Seoul Robotics, a Korean-based software company, has launched its first 3D perception software that utilizes deep learning to detect fast-moving, clustered, and partly visible objects accurately.
Deep learning is pushing the boundaries of possibility in the LiDAR industry, enabling perception accuracy that has never been seen before. While other 3D computer vision software relies on machine learning and rule-based systems, Seoul Robotics now uses deep learning to track more than 500 objects concurrently and with an accuracy range of within 10 cm—something that rule-based systems have yet to achieve. Weather-filtering AI is also included in SENSR 2.2, allowing the software to track and detect objects even in severe weather conditions such as heavy rain and snow.
SENSR 2.2 is sensor-independent and works with over 75 different types of 3D sensors on the market today, including LiDAR, 3D cameras, and imaging radar. Smart intersections, wrong-way detection, speeding, smart railroad crossings, crowd management, and smart retail all benefit from SENSR 2.2’s increased accuracy.
Seoul Robotics is rapidly expanding globally and currently partners with several top-tier organizations, including BMW, Mercedes-Benz, the Chattanooga Department of Transportation, Emart, and many others. Earlier, Seoul Robotics partnered with Israel-based Hyper-Technology to deploy 3D vision for autonomous vehicles. It has also launched the robotic LiDAR perception system called Voyage.
According to HanBin Lee, CEO of Seoul Robotics, the incorporation of deep learning into 3D perception software may be one of the last game-changers in the LiDAR industry. Historically, the emphasis has been on improving LiDAR sensors, but this is changing.
In the future, significant investments will be made in 3D perception software that interprets data and converts it into actionable solutions. He went on to say that the introduction of SENSR 2.2 is hastening the adoption of solutions that will fuel autonomy around the world.
According to Kevin Comstock, the City of Chattanooga’s smart city director, Seoul Robotics has specifically assisted the City of Chattanooga in seamlessly monitoring pedestrian traffic, and they are currently gathering data that will inform future capabilities of wrong-way detection. These efforts save the city money, residents’ travel time, and, most importantly, lives.
SENSR 2.2 will be on display at the IAA Mobility Conference in Munich, Germany, from September 7 to 12, 2021.