence-based
Our work is rooted in rigorous scientific analysis and is thoroughly documented.
Innovative
We draw from a pool of cross-disciplinary science and research to think outside the box.
Full-circle
We create solutions from start to finish for real-world problems.
Sci-eye is led by an interdisciplinary team of experienced scientists.
Ecology | Sensor Platforms | Statistics
Dr Andrew Colefax has a strong background and 18 years’ experience in animal behaviour, coastal ecology, hardware/software application such as sensors and drones, and methods development. He is a Founding Director at Sci-eye and maintains a large cross-disciplinary network through commercial industry and academia.
Data Science | Machine Learning | Real-World Solutions
Dr Andrew Walsh has extensive experience in, AI, computer vision, data science, hardware/software interfacing, app development, customer experience and is Founding Director at Sci-eye. He has over 30 years’ experience as a professional scientist, with skills in cutting edge research, with a finely tuned knowledge of image processing.
External Technial Advisor
Dr Cormac Purcell is a machine learning engineer and developer with a background in astrophysics. He was a Sci-eye co-founder and helped develop the core technology behind Sci-eye's shark-tracking AI. Cormac is an external technical advisor for Sci-eye, contributing to some of our scientific and machine learning research.
We are working on a diverse range of projects. See below for some current project highlights.
Portable AI in real-time
This AI solution is a mix of both hardware and software, and is designed to provide species-level identification of coastal marine fauna (such as species of dangerous sharks) in real-time without bulky equipment or an internet connection. The ultrabright screen with AI connects to the drone controller and can improve the reliability of drone-based shark spotting and marine fauna surveys.
Sharkie App
The Sharkie App runs under the Android operating system and was designed in collaboration with NSW Department for Primary Industries for use by NSW Surf Life Saving. It has a display that has similar functionality to standard DJI apps, allowing pilots to control and monitor their drone in-flight. The App also contains an AI model that can be used to identify marine fauna in real-time.
DaBox
Dabox is a portable device, designed and built by us, to automatically and securely upload drone footage to a cloud service. The drone pilot inserts the drone’s micro-SD card into DaBox and the video data is automatically transferred overnight, making the micro-SD card flight-ready for the next day - all without further interaction from the pilot. It has an easy-to-use touch screen so the pilot can monitor data transfers.
Wildlife Documentary Filming
Our knowledge of wildlife behaviour, along with our drone and sensor capabilities, provide us with unique capabilities for delivering high-end imagery for blue-chip and standard wildlife documentaries. We enjoy the creativity and challenge of ‘getting in amongst it’ to film wildlife in their natural behaviour without causing disturbance.
Identifying marine fauna with drones and AI
We have spearheaded a major research paper (in peer review) which demonstrated the utility of using AI with drones to help detect and classify marine species. We also empirically investigated how AI can perform on an ‘edge’ device in a beach setting, compared to a ‘desktop’ model, and demonstrated the importance of testing models against completely new datasets.
Review of Drone Technology
In a multi-disciplinary team, we published The Drone Revolution of Shark Science: A Review in "Drones", a scientific peer-reviewed journal. This paper highlights how drones can be used to greatly improve our understanding of shark movement and behaviour. Check out our open-access paper, featuring some of the leading ecologists worldwide.
Methods for Improving Detectability
After publishing Identifying optimal wavelengths to maximise the detection rates of marine fauna from aerial surveys in the scientific journal "Biological Conservation", where we custom mounted a hyperspectral camera on a drone, we are extending this research to empirically test a potential cost-effective real-world application that can be used on consumer drones.
Social Research on Shark Management
Recent research by Dr Andrew Colefax has looked at the interface between strategies used to mitigate shark incidents and the public perception of these strategies. This work shows that the public strongly supports the use of drones as a shark management tool. You can find more details at The Conversation and the journal of Marine Policy.
Ongoing data collection
Whenever we have spare time, we will fly our drones over beaches to collect new footage. Our previous work has shown that the best AI models require training data from many different beaches, under different weather and water conditions. We can also use the new footage to deliver AI models that are dedicated to specific use cases.