according to Daan Arscottdata collection reader Nirickson, “Machine learning makes it more efficient and cost-effective compared to traditional inspection processes.” Based on experience working on projects involving: He’s seen first-hand how machine learning is doing drone-based infrastructure health assessments.More Accurate, Efficient and Safer.
“Machine learning has some features that make it a great tool for inspection,” Alscott told Commercial UAV News. “One is to look unbiased at the data we are collecting. That is, if we can see the surface and the image quality and resolution are good enough, we can identify and quantify all defects in an unbiased way. It means it can be transformed.”
Machine learning automation will also enhance drone inspection work. “These systems take us one step further than just digitizing assets,” Arscott explained. “Obviously, it’s nice to have a 3D model of the structure, but one day someone will have to manually pour over the entire structure and identify all the cracks and imperfections in order to do a proper engineering evaluation. Nirickson will automate this process for faster and higher quality evaluations.”
Nirickson has been working on drone data collection and inspection projects in the United States, Australia and New Zealand in recent years, as well as in its home country of Canada. “We see ourselves as a software company that specializes in automatically identifying and quantifying defects in concrete assets using machine learning and AI,” Arscott said. rice field. “We are focused on being a software company, but we invest heavily in data collection because machine learning and AI are all but devalued without quality data.”
Regarding this research, Arscott emphasized the importance of Niricson’s unique acoustic acoustic payload. “We are not only simulating the engineer’s eyes through a camera mounted on the drone, but we are also simulating the engineer’s ears through an in-house developed acoustic payload,” he said. “By simulating the traditional concrete hammer test, we can physically hit structures with the help of drones to identify concrete delamination.”
In his role at Nirickson, Arscott helps coordinate the various systems and personnel required to inspect and monitor these large assets. “I serve between operations his pipeline, which processes data and uploads it to his platform in the cloud for customers, and the sales team, which is involved in business development and contracts,” he said.
A key element of Arscott’s work involves working with drone pilots. “I work directly with my clients to hire third-party pilots to collect data,” he explained. “This allows us to be global in nature, so we don’t need to have service teams around the world. Instead, we leverage relationships and partnerships with third-party providers to ensure high-quality data. We maintain collections and execute projects quickly, although in some cases our clients have in-house drone teams that we directly support and train in data collection methodologies. This will allow asset owners to use their own drones to collect their own data, and Nirickson will move strictly to software solutions in the future.”
Alscott argued that finding “reliable and qualified pilots” for these projects is key. “We collect very high-resolution images so that we can see all these microscopic defects,” he explained. “We create very complex flight plans to collect this data autonomously. I go to the field, spend a week with the pilots, share my knowledge and do some QA/QC work on site before handing the data over to the processing team, and then my internal team builds the 3D model. to perform machine learning and upload AI-generated defect maps to our cloud platform.”
Alscott reports that Nirickson has worked on more than 50 drone-based data collection projects over the past three years. The effort has enabled the company to “create a baseline for large infrastructure assets with associated quantifiable data, so that each year we can return to the site and rescan the entire structure.” he said. “We can then compare the two datasets and monitor changes over time. Accurate change detection is the real added value of our solution.”
Nirickson’s experience with these advanced technologies drives growth. “2023 is going to be a good year for us,” Arscott declared. “We have already completed several projects in North America this year and have just completed several more in Australia and New Zealand, working on some of the largest, most iconic and most complex infrastructure assets in the world. .”
Click here to learn more about Nirickson.
