Modern robotics has often grappled with detecting transparent and reflective surfaces like glass or shiny metals. Traditional 3D sensors can easily miss the mark by distorting or absorbing light, a real headache when precision and speed are essential.
The Fraunhofer Institute for Applied Optics and Precision Engineering IOF has developed goROBOT3D to tackle this challenge head on. Instead of a lengthy 15‑second scan, this system takes just 1.5 seconds by employing structured thermal imaging. By projecting heat patterns onto surfaces and capturing them with thermal cameras, it creates accurate 3D images regardless of the object’s appearance.
If you’ve ever wrestled with slow or unreliable measurements, you’ll appreciate that goROBOT3D now needs only one pair of images to generate a full 3D model—a vast improvement over previous methods that required hundreds of shots. As Dr Martin Landmann explains, the process involves heating the object in a structured way, then using spatial cross-correlation between two thermal cameras to produce a precise model.
Moreover, the team replaced the outdated fringe projection technique with a rapid, irregular dot pattern generated by diffractive optical elements. This change means the laser beam instantly covers even complex surfaces, making accurate, high-speed measurements possible.
After the scan, an AI system steps in to analyse the object and suggest the best gripping points using a dynamic bin‑picking process. This efficient approach helps production lines run smoothly, reducing downtime and optimising operations.
