Many insects have compound eyes, which provide them with a wider dynamic field of vision, enhancing their ability to evade enemies and capture prey. Inspired by the visual structure of insect eyes, South Korean scientists have developed a new biomimetic camera with ultra-high frame rates and sensitivity.
Traditional high-speed cameras excel at capturing fast movements, but their sensitivity decreases as frame rates increase. This is because as shutter speed increases, the amount of light collected decreases, making it challenging to obtain high-quality images in high-speed and low-light conditions.
However, scientists found that the operation principle of insect compound eyes may solve these issues. Compound eyes, composed of numerous small eyes, can synchronously detect fast-moving targets to enhance visual sensitivity. By integrating received signals, insects can observe targets in low-light conditions, effectively foraging, navigating, and evading threats.
Researchers at the Korea Advanced Institute of Science and Technology (KAIST) were inspired by these biological mechanisms to imitate the visual structure of insect eyes, successfully developing a low-cost high-speed biomimetic camera that overcomes the limitations of frame rates and sensitivity faced by conventional high-speed cameras. The study was published in the international journal “Science” in January 2025.
Using a structure similar to insect compound eyes, the researchers employed multiple tiny optical channels for signal reception, capturing signals at different times and combining them to create a larger signal. This approach allows rapid overlap of light frames, accumulating them into a cohesive image to reduce signal noise and improve image quality.
Through this method, they created a high-speed high-sensitivity MicroLens Array Camera (HS-MAC) capable of low noise, capturing 9,120 frames per second (fps), and providing clear images in low-light conditions.
The HS-MAC camera features a single rolling shutter with a complementary metal-oxide semiconductor (CMOS) image sensor measuring 10.5 mm long, 8.35 mm wide, and 1.53 mm thick. The CMOS image sensor includes over 1000 tiny lenses forming a lens array (oMLA).
By utilizing channel division technology, the CMOS image sensor rapidly exposes in sequence to enhance the overall frame rate of the camera in low light. Once the camera captures these high frame rate images, an algorithm for compressed image recovery can eliminate image blur caused by frame rate accumulation, ensuring clear images for viewing and usage.
In a comparison with a traditional high frame rate MicroLens Array Camera (MAC), the HS-MAC camera achieved a maximum frame rate of 9,120 fps, 19 times higher than the MAC’s highest frame rate of 480 fps, and exhibited superior image quality and clarity.
Experimenters tested the HS-MAC camera by filming the swaying pattern of a low-light (0.88 lumens) flame. Results showed the camera could accurately and clearly capture the flame’s swaying rhythm at a distance of 20 cm, even capturing the rapid changes as the flame was snuffed out.
Furthermore, the research team plans to expand this technology by developing advanced image processing algorithms for 3D imaging and super-resolution imaging applications, aiming to apply them in biomedical imaging, smart devices, drones, cars, and other image-related uses.
The experimenters stated they will continue to enhance the performance of the HS-MAC camera by improving optical design, computational correction, addressing image distortion issues commonly encountered in close-up shots, and reducing the computational cost of compressed image recovery to achieve clearer and higher resolution images.
Hyun-Kyung Kim, a doctoral student in Bio and Brain Engineering at the Korea Advanced Institute of Science and Technology and the first author of the study, expressed to the university’s newsroom, “Through experiments, we have confirmed that this insect compound eye-inspired camera can capture clear photos in high-speed and low-light conditions. Its compact size and convenience make it suitable for various applications such as surveillance and medical imaging.”
This research received support from the Defense Acquisition Program Administration (DAPA), the Defense Technology Planning and Development Institute (KRIT) of South Korea, the Ministry of Science and ICT, and the Ministry of Trade, Industry, and Energy (MOTIE).