On construction sites with blind spots, a construction machine equipped with the ifm O3M AI 2D/3D intelligent sensor system can reverse more quickly. A municipal vehicle can stop reliably, even if the person crossing is difficult to detect. At airports or container terminals, the sensor system guides the positioning of a passenger boarding bridge or a container to within centimetres without risk of collision. The combination of the two technologies virtually eliminates the annoying false triggers that are almost inevitable with conventional 3D systems. The number of false alarms is significantly reduced. This is because the system can detect whether there is a person or a pile of earth in the vehicle’s path. The result is increased efficiency in the work process.
Improved results in a radius up to 25 meters
Such is achieved through the unique combination of a 3D PMD sensor and 2D AI camera in an embedded area monitoring system. System reliability is greatly enhanced by fusing data from two different types of sensors on a fully integrated and calibrated platform. AI analysis of the 2D camera data improves the quality of object detection, enabling it to distinguish between people and other obstacles. The new collision warning system works much better and more accurately than conventional solutions. The O3M AI smart sensor system uses the data from the 2D and 3D cameras to generate a video that displays the surroundings and simultaneously overlays information on the exact distances to all objects and people within a radius of up to 25 metres onto the image. The O3M AI can always tell whether emergency braking, a controlled stop or deceleration would be appropriate in a given situation.
The high-resolution digital Ethernet camera is equipped with a powerful processor with AI accelerator (NPU). This enables ultra-fast processing of the AI algorithms for person detection directly in the system. In addition to the Ethernet camera, a 3D PMD camera is integrated. It provides an accurate distance measurement to the object in every single pixel based on the PMD time-of-flight principle. Depth mapping from the 3D system and video from the Ethernet camera can be combined or used on their own, depending on the application.
Trained for industrial environments
The O3M AI’s person and obstacle detection has been optimised for mobile machines. The collision warning system can accurately distinguish between a distant obstacle and a person in close proximity, even if that person is lying on the ground, is wearing dark clothing or is difficult for conventional systems to detect due to larger tools or unusual posture. It works in both bright sunlight and twilight and has a maximum range of 25 metres with an accuracy of 10 centimetres.
The AI, hardware and software have been developed in Germany by ifm. A wide variety of people, equipment and work positions from the typical mobile machine work environment are included in the AI training examples. The result is a far greater gradation of possible alarms and warnings from the system. When it comes to striking the right balance between efficiency and safety, this is an important criterion.
The new combined sensor system for mobile machines is available in several versions.
