Computer Vision News - April 2021

Drones share the same airspace as crewed aircraft, so Iris Automation works with regulators around the world to make them safer and more accessible as airspace regulation continues to evolve. Detect-and-avoid technology is a vital part of this. “We add eyes and brains to drones,” he explains . “While they are flying by themselves, we will keep them safe, looking around to make sure there is nothing else in the same airspace, and preserving the required safety buffer. If there is something in the vicinity, we can classify what it is and get out of its way.” These capabilities bring a host of new possibilities for drones, such as package delivery and railway inspection. They can support surface mining, precision agriculture, and search and rescue. Potential economic savings are huge, but they can also save lives , both indirectly and directly. A number of people every year lose their lives during power-line inspections, for example, 250 Computer Vision Application which is something drones can perform autonomously. The team has extensively tested the technology on more than 12,000 real- world encounters , but this is costly and not without risk. To complement this, they have built their own simulator which has created more than 50,000 synthetic encounter scenarios , allowing them to test very dangerous events like head-on collisions. “This is not just your vanilla simulator,” Alejandro tells us proudly. “Our Simulation Architect worked for NASA and Industrial Light & Magic in the movie industry, amongst other places, and he developed a simulator that can model all the things that affect computer vision, such as lens distortion, noise, and illumination . The simulator can try to understand what a scenario will look like and how the system would sense what is happening. We’ve been developing this for four years and now we’re at the point where we can generate data at a large scale .”