A computer vision scientist says the resources sector could benefit from his startup’s deep learning technology.
A computer vision scientist says the resources sector could benefit from his startup’s deep learning technology.
The founder of geospatial software startup Mapizy intends to use his award-winning tech to help cut costs and improve working conditions for resources companies.
Having emerged from the School of Computer Science at the University of Western Australia, Mapizy extracts image data from sources such as satellites, drones, and underwater vehicles to automate the presentation of changes to natural and built environments.
Mapizy has used its ‘deep learning technology’ in forestry, agricultural, urban infrastructure and mining projects to identify weeds, measure forest growth, assess the risk to buildings, and monitor mine rehabilitation.
Last month, the company and local startup Camp Connect were awarded three months of co-working space at CORE Innovation Hub in Perth, after proposing their technology solutions at the Pitch @ Core event.
Pitch @ Core serves to highlight technology solutions for the resources industry and provide startups with the added benefits of networking, profiling and promotion.
Mapizy founder and chief executive Mehdi Ravanbakhsh said the company would take advantage of the networking opportunities at Core over the next few months to increase its engagement with the resources sector.
“There is great potential for our technology in oil and gas,” Mr Ravanbakhsh told Business News.
“For example, underwater asset inspection is a very risky task and normally a diver needs to go and inspect the asset.
“We can send underwater vehicles like a remote-operated vehicle to capture videos and then use these videos and artificial intelligence to find anomalies in assets.”
Mr Ravanbakhsh said despite the opportunities available through the Core award, approaching big oil and gas businesses would be challenging.
“Normally, it’s an environment where people are busy and under high pressure to deliver and sometimes they do not have time to explore other technologies,” he said.
“We have to approach them and explain the efficiency and productivity they would gain if they use our technology.”
Established in 2014, Mapizy has a team of four based at the UWA Crawley campus, which Mr Ravanbakhsh said was part of an agreement with the university.
He and the team provide teaching support, and PhD and course supervision in exchange for office space.
Mr Ravanbakhsh said he established Mapizy while working as an associate research professor at the university.
One of his projects, which involved the use of underwater stereo videos to estimate fish abundance and size, received a technology award from the UK Society of Photogrammetry and Remote Sensing.
In 2016, he and the team participated in UWA IQ Start Something, targeted at researchers and postgraduate students, which helped the startup validate and customise its business model to address market needs.
“That was really a starting point to get more involved in industry activities,” Mr Ravanbakhsh said.
“We have had fantastic organic growth so far.”
Its main ventures started two years ago in a project with WA forestry service provider PF Olsen.
Mapizy was the first Australian company to use drone and satellite image data to count trees by producing a forest inventory – or a tree stocking map – for the number of trees per hectare.
The startup has also worked with sandalwood plantation company Quintis, for which it produced an inventory of sandalwood trees and their health status.
In 2016, Mapizy worked with BHP, combining data from satellite and aerial images to assess the progress of BHP’s mine rehabilitation.
“Mining companies have a commitment to make sure that, after mine closure, the site is restored to the natural environment,” Mr Ravanbakhsh said.
“They need to monitor the rehabilitation of the natural plant fissures over the course of two, three and sometimes even five years.”
Mapizy is also working with Chevron to identify types of road cracks and potholes, and is currently exploring opportunities with Woodside Petroleum’s robotics lab.
“Our technology removes humans from hazardous environments, whether in mine sites or oil and gas plants,” Mr Ravanbakhsh said.
“Human operators are involved in sample data collection, analysis, asset inspection and repairs, which are costly and risky.
“Except for complex repairs, all other tasks can be done through AI and image data at a much lower cost, faster and more frequently.”