With our population and boating registrations increasing, we’re turning to science to give us a clearer picture of activity on our waterways. This will help us improve planning, waterways management and safety.
We’ve kicked off some research using smart cameras to identify the number and types of vessels using our waterways.
Researchers from the Queensland University of Technology (QUT) will combine the images with information on marine incidents and weather conditions to establish trends and provide insight into future patterns of use.
Q What research is being undertaken for Gold Coast Waterways Authority (GCWA)?
We are using smart camera technology and data science methods to build a more detailed picture of how our waterways are being used to improve planning, network management and safety.
The research will use advanced machine learning and statistical methods to estimate the number and types of vessels using the waterways. In combination with information on marine incidents and weather conditions, this will establish usage trends and provide insight into future patterns of waterways use and potential pressure points. The researchers will use machine learning to train computers to differentiate between types of vessels, for example, to separately identify personal watercraft from tinnies.
A second component of the project is exploring the feasibility of using the same camera technology to develop a more accurate way of measuring vessel speed on the water.
Q Why is the research being done?
This project builds on a study we completed last year into users and uses on our waterways.
The Gold Coast is experiencing rapid population growth and increasing boat registrations. Our ‘Users and Uses’ study shows boating traffic in the northern Gold Coast waterways region (the area between the Logan River and the Broadwater north of the Seaway) is expected to triple in line with population increases in the next 25 years
We need to ensure our planning keeps pace with this growth to ensure locals and visitors can continue to access and enjoy our beautiful waterways.
To properly plan we need access to data with integrity and science behind it, which is why we’ve engaged with the team at the Queensland University of Technology (QUT.)
Q Who is conducting the research?
The research is being undertaken by academics from QUT’s Science and Engineering Faculty. The team includes specialists in statistics, machine learning, coastal management, engineering and the built environment.
Q When is it being undertaken?
The cameras will start taking pictures on the Australia Day long weekend and will be in place for about two months. We’ve chosen to start the trial on the long weekend because this is traditionally one of the busiest weekends on the waterways. The researchers will download pictures from the cameras at regular intervals during the trial so they can start the process of interpreting the data.
Q How many cameras will be in place and where will they be?
20 cameras will be installed across the waterways network including locations in the Southern Broadwater and the Coomera River.
We’re using a mix of public and private properties to host the technology.
We won’t disclose the exact locations for security reasons.
Q Doesn’t this technology already exist?
As far as GCWA is aware, there is no camera technology in widespread use that can accurately, autonomously and remotely measure on-water speeds, taking into account factors like strength and direction of currents and the directions vessels are travelling in, as well as identifying registration numbers of vessels.
Q Why can’t you measure speed using a radar gun?
Measuring speed on-water is not the same as measuring point-to-point speed on land because a number of other factors need to be considered.
This includes strength and direction of water currents, the direction a boat is travelling in, prevailing weather conditions such as wind and the distance of the vessel from the camera.
The research work will involve taking those factors into account to develop a method for measuring on-water speed. For example, we’ll use current sensors placed at several locations to measure current strength and direction. We’ll be using advanced machine learning to train computers to recognise registration numbers on vessels.
Q What will GCWA do if photos capture people doing the wrong thing on the water?
The information is being gathered for research purposes, not for law enforcement or compliance. The images collected will only be used for the intended purposes of the project.
Q Will I be fined?
The information is being gathered for research purposes, not for law enforcement or compliance. The images collected will only be used for the intended purposes of the project.
Q What about in the future?
We don’t even know if we can accurately measure speed at this stage, this is why we’re undertaking the research.
If we do come up with a method that works we’d have to do a lot of field testing with other agencies like the water police and Maritime Safety Queensland before it was put to use.
Q How does the camera technology work?
Each camera will take high frequency images of the waterways. From these images, data will be extracted to determine type and number of users on the waterways. Data from consecutive images will be used in estimating vessel speed.
Q What will you be doing with the images?
The images of the waterways will form the majority of the data for the study. This data will be input variables into algorithms, which will be developed to automatically identify vessel type and estimate vessel speed. To train these algorithms, known vessel types in images and known speeds of vessels will be used. To capture this additional data, a proportion of images will be visually inspected to manually identify vessels in images. For estimating vessel speed, we will engage with the Queensland Police Service to capture images of police vessels at various speeds throughout the project.
The photos taken by the cameras will only be assessed by the QUT research team.
Q How will machine learning be used in this project?
Machine learning methods will be used to determine what image data will be most useful for determining vessel type and speed. Such methods should also be useful in developing a predictive model.
Q Where else has machine learning effectively been used?
Machine learning is used in many applications including in your day-to-day life, such as in virtual personal assistants like Siri, video surveillance, online advertising, detecting spam emails and more.
Q What about privacy issues?
GCWA has signed a formal agreement with QUT. The photos taken by the cameras will only be assessed by the QUT research team.
Q Will GCWA be acting like Big Brother on the water?
This is a research project, not a law enforcement or compliance exercise.
We’re using the technology to estimate traffic movements so that we can improve our planning for the future use of the waterways network.
Q What was the Users and Uses study?
The Assessment of Congestion and Conflicting Use Management for the Gold Coast Waterways also known as the ‘Users and Uses’ study was completed in 2019.
This initial study was the first serious attempt to understand past and present patterns of use on the Gold Coast’s waterways and pinpoint areas where access could come under pressure.
The assessment looked at past and present uses of the waterways and drew on publicly available information sources such as:
- Geographic Information Systems (GIS) such as the City of Gold Coast’s data portal, Australian Bureau of Statistics and Geoscience Australia
- GCWA reports and projects – for example the NCEconomics report into the social and economic value of the waterways and our report into the Coomera River South Branch speed limits trial
- online social media portals – such as geotagged Flickr photos, geotagged Tweets, MapMyFitness, and the STRAVA fitness tracking app
- online Automatic Identification System (AIS) vessel tracking data
- online high-resolution satellite imagery and navigation charts.
Q How much is the project costing?
We’re investing $230,000 in this trial. The funding is coming from our $27.6 million three year rolling Waterways Management Program.
