Person detection
Driver assistance system reacts to reflective clothing
The Swedish start-up Retenua has developed the first fully integrated driver assistance system for detecting people on mobile machines and forklift trucks based on an embedded camera from Vision Components. It is trained on reflective warning clothing and uses an NIR flash.
The patented process offers reliable detection both in the dark and under blinding sunlight and considerably increases occupational safety. The development was carried out together with companies from the mining and logistics sectors. Other conceivable target applications include use on landfill sites and all other areas where the viewing angle for vehicle drivers is restricted and people are equipped with high-visibility clothing.
The embedded design makes the driver assistance system compact and robust. The integrated VC-Z stereo camera from Vision Components detects reflectors, calculates the 3D position and speed of people from the images and triggers an alarm if there is a risk of an accident. The board camera with two remote sensor boards offers a flash trigger and a 1 Gbit Ethernet interface as well as a Xilinx Zynq SoC, which carries out all the image data processing. The Linux operating system and the program libraries provided by Vision Components made software development easier for Retenua.
According to the company, the innovative solution, which Retenua markets throughout Europe under the name Emitrace, has advantages over existing sensor and camera systems, which each have different weaknesses: Lidar, radar and ultrasound are often unable to reliably distinguish between people and objects - many false alarms disqualify them for industrial environments. RFID technology is better suited in this respect, but the additional equipment required with RFID transponders and readers can result in considerable costs, depending on the number of devices. Current camera solutions detect people based on their silhouette and only work reliably when they are standing upright. The algorithms are still a long way from being able to recognize an unconscious, lying person in real time, for example. Emitrace, which recognizes the reflectors on work clothing, is the only system that works perfectly regardless of lighting conditions and body posture. In contrast to existing camera-based products, Emitrace does not require an additional image processing PC. It is therefore also particularly low-maintenance and cost-efficient. as








