Motivation for New Image Sensor Technologies:
While conventional CMOS detectors for visible light are well-established and somewhat widespread, at least for low-value applications, there is a significant opportunity for more complex image sensing hardware that offers capabilities beyond simply acquiring red, green, and blue (RGB) intensity values for each pixel. This is largely due to the increasing adoption of machine vision, where image analysis is performed using computer algorithms. Machine learning requires as much input data as possible to establish correlations that facilitate object identification and classification, so acquiring optical information across different wavelength ranges, or with spectral resolution, for example, is highly advantageous.
Of course, new image sensor technologies offer many other advantages. For example, similar capabilities at a lower cost, greater dynamic range, improved temporal resolution, spatially variable sensitivity, high-resolution global shuttering, reduced unwanted dispersion, flexibility/conformability, etc.
The new IDTechEx report, "Emerging Image Sensor Technologies 2021-2031: Applications and Markets", covers many emerging technologies, specifically
: Quantum dots on silicon hybrid image sensors
; Organic photodetectors on silicon hybrid image sensors
; Emerging SWIR image sensor technologies, including extended-range silicon;
Organic and perovskite photodiodes (OPDs and PPDs)
; Event-based vision
; Hyperspectral imaging
; Flexible X-ray image sensors
; and Wavefront imaging.
The emerging image sensor technologies included in the report are as follows:
Hybrid image sensors: adding an extra light-absorbing layer on top of a CMOS readout circuit is a hybrid approach that uses organic semiconductors or quantum dots to increase spectral sensitivity in the SWIR region. This new technology, currently dominated by expensive InGaAs sensors, promises a significant price reduction and, therefore, the adoption of SWIR imaging for new applications such as autonomous vehicles.
Extended-range silicon: Given the high price of InGaAs sensors, there is considerable incentive to develop much lower-cost alternatives that can detect light toward the lower end of the SWIR spectral region. These SWIR sensors could be used in vehicles to improve visibility through fog and dust by reducing scattering.
Thin-film photodetectors: Detecting light over a wide area, rather than in a single small detector, is highly desirable for acquiring biometric data and, if flexible, for imaging through the skin. New approaches using solution-processable semiconductors offer a compelling way to produce large-area conformal photodetectors for applications such as under-display fingerprint detection.
Event-based vision – Autonomous vehicles, drones, and high-speed industrial applications require image detection with high temporal resolution. However, with conventional frame-based imaging, high temporal resolution produces large amounts of data that require computationally intensive processing. Event-based vision is a new way of acquiring optical information that combines higher temporal resolution of rapidly changing image regions with significantly reduced data transfer and post-processing requirements.
Summary
In summary, the increasing adoption of machine vision presents a great opportunity for new image sensing technologies that offer capabilities beyond conventional CMOS sensors.
