Organization � lectures � discussion of research papers � student projects Imaging Sensors � (1-2 student(s) per group) � list of possible ideas � presentation of ideas � project proposal (2 pages) � implementation � presentation of results � report (like a conference paper 6-8 pages) Computational Photography Hendrik Lensch, Summer 2007 Computational Photography Hendrik Lensch, Summer 2007 Image Sensors Image Sensors Photodetection CCD’s vs CMOS Sensor performance characteristics Noise Color Sensors Exotic Sensors CCD CMOS Computational Photography Hendrik Lensch, Summer 2007 Computational Photography Hendrik Lensch, Summer 2007 Photogeneration Integration Silicon Measuring one electron is really hard! � “Band gap” of 1.124eV between valence band (Doesn’t have much energy…) and conduction band . Fortunately, the electrons hang around for a while. Incident photon > 1.124eV (hc/ λ ) may be absorbed, So integrate the charge over a period of time. causing election to jump to conduction band. � 10’s to 1000’s of electrons. Visible light ( λ =400 to 700nm) Two fundamental structures… � λ = 400nm (violet) E = 3.1eV � λ = 700nm (red) E = 1.77eV � λ = 1100nm (infrared), E=1.12eV Computational Photography Hendrik Lensch, Summer 2007 Computational Photography Hendrik Lensch, Summer 2007 Page 1
Photodetectors Photodetector Performance Metrics (a) photodiode, (b) photogate Pixel size All electrons created in depletion region are collected, plus some from surrounding region. Fill factor Full well depth Spectral quantum efficiency Sensitivity (Saving noise & dynamic range for later) image: Theuwissen Computational Photography Hendrik Lensch, Summer 2007 Computational Photography Hendrik Lensch, Summer 2007 Pixel Size Fill Factor Large pixels means more light collected. Percent of pixel area that captures photons. Typically 3 � m-10 � m Typically 25% to 100% 20 � m for astronomy Smaller for photogate than photodiode. Pixels getting tiny for cell phones, digital cameras Reduced by non-light gathering components in pixel (see CMOS sensors…) � 2 � m x 2 � m is probably the smallest CMOS pixel today (Matsushita, ISSCC 2005) Can be increased using microlenses: � Optics will get you eventually. Computational Photography Hendrik Lensch, Summer 2007 Computational Photography Hendrik Lensch, Summer 2007 Lenslets Full Well Depth Increase effective fill factor by focusing light “Saturation charge” 45 to 100 ke – Can double or triple fill factor � depends on the pixel size Limits dynamic range (more about this later) Once you fill up your well, can overflow into your neighbors. This is called blooming. Blooming almost irrelevant for CMOS image: Kodak application note DS00-001 Computational Photography Hendrik Lensch, Summer 2007 Computational Photography Hendrik Lensch, Summer 2007 Page 2
Blooming Extra Overflow Drain http://www.ccd-sensor.de/assets/images/blooming.jpg Computational Photography Hendrik Lensch, Summer 2007 Computational Photography Hendrik Lensch, Summer 2007 Absorption Coefficients Penetration Depth image: Theuwissen Computational Photography Hendrik Lensch, Summer 2007 Computational Photography Hendrik Lensch, Summer 2007 Spectral quantum efficiency Filtered Spectral Quantum Efficiency source: Kodak KAI-2000m data sheet source: Kodak KAF-5101ce data sheet Computational Photography Hendrik Lensch, Summer 2007 Computational Photography Hendrik Lensch, Summer 2007 Page 3
Factors for Quantum Efficiency Extended Sensitivity Color filters Absorption coefficients & depletion depth � Blue light is absorbed quickly, red wavelengths penetrate more deeply. � Photogate detectors have poor blue response because the gate absorbs blue light, too. Fill factor � blue plus – applies a phosphorescent layer � back illuminated CCDs – decrease thickness Computational Photography Hendrik Lensch, Summer 2007 Computational Photography Hendrik Lensch, Summer 2007 Back Illuminated CCDs Sensitivity Sensitivity = quantum efficiency * conversion gain Conversion gain is basically volts per electron. � You don’t want to know about this… � Depends on device process, topology, etc. Sensitivity is often expressed as Volts/lux � 1 Lux = (1/683)W/m 2 at λ = 555nm � 1 Lux (or lumens/m 2 ) = 4.09E11 photons/(cm 2 sec) � Clear sky ~= 10E4 Lux � Room light ~= 10 Lux � Full moon ~= 0.1 Lux Computational Photography Hendrik Lensch, Summer 2007 Computational Photography Hendrik Lensch, Summer 2007 CCD’s vs CMOS Image Sensors Charge Transfer for CCD’s Differ primarily in readout—how the accumulated charge is measured and communicated. CCD’s transfer the collected charge, through capacitors, to one output amplifier CMOS sensors “read out” the charge or voltage using row and column decoders, like a digital memory (but with analog data). image: Theuwissen Computational Photography Hendrik Lensch, Summer 2007 Computational Photography Hendrik Lensch, Summer 2007 Page 4
Example:Three Phase CCD’s Full Frame CCD Photogate detector doubles as transfer cap. Simplest, highest fill factor. Must transfer quickly (or use mechanical shutter) to avoid corruption by light while shifting charge. image: Curless image: Theuwissen Computational Photography Hendrik Lensch, Summer 2007 Computational Photography Hendrik Lensch, Summer 2007 Frame Transfer Smearing vertical streak memory area is shielded wikipedia image: Theuwissen Computational Photography Hendrik Lensch, Summer 2007 Computational Photography Hendrik Lensch, Summer 2007 Smearing Interline CCD Charge simultaneously shifted to shielded gates. Provides electronic shutter—snapshot operation Uses photodiodes (better detectors) Most common architecture for CCDs image: Theuwissen http://www.astrosurf.com/maugis/topo_ccd/smearing.jpg Computational Photography Hendrik Lensch, Summer 2007 Computational Photography Hendrik Lensch, Summer 2007 Page 5
Charge Transfer Efficiency Advantages of CCD’s CCD charge transfer efficiency, η , is the fraction of Advantages: charge transferred from one capacitor to the next. � Optimized photodetectors (high QE, low dark η must be very close to 1, because charge is current) transferred up to n+m times (or more for 3-phase…) � Very low noise. For a 1024 x 1024 CCD: � Single amplifyer does not introduce random noise or fixed pattern noise. Disadvantages η Fraction at output � No integrated digital logic 0.999 0.1289 � Not programmable (no window of interest) 0.9999 0.8148 � High power (whole array switching all the time) 0.99999 0.9797 � Limited frame rate due to charge transfer Computational Photography Hendrik Lensch, Summer 2007 Computational Photography Hendrik Lensch, Summer 2007 CMOS Sensors (active pixel sensor - APS) CMOS Sensors • charge converted to a voltage at the pixel • pixel amp, column amp, output amp. row select bitline Image : EE392B, El Gamal Computational Photography Hendrik Lensch, Summer 2007 Computational Photography Hendrik Lensch, Summer 2007 Example CMOS Pixel Rolling Shutter Photo sensitive area is reduced by additional circuitry. Source: Stanford EE392B notes Computational Photography Hendrik Lensch, Summer 2007 Computational Photography Hendrik Lensch, Summer 2007 Page 6
Rolling Shutter Distortion CMOS Sensors Advantages � Integrated digital logic � Fast � Mainstream process (cheap) � Lower power Disadvantages � Noise & quality Most high quality cameras still CCD’s. Computational Photography Hendrik Lensch, Summer 2007 Computational Photography Hendrik Lensch, Summer 2007 CMOS with Integrated Logic CMOS vs CCD, bottom line CCD’s transfers charge to a single output amplifier. Inherently low-noise. CMOS converts charge to voltage at the pixel. � Read out like a digital memory - windowing � Reset noise (can use correlated double sampling CDS) � Fixed pattern noise (device mismatch) [micro.manget.fsu.edu] Computational Photography Hendrik Lensch, Summer 2007 Computational Photography Hendrik Lensch, Summer 2007 Sources of noise Noise Sources Photon shot noise Dark current shot noise [Reibel2003] Fixed pattern noise [Janesick97] Readout noise … readout noise Computational Photography Hendrik Lensch, Summer 2007 Computational Photography Hendrik Lensch, Summer 2007 Page 7
Recommend
More recommend
