Low-light Imaging Applications Research Articles

Progressive Joint Low-Light Enhancement and Noise Removal for Raw Images.

Low-light imaging on mobile devices is typically challenging due to insufficient incident light coming through the relatively small aperture, resulting in low image quality. Most of the previous works on low-light imaging focus either only on a single task such as illumination adjustment, color enhancement, or noise removal; or on a joint illumination adjustment and denoising task that heavily relies on short-long exposure image pairs from specific camera models. These approaches are less practical and generalizable in real-world settings where camera-specific joint enhancement and restoration is required. In this paper, we propose a low-light imaging framework that performs joint illumination adjustment, color enhancement, and denoising to tackle this problem. Considering the difficulty in model-specific data collection and the ultra-high definition of the captured images, we design two branches: a coefficient estimation branch and a joint operation branch. The coefficient estimation branch works in a low-resolution space and predicts the coefficients for enhancement via bilateral learning, whereas the joint operation branch works in a full-resolution space and progressively performs joint enhancement and denoising. In contrast to existing methods, our framework does not need to recollect massive data when adapted to another camera model, which significantly reduces the efforts required to fine-tune our approach for practical usage. Through extensive experiments, we demonstrate its great potential in real-world low-light imaging applications.

Hot pixel classification of single‐photon avalanche diode detector arrays using a log‐normal statistical distribution

CMOS single-photon avalanche diode (SPAD) detector arrays are commonly used in low-light imaging applications, and are known to suffer from certain defects which cause `hot pixels'. These are detectors which exhibit a significantly larger than average dark count rate, adding noise to the data. Typically, data from these detectors are removed by post-processing or the detectors can, in some cases, be switched off prior to measurement. Users must define and identify these hot pixels, however there exists no consistent methodology of doing so. The authors present a self-consistent method of defining a hot pixel by fitting a log-normal distribution to a histogrammed dark-count map of the array. The approach has proven a robust method of classifying hot pixels in a number of different detector arrays, providing a threshold based on statistical analysis rather than human intuition. This definition provides a reliable and standardised figure of merit, facilitating a more accurate comparison between different single-photon detector arrays.

Highly Sensitive Graphene–Semiconducting Polymer Hybrid Photodetectors with Millisecond Response Time

Graphene–semiconducting light absorber hybrid photodetectors have attracted increasing attention because of their ultrahigh photoconductive gain and superior sensitivity. However, most graphene-based hybrid photodetectors reported previously have shown a relatively long response time (on the order of seconds) caused by numerous long-lived traps in these hybrid systems, which greatly restricts device speed. In this work, graphene–thieno[3,4-b]thiophene/benzodithiophene polymer hybrid photodetectors fabricated on self-assembled-monolayer (SAM)-functionalized SiO2 substrates are demonstrated with a maximum responsivity of ∼1.8 × 105 A W–1 and a relatively short photocurrent response time of ∼7.8 ms. The fast and highly sensitive device characteristics provide great potential in low-light imaging applications. The hybrid photodetector on the SAM-coated SiO2 substrate shows better performance in responsivities and response times as compared with those of the device on the bare SiO2 substrate. The improved resp...

Measurement and modeling of microlenses fabricated on single-photon avalanche diode arrays for fill factor recovery

Single-photon avalanche diode (SPAD) imagers typically have a relatively low fill factor, i.e. a low proportion of the pixel's surface is light sensitive, due to in-pixel circuitry. We present a microlens array fabricated on a 128×128 single-photon avalanche diode (SPAD) imager to enhance its sensitivity. The benefits and limitations of these light concentrators are studied for low light imaging applications. We present a new simulation software that can be used to simulate microlenses' performance under different conditions and a new non-destructive contact-less method to estimate the height of the microlenses. Results of experiments and simulations are in good agreement, indicating that a gain >10 can be achieved for this particular sensor.

Column-Parallel Digital Correlated Multiple Sampling for Low-Noise CMOS Image Sensors

This paper presents a low-noise CMOS image sensor using column-parallel high-gain signal readout and digital correlated multiple sampling (CMS). The sensor used is a conventional 4T active pixel with a pinned-photodiode as detector. The test sensor has been fabricated in a 0.18 μm CMOS image sensor process from TSMC. The random noise from the pixel readout chain is reduced in two stages, first using a high gain column parallel amplifier and second by using the digital CMS technique. The dark random noise measurement results show that the proposed column-parallel circuits with digital CMS technique is able to achieve 127 μ <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">V</i> <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">rms</sub> input referred noise. The significant reduction in the sensor read noise enhances the sensor's signal-to-noise ratio (SNR) with 10.4 dB. Such sensors are very attractive for low-light imaging applications which demand high SNR values.

Improved Single-Molecule Imaging Based on Photon Counting with an EMCCD Camera

New breakthroughs in digital imaging have historically opened opportunities to scientists in bio-medical research. One important landmark of the last decade has been the development of the electron-multiplying charge coupled device (EMCCD), an imaging detector that uses an avalanche phenomenon to amplify the weak photo-electron signals. EMCCDs are well suited for ultralow light applications in that a single photo-electron can be amplified sufficiently to allow the possibility of photon counting. However, as with all electronic detectors, the noise increases greatly with an increase in gain, diminishing the overall sensitivity and dynamic range and thus compromising images.Recently, a new camera has been developed by Nuvu Cameras that significantly reduces the noise generated during the read-out process (up to 10 times) and allows more efficient photon counting. This major reduction of the noise threshold represents an opportunity for various low-light imaging applications including single-molecule studies.We present a demonstration of its applicability in a total internal reflection fluorescence microscopy (TIRFM) assay to visualize single biomolecules. In molecular biology, it is well established that most proteins exert their biological roles in complex with other proteins. Thus, detecting individual components and the structure of these complexes is essential to understand the mechanism of biological events. To this aim, we employ TIRFM imaged in photon counting mode for visualization of single centromeric nucleosomes, containing a histone variant known as CENP-A. We show that this technique is well suited to TIRFM and that it is capable of detecting single fluorescently labelled CENP-A in the nucleosome with an enhanced signal to noise ratio. The desired and improved properties of the assay such as high sensitivity and photon counting shall pave the way for comprehensive studies of single molecules in this and other biological contexts.

An electron bombardment CCD-camera as detection system for an echelle spectrometer

The electron bombardment CCD (EB-CCD)-camera system is a very sensitive detector designed mainly for low-light imaging applications. The major advantage compared to a microchannel plate intensified CCD system is the superior spatial resolution due to the special amplification system. This property would make it an ideal detection system for a high-resolution echelle spectrometer, where high sensitivity in combination with high spatial resolution is required. However, we found that part of the light focused on the photocathode of the EB-CCD is transmitted through it and detected on the CCD chip as a diffuse spot. Thus, it lowers the signal to background ratio and can lead to ghost peaks. The problem becomes particularly serious when detecting emission lines of very different intensities.