733667-001-011-0121
AI Summary
This document discusses various aspects of radiometric distortion and noise in imaging systems, focusing on definitions and characteristics of different types of noise and their impact on image quality.
Key Findings
- Radiometric distortion refers to deviations in spatially imaged points from their intended energy values. - Types of noise discussed include temporal noise, fixed pattern noise, and chromatic noise. - Various definitions and metrics for assessing noise, such as Noise Power Spectrum (NPS) and root mean square deviation, are provided. - Factors affecting image quality include sensor performance, optical design, and environmental conditions.
OCR Text
Federal Agencies Digitization Initiative Stil Image Working Group — August 2010 —" — Radiometric Distortion - ? aitite 9 definition: The deviation of any given spatially imaged point from an aim radiant energy value relative to the input object. w Noise Power Spectrum (NPS) Chromatic Noise 5 Total Noise eee 5 ul M UM Definition : The inter-color channel! radiometric deviations 5 g Definition : A spatial frequency descriptor of the vit em of radiometric noise of an imaging relative to an identified aim component or m Tem ral Noise Fixed Pattern Noise Color Uniformity Color SFR uniformity Random Banding/ Defects = Definition ; A difference in large | Definition: li woleolpel (stochastic) , (stochastic) ‘ area uniformity/shading spread of light between color 2 8 Definition : The root streaking Definition point or shadin between color channels channels. 8 | mean square deviation | (deterministic) clusters of {determin'stic) ; Definibon : One Definition: A deviation in = ( std. deviation) of both di defective or poorly alli iin mensional ‘ the effective illumination temporal and fixed corrected pixels ‘. patterns over a capture device's pattern noise for a , venriniemneapiee single color channe! ield of view; rang yw lower illumination near the field's outer extent. - Temporal noise « Stripes - Hot, Cold, or - Vignetting - Rainbows — Colored edges ue - Grain - Banding Dead Pixels - Relative illumination ~ Color Bleed £ SE | - Shot noise - Streaking - Wounded Pixels ~ Fringing 2% £ - Read noise ~ Blinkers U - White noise e @: RMS deviation of | @: The relative @: The number or | @: The percent deviation @: The percent deviation of @: The difference in SFR @ || 3 pixel values in terms of | amount of size of defects of several large area several large area chroma response between selected Selected metric(i.e., | variance or noise | per unit sensor luminance measurements | measurements over the field of | color channels, aos counts, density, | power that a area, over the field of view | view relative to the average of | units: % deviation in SFR q pt Luminance) over an | selected spatial units: # of relative to the average of | those chroma measurements. response relative to the : Se identified region of || frequency band defects/unit those measurements, units: % chroma difference highest measured SFR i 5 fe) interest contributes to senate bran units: % Luminance (unit less) ( unit less) § the total noise. difference (unit less) & units: counts, density, | units: TBD 2) Luminance - Aggressive digital ~ Poor sensor ~- dust on sensor - poorly designed optics - Chief ray angle (CRA) - Poor optical design or - signal amplification or calibration ~ poor sensor - non-uniform lighting mismatch between optics and performance 3 w processing - dust/dirt on fabrication sensor & « High ISO sneed linear array - Non-uniform color coatings at = 3 selection sensor - poor sensor sensor fabrication. g - High throughput “ poor sensor calibration | 2 | workflows calibration
Metadata
- Agency
- —
- Classification
- UNCLASSIFIED
- Department
- National Archives and Records Administration
- Catalog source
- View NARA catalog record
NARA Source
- NAID
- 733667
- File
- 733667-001-011-0121.jpg
- Type
- image/jpeg
No machine-readable OCR text for this asset. Photographs without captions may have no extractable text.