Image sensors convert light into electrical signals and measure light intensity, but for machine vision to improve significantly—essential to advances in autonomous vehicles, detection with biomedical imaging, astronomical discoveries, and more—angle, spectrum, and other light aspects must be extracted as well. As image sensors are essentially “millions of pixels on a single chip,” advancement challenges depend on developers’ ability to integrate nanostructured components into image sensors chips.
Researchers from St. Louis’s Washington University, the University of Wisconsin-Madison, and OmniVision Technologies have created an on-chip spectrometer to detect multiple-band spectra. The team created intricate communication between incident light and the sensor by placing a photonic crystal film, composed of silicon, atop the pixels, which then record light energy distribution. From this, the researchers can deduce light spectral information.
Using directional hearing sensors found in animals as inspiration, the researchers built a component that detects angular information to measure depth and construct 3D shapes at subcellular scales. The team’s design uses silicon nanowire pairs as resonators to support optical resonance so that the stored optical energy is sensitive to incidental angle, and by comparing the currents in each wire, the researchers could identify the angle of light waves.
The device, which is smaller than one-one hundredth of a square inch, can hold millions of the nanowires. Programmable to meet various dynamic ranges, resolution levels, and almost any spectral regime, the device has implications for innovation in lensless cameras, augmented reality, and robotic vision.
The study’s co-author Yurui Qu of the University of Wisconsin-Madison explains, “Image sensors will gradually undergo a transition to become the ideal artificial eyes of machines.”