The first ADAS detecting people by means of reflective workwear
Advanced Driver Assistance System with Embedded Vision increases occupational safety
The Advanced Driver Assistance System (ADAS) from Retenua detects people by means of reflective workwear. Thanks to embedded machine vision it is particularly compact and fail-safe.
Where people and heavy equipment operate in the same area, accidents still cannot be avoided, unfortunately. In mining, construction sites, logistics and many other places and situations, drivers navigate machines with huge blind spots. Retenua presents the first fully integrated system that can be mounted on the machines themselves, detecting reflective workwear by using an infrared camera.
Vision Components contributes the intelligent stereo camera. This embedded vision system with a powerful integrated processor both captures the images and processes them. This renders additional peripheral devices for machine vision tasks unnecessary. Also, it makes it an extremely robust solution mechanically and functionally.
Retenua holds the patent for a technology detecting retroreflectors, the reflecting stripes on high-vis vests and protective clothing. Based on this technology the company has developed the advanced driver assistance system emitrace, consisting of a sensor and CPU that is typically mounted on the rear of industrial vehicles. The operator's cab is equipped with an opto-acoustic signal device, more devices are not necessary. The integrated embedded vision system detects reflectors, calculates the 3D position and speed of the persons based on the stereo captures and triggers an alarm, if an accident threatens. Additionally, there is an option, to connect emitrace directly with the vehicle control system, to allow for automated brake activation. The powerful NIR flash (near infrared) allows for short exposure times with only minimal infrared radiation for the environment. The infrared flash is not disturbing persons, but ensures a stable detection performance under any lighting condition instead.
Rafael Mosberger and Henrik Andreasson started on the scientific side: They were both researching person detection with industrial vehicles at the Örebro University in Sweden. It was also there, they developed their innovative approach for the detection of retroreflectors be means of an infrared camera. Besides EU subsidies they were also funded by companies from the industry sector, which lead the two scientists to found Retenua. The start-up provided some vehicle manufacturers with an early emitrace version as a Developer's Kit. The lively exchange with the engineering departments was used to determine the demands and to develop a standard plug&play product with preconfigurated parameters.
The product now availalbe is designed for a simple retrofit on any mobile machine and for maintenance free operation. It is also successfully field-tested. Retenua's C++ vision library and the API provided, allow for a flexible integration in different software environments. With that, raw data and image processing results can also be transferred to other systems for further processing or documentation.
emitrace, the advanced driver-assistance system, is fully integrated - this means, it renders any other divece for data processing unnecessary. The developers were looking specifically for embedded vision systems for real-time stereo imaging and found the perfect fit at Vision Components'. "Automotive industry and its related sectors prefer compact sensors, that can quickly and easily be installed", explains Rafael Mosberger, CEO of Retenua.
The intelligent VC Z series stereo camera fulfills all requirements: As a board camera with two remote sensor heads, it is easy to integrate and provides a Xilinx® Zynq® SoC with a dual core ARM® processor taking care of the complete machine vision data processing. A flash trigger and a 1 Gbit Ethernet interface make the optimal connection.
With a Linux operating system and the VCLib software library the camera offers OEM ideal conditions, as Rafael Mosberger acknowledges: "This is a huge advantage for, because we were able start right away. For sensor configuration and image readout we used the functions of the included software components VCLinux and VCLib. At the moment, our programm is executed completely on the ARM processor. We now investigate with Vision Components, how we can accelerate the algorithms by relocating certain computationally intensive tasks on the FPGA. This will make the next generation emitrace."
The FPGA is part of the Zynq SoC modules, the core of VC Embedded Cameras.This module enables faster processing rates and is freely programmable with a hardware discription language.
Retenua AB is a product development and tech consulting company from Örebro, Sweden.
With a strong technological background in sensors, computer vision, machine learning, AI and robotics, Retenua is striving towards the most efficient technology transfer from state-of-the-art research to real-world industrial machine vision and robotics applications. Founded in 2015, as a result from a research project on person detection with industrial vehicles, the company moved 2019 in a new building. http://www.retenua.com