Charge-coupled device (CCD) area image sensors of the interline-transfer type include design features which enable the transfer and detection of signal charge packets of the order of tens of electrons. These sensors, which include devices with 190 x 244 and 380 x 488 elements, have application in all-solid-state TV cameras designed for use with very low levels of scene illumination. The designer of TV cam-eras which fully exploit this low-light performance potential is faced with a challenging constraint; extraneous noise effects must be mini-mized. In order to assess progress to date, both for sensor design and utilization techniques, the theoretical performance limitations imposed by non-cancellable noise of the detection process, and the effects of sensor CTF, are examined. This paper describes an analytical model for predicting the sensor-limited resolution-irradiance characteristics of CCD-TV camera systems as a function of responsivity, CTF, dark charge, on-chip amplifier NES, optical image contrast, and the required threshold signal-to-noise ratio. The results of the analysis are compared with recent experimental results approaching the predicted sensor-limited camera performance.
To fully realize the potential of charge coupled devices for solid state low light level imaging additional gain before the array appears necessary. This can be accomplished either by coupling the device to an image intensifier tube optically or by operating in the electron bombarded mode (i.e. EBS). As in the case of the electron bombarded silicon diode array vidicon the performance in the electron-in mode is expected to be superior to the performance in the optically coupled mode. 100 x 160 element thinned CCD arrays have been operated in the EBS mode and the results indicate that substantial improvement in low light level performance can be realized when compared to direct photon-in performance. Imaging has been demonstrated at an equivalent signal level of 5 electrons/pixel (i.e. Actual No./Gain).
Solid-state imagers can displace electronic beam scanned imagers primarily in applications which con-ventional TV has not satisfied. Included are imaging from a moving platform like a weather satellite, imaging with very large intrascene dynamic range, imaging of very low contrast scenes including imaging in the infared where solid-state arrays in optomechanical scanners provide basically better performance and TDI techniques are increasing their advantage, and very low light imaging where "photon counting" techniques using an ICCD with an external computer memory can eliminate problems of dark current and of preamplifier noise in long time exposures and provide higher signal-to-noise ratio for improved radio-metric accuracy. In direct comparison, solid-state imagers lead for larger effective dynamic range, greater maximum signal-to-noise ratio, freedom from lag, geometric fidelity and metricity, and stability of characteristics, as well as the obvious compactness, long life, and low operating voltages. Tubes lead for the greatest number of resolvable elements per frame, for producing images with better element-to-element uniformity, and for maximum output data rate on a single terminal.
The basic purpose of this paper is to review the performance improvements that have been made over the past few years and to highlight the strengths and weaknesses of the present microchannel plate (MCP) inverter. On reviewing these strengths and weaknesses, present applications will then be surveyed and a few potential applications will be highlighted.
Infrared imagers have been used to remotely sense several potentially hazardous conditions in mining operations; among these are loose rock, misfires, shorted power cables, overheated bearings, and combustion (both existing and incipient). In addition, the potential use of infrared imagers as mine rescue tools has been evaluated.
Infra-red thermography by itself, despite much wishful thinking, will not quantify heat losses, nor is it always cost-effective when used for other purposes, but it does have a place as a useful tool when used with discretion.
Within a year after x-rays were discovered by Roentgen, x-ray imaging was being used not only for medical applications, but also by Customs and Security people for examining the contents of packages and parcels. The full potential of the penetrating nature of x-rays with its ability to produce easily recognized images was quickly recognized, but its implication was impractical and proved less valuable for security use than for use in the medical profession. Hence, x-ray security systems faded away and were employed only on a very limited basis for about 70 years (1900-1970). During this period, manufacturers were offering industrial or medical x-ray units primarily designed for other functions. These machines failed to develop a viable market and no more than a few dozen units were actually employed in security functions. Even though the machines presented an x-ray image of the parcel, why did security officers avoid them?
For the last two years, the USDA Forest Service has been engaged in an R&D program to expand helicopter firefighting operations into the nighttime by use of current night vision technology. As a result, Land Managers are beginning to utilize some of the systems and devices for other tasks as well. These include law enforcement on National Forests, surveying techniques, nocturnal game studies, search and rescue, and reconnaissance duties. This paper describes the equipment in present use, training requirements, and typical operations.
Baird-Atomic has developed and marketed image intensification systems to the U. S. Military and commercial customers. In order to be successful in the latter market, maximum utilization must be made of military-funded components, existing sales and service organizations, and company commitment of personnel and front-end money. Several examples of commercial projects are examined and con-trasted to military night vision programs.
Much effort has recently been devoted to Charge Coupled Devices (CCD's) as imaging detectors. This paper describes a large imaging array CCD program currently underway at the Jet Propulsion Laboratory and Texas Instruments and presents test results obtained on both 100 x 160 element and 400 x 400 element arrays. Expected low light level performance is also given, along with a brief description of future plans.
During the past year several types of silicon self-scanning arrays have become available. The Charge Injection Device (CID) which is processed in a 100 x 100 format has been examined in detail and is presently being used on a regular basis at Kitt Peak National Observatory for a special class of astronomical problems. The random address and non destructive readout features make the CID a unique member in the family of array devices. Linearity, MTF, and Quantum Efficiency have been measured and detailed results are given.
The development of high performance solid-state sensors is important for future astronomical space flight missions. Although sensitivity in the ultraviolet is obviously advantageous, the potential for superb imaging afforded by space telescopes recommends a sensor design covering a broad spectral range including the visible and near infrared. Additionally, the sensor must be rugged, reliable, and relatively insensitive to corpuscular radiation. With these considerations in mind, the Laboratory for Optical Astronomy has begun the development of a sensor incorporating a CCD and photocathode for operation in the electron bombarded mode in both a magnetic and an electrostatic focussed, gated configuration. Sufficient gain will be available to provide adequate signal-to-noise for the detection of individual photoelectron events. Initial format size is 100 x 160 pixels with eventual growth to 400 x 400 pixels. The sensor is part of a digital camera system which includes the low level video condit-ioning electronics, a camera controller, a high speed buffer memory, and digital recording and display electronics. The memory uses CMOS/SOS and has a capacity of 1.6 Mbits with operating rates of 48 Mbits/sec. Individual frames are co-added to provide wide dynamic range and photometric precision better than 1%. A 4-bit video quantization is used to increase the photon counting detection rate before coincidence losses become serious.
CCD imaging arrays have been evaluated for application in a photon counting astronomical sensor. The results obtained show that CCD devices presently being manufactured exhibit noise corresponding to less than 100 rms electrons and gains of approximately 103 at an accelerating voltage of 10kV and greater than 5 x 103 at 25kV. These results demonstrate that the CCD devices, when operated in an intensified mode, can be used for photon counting. In addition, the measured charge transfer efficiencies are in excess of 0.9999, signal saturation levels are as high as 9.9 x 105 electrons per pixel and peak quantum efficiencies are greater than 75%. With these characteristics, the arrays will make excellent analog detectors surpassing commercial vidicon performance in small format applications. Successful proximity focussed intensified charge coupled devices (ICCD) have been fabricated, although problems still exist with contamination and possible structural failures in the CCD's during tube manufacture.
The operation and performance of an array photometer with the ability to discriminate on single photoelectrons will be discussed. The University of Maryland Array Photometer is a system which consists of a 100 by 100 array of channels which count single photoelectrons. The photosensor is an Intensified Charge Coupled Device using a Fairchild CCD201 which was fabricated by the Electronic Vision Company. This system has demonstrated a noise level which permits single photoelectron discrimination, high scan rate (400 frames/ sec), and low lag. The system and its performance will be discussed. The UMAP system has been successfully operated in the photon counting mode on a 36-inch telescope and on a 48-inch telescope. Data from these observations will be discussed, to illustrate the system sensitivity and the spacial resolution which is better defined by the size of an individual pixel element on the CCD.
Electronic imaging devices based on the magnetically-focused, internal-optic Schmidt image converter concept, previously used at NRL for the vacuum ultraviolet, have been developed with cesium telluride photocathodes for use in the middle ultraviolet (2000-3000 A). These devices are intended primarily for flame and mid-UV source obser-vations, but also have applications to astronomy and to planetary atmosphere studies. Three versions of these devices have been constructed and tested: (1) a single-stage image converter with a phosphor/fiber optic output, for film recording or coupling to a low-light-level television camera tube, (2) an image converter incorporating a micro-channel intensifier stage with proximity focusing of the output onto a phosphor screen, and (3) same as (2) except with magnetic focusing from the microchannel plate output onto the phosphor screen. The latter two devices can be used with direct viewing of the phosphor output, or with lens coupling to a film camera or television camera tube. They can also be fiber-optically coupled to a CCD array. In each case, single-photo-electron events are detectable. Cesium telluride photocathodes appear to be significantly more stable in the presence of small leaks and outgassing than are cesium antimonide photocathodes. Therefore, electrographic recording should be easier with the Cs2Te photocathodes, and preliminary investigations are being initiated.
The disadvantages of full size film and film screen techniques have stimulated research during recent years for an alternate approach. For our technique, we have been developing a low light level television system to eliminate many of the disadvantages thereby yielding the potential for improved diagnosis of early pathology at reduced radiation, cost, procedure time and storage and retrieval expense. This isocon imaging system is being developed for early breast cancer, lung cancer and pneumoconiosis detection.
Various scintillation cameras using image intensifiers are described and compared in their important imaging characteristics with existing instrumentation. The main advantages of image intensifier cameras are in the areas of positional resolution, count-rate capability and system simplicity. The main problem in the past has been the general lack of good gamma-ray energy discrimination which results in low contrast images in clinical situations. Good energy discrimination is in principle possible with several of the described systems.
A complex system featuring a video-camera connected to a video disk, cine (medical motion picture) camera and PDP-9 computer with various input/output facilities has been developed. This system enables the performance of quantitative analysis of various functions recorded in clinical studies. Several studies are described, such as heart chamber volume calculations, left ventricle ejection fraction, blood flow through the lungs and also the possibility of obtaining information about blood flow and constrictions in small cross-section vessels.
Our present understanding of breast cancer indicates that increased metabolic activity occurs and thus produces a local temperature increase. This paper evaluates present thermographic techniques used to detect these temperature increases and examines the problem of skin emissivity variations that can produce erroneous temperature measurements. These false temperature variations are on the order of the decision level used by radiologists, and therefore they can cause significant confusion in the interpretation of the thermogram. The ratio temperature thermograph is shown to reduce the effects of emissivity by measuring the spectral radiance at two prescribed wavelengths and ratioing the results. A dual-channel ratio thermo-graph was built using state-of-the-art detectors and electronics to prove its feasibility. The ratio temperature thermograph was quantitatively evaluated for small emissivity variations. This evaluation demonstrated the instrument's capability of minimizing emissivity effects. It was also evaluated for the detection of temperature changes.
The Honeywell Medical Scanner has been developed to accurately measure human body temperatures in a laboratory environment. Specifically, it detects abnormal blood profusion, an indicator of cancer, in the breast and surrounding areas of the body. The scanner uses Honeywell's high performance TV compatible FLIR technology, similar to the AN/AAQ-9 and Chaparral FLIRs, to produce radiometrically calibrated FLIR imagery which may be digitized, for computer processing, or stored on video tape for playback at a later date. The calibrated imagery is produced by referencing the image seen by the front objective to two internal standards. These standards set the gain and level of the output video such that voltage out corresponds to a given observed temperature. By sampling a section of the frame and integrating over several frames, thermal resolution of 0.1°C may be achieved.
Second Generation Night Vision Devices have received some consideration as comilmer cial products, but rarely as consumer goods, not from lack of potential applications, but because their high cost was inevitably considered to reduce the size of the potential market nearly to the vanishing point. Working with the Retinitis Pigmatosa Foundation, ITT has designed a Second Generation Intensifier Monocular as a medical prosthetic aid for persons suffering from retinal degenerations causing night blindness. The proposed use environment for such an instrullent posed an altogether different set of contraints on the design from those encountered in designing military instruments; some require ents of course are relaxed, but others, surprisingly are at least as stringent as those for military application, though often for different reasons. This paper describes the resulting instrument, showing the tradeoffs among cost, performance, human engineering, and environmental considerations, and indicating the rationale behind the design deci-sions made.
Photoelectron microscopy is a surface technique which provides topographical information using the photoelectric effect as a basis for contrast. Progress in the biological applications of this technique is briefly reviewed. Due to relatively low quantum yields, photoemission from biological samples is weak and an image intensifier is used in order to visualize and record the photoelectron image. Currently the limiting magnification is determined by UV power incident on the sample. Power requirements for high-magnification imaging are calculated in terms of microscope, sample, and image intensifier parameters. To approach 40 R resolution, an instrument magnification of 12,000-50,000 is required along with a UV intensity of 0.01 to 10 Watts/cm2 depending on the wavelength and sample. For a tightly focused laser source the total power requirement is 1 mWatt or less.
Low-light-level video techniques are applicable to the non-destructive study of biological phenomena at both the organ or macroscopic level and the cellular or microscopic level. At the macroscopic level, a low-light-level video fluorometer has been employed to study transient NADH changes occuring during cortical brain stimulation and ischemic changes occuring during myocardial infarction. At the microscopic level, low-light-level video techniques have been employed to study rapidly occuring transient events in living cells that are photosensitive at light levels normally employed for visualization.