Researchers have created a quicker and more precise technique for obtaining and reconstructing high-quality 3D surface measurements. This innovation has the potential to significantly enhance the speed and precision of surface measurements used in various fields such as industrial inspections, medical uses, robotic vision, and beyond.
A team of researchers has introduced a streamlined and accurate way to capture and construct high-quality 3D surface measurements. This new method could greatly enhance both the speed and accuracy of measurements utilized in industrial inspections, medical applications, robotic vision, and more.
“Conventional 3D imaging relies on comparing two different viewpoints, much like how our eyes collaborate to perceive depth,” explained Ce Zhu, the leader of the research team from the University of Electronic Science and Technology of China. “In contrast, our innovative method ‘feels’ the surface by projecting light patterns, akin to running a hand over it to sense changes. This reduces the quantity of patterns needed by more than two-thirds, accelerating the scanning process and surprisingly, yielding even better accuracy than previous methods.”
In Optica, the journal from Optica Publishing Group focusing on high-impact research, the team discusses their new fringe photometric stereo technique, which achieves high frame rates and precision at the micrometer level while cutting the noise variance in half.
“Our technique is perfectly suited for real-time scanning applications, including industrial uses such as identifying flaws in printed circuit boards, batteries, or oil pipelines, as well as medical functions like diagnostics and customizing implants,” Zhu added. “It can also enhance robotics, leading to improved human-robot interactions or vision guidance for tasks like folding laundry.”
Quicker Acquisition Time
High-precision 3D surface measurements and reconstructions are typically performed using a technique known as phase-shifting profilometry, which is a form of fringe projection profilometry. This method involves projecting a series of phase-shifted light patterns onto an object’s surface and capturing the reflected images. The phase differences are then analyzed to create a highly detailed 3D representation of the surface.
However, this technique presents challenges for many applications due to its lengthy scanning time, mainly because it requires a large number of multiple-frequency fringe images to analyze phase differences. Triangulation is then employed to convert the phase data into continuous values, allowing for an accurate depiction of the shape or surface. In their new approach, the researchers found a way to bypass this lengthy process and notably reduce the number of required fringe images by using a single frequency.
To evaluate their new fringe photometric stereo method, the researchers constructed an experimental setup that included a 1280 × 960 camera with an 8-mm lens and a projector of resolution 912 × 1140. They utilized this system to measure various objects and groups with continuous surfaces, such as a human hand, a paper mask, a soft toy, gypsum shapes, and clay crafts. They verified the effectiveness of the method against standard plane and sphere models, showing it successfully mitigates noise compared to traditional fringe projection profilometry techniques.
Enhancing Prosthetic Design
“One area where this new method may be particularly beneficial is in customizing prosthetics,” Zhu stated. “It can swiftly gather high-accuracy surface data from the residual limb, minimizing manual measurement errors and improving the comfort of the prosthetic fit. This method also removes the necessity of applying plaster or other materials directly onto the skin, making the process far more comfortable for patients.”
While the method currently excels at scanning scenes featuring continuous surfaces, it still faces challenges in reconstructing depth for objects with sudden changes in depth. The researchers are actively working on overcoming this challenge by integrating established surface reconstruction methods from photometric stereo into their new technique. This enhancement is expected to broaden the scope of useful applications and analyze more complex scenes.
