Electro-optical and infrared (EO/IR) sensor models are useful tools that can facilitate understanding a system’s behavior without expensive and time-consuming testing of an actual system. EO/IR models are especially important to the military industry where truth data is required but is sometimes impractical to obtain through experimentation due to expense or difficulties in procuring hardware.
This research describes implementation of a focal plane array (FPA) model of charge-coupled device (CCD) and complementary metal-oxide semiconductor (CMOS) photodetectors as a component in the Air Force Institute of Technology (AFIT) Sensor and Scene Emulation Tool (ASSET). The FPA model covers conversion of photo-generated electrons to voltage and then to digital numbers. It incorporates sense node, source follower, and analog-to-digital converter (ADC) components contributing to gain non-linearities and includes noise sources associated with the detector and electronics such as shot, thermal, 1/f, and quantization noise.
Advances in solid-state technology have increased image quality in recent years. Reductions in photodetector sizes have enabled an increase in the number of elements in a pixel array of constant area, allowing proportional increases in the resolution of the system. However, this decrease in pixel size has also reduced the collection area available to convert incident photons to photoelectron signal, which often results in a reduced signal to noise ratio (SNR). This makes it increasingly difficult to sense dim signals or distinguish radiometric characteristics from noise. New and better algorithms for signal processing are one way to compensate for the potential loss in sensitivity and noise performance.
ASSET was created to produce synthetic data representative of real electro-optic and infrared (EO/IR) sensors. Its development was driven by the need for realistic data under controlled conditions without the expense of conducting field or laboratory experiments. ASSET can produce data suitable for signal processing and algorithm development for both real and hypothetical systems by accurately modeling scene and sensor characteristics. The latter is especially important, as military researchers increasingly use modeling for assessing theoretical system performance to keep pace with sensor technology development, without the need to build an entire sensor system.
Many modern image sensors use charge-coupled device (CCD) or complementary metal-oxide semiconductor (CMOS) solid-state photosensors to convert light into a digital signal. Due to imperfections of photosensors, such a conversion is not ideal and leads to noise in the measured signal. Therefore, one can either estimate and reduce the impact of noise from the image sensor, or simulate and predict what the performance of a given sensor will be in terms of image appearance, given the device specifications and key design parameters.
The high-level simulation of noise in photosensors, however, is still an area of active research. The main problem is that photosensors are affected by many different sources of noise, some of which cannot be modeled effectively using only Gaussian noise. The growing use of new pixel architectures, especially staring focal plane array (FPA) technology, introduces new parameters affecting system performance that are not adequately addressed by current models (e.g. offset FPN due to pixel’s transistors). There is a need for high-level image sensor system modeling tools that allow engineers to see realistic visual effects of noise where the user can change individual design or process parameters to quickly see the resulting effects on image quality. This research attempts to address the issue of the lack of high-level photosensor modeling tools that enable the user to simulate realistic effects of noise on CCD and CMOS image sensors.
This work was done by Fernando D. Fernandez for the Air Force Institute of Technology. AFRL-0288
This Brief includes a Technical Support Package (TSP).
A Focal Plane Array and Electronics Model for CMOS and CCD Sensors in the AFIT Sensor and Scene Emulation Tool
(reference AFRL-0288) is currently available for download from the TSP library.
Don't have an account? Sign up here.