IMAGING FROM NONDEDICATED SATELLITES
This report presents the results of a MITRE-sponsored research project that investigated the use of new technology for creating multispectral images of the earth's atmosphere and surface from geostationary orbit. The use of commercial communications satellites to carry the new-technology imaging instruments also was investigated. The research indicates that it is possible to achieve improved imaging instrument performance compared with current technology. The Geostationary Operational Environmental Satellite (GOES) system was used as a baseline for much of the work in this project since it is the primary source of earth images from geostationary orbit used by the National Weather Service (NWS) for weather warning and forecasting.
The National Oceanic and Atmospheric Administration (NOAA) implements and operates the GOES system for meteorological data collection. Each satellite provides a platform for meteorological data collection instruments whose outputs are used by the NWS and other government agencies, as well as domestic and international meteorological researchers. NOAA, through the National Aeronautics and Space Administration (NASA), has contracted for advanced versions of the GOES spacecraft, including an advanced scanning radiometer imager to provide multispectral images of the earth's atmosphere and surface. The imager for the advanced GOES I through M series spacecraft is being developed to meet performance requirements expressed in 1983 to support the NWS Modernization program. The imager design is an evolution of the imaging radiometer first launched into geostationary orbit in 1975.
In this project, MITRE assessed the technical feasibility of using advanced focal plane array (FPA) detector technology to develop an imager that would achieve NWS performance requirements for imagers for the year 2000 and beyond. In consideration of the large costs and risks associated with space-based programs, an assessment was made of the economic, institutional, and operational ramifications of carrying remote sensing instruments on geostationary satellites other than those dedicated to the meteorological mission.
Analyses were conducted of four areas associated with development of an FPA detector imager designed around state-of-the-art charge-coupled-device (CCD) detector arrays and some of the ramifications of having such an imager included in the payload of a non-dedicated satellite, such as a commercial communications satellite. The sections that follow summarize findings in each of these four areas: (1) Conceptual design for an FPA imager, (2) image navigation and registration (INR), (3) ground-based processing of FPA instrument output data, and (4) deployment opportunities and costs. The final section presents key conclusions from the four areas.
IMAGER TECHNOLOGY AND DESIGN CONSIDERATIONS
A conceptual design was developed for an FPA imaging instrument, incorporating advanced, second generation FPA detectors and mechanical cryogenic coolers. The FPA imager incorporates a 30 centimeter (cm) f/12 Ritchey-Chrétien reflector telescope with dichroic beam splitting aft optics to select visible and four infrared (IR) bands of interest. It incorporates redundant array detectors of 1024 x 512 pixels for the visible channel, 256 x 128 pixel detectors for the short-wavelength IR channel 2, and 64 x 32 pixel detectors for the medium- and long-wavelength IR channels 3-5. Corresponding earth resolutions are 0.5 km for the visible and 2 and 4 kilometers (km), respectively, for the IR channels 2 and 3-5. The proposed instrument is projected to weigh 235 lb and require 200 watts of power.
A tradeoff study of FPA scanning was conducted, in which the relative advantages of using step-stare scanning or a time delay and integration (TDI) method were weighed. TDI scanning offers improved radiometric performance, while the step-stare method provides a basis for improved INR performance as discussed below. With either approach, a full earth disc image containing 6.9 x 10^(8) scene pixels encoded with 11 bits/pixel will result in a 50 megabits per second (Mbps) raw data rate without compression. Given sufficient downlink bandwidth, a full earth disc image can be created and transmitted to the ground every 3 minutes. This refresh rate is an order-of-magnitude improvement over the current and the planned GOES I rate of 30 minutes per image, which is governed by the imaging instrument.
The FPA imaging instrument is predicted to exceed the radiometric performance requirements expressed in 1991 by the NWS for imagers for year 2000, and to achieve the desired performance objectives for imaging beyond the 2000 requirements.
IMAGE NAVIGATION AND REGISTRATION (INR)
For the GOES I through M system, tailored to the remote sensing mission, INR is the name given to a collection of measures used to locate or assign latitude and longitude values for each pixel of an image, and to maintain absolute and mutual latitude and longitude spatial relationships within and among images. INR for the step-stare FPA imaging instrument consists of acquiring overlapping image frames covering the earth view, creating a mosaic of the frames to form an image, and providing annotation for each frame that links together the frames and provides for calibration. Studies of FPA imaging instrument INR were conducted, examining the use of step-stare scanning and contrasting it with TDI and GOES as noted above. While TDI scanning offers improved radiometric performance, the step-stare method provides a basis for improved INR performance.
The FPA instrument can capture frames of up to half a million pixels simultaneously, all pixels in a frame having fixed mutual spatial relationships. The imager allows for overlapping frames captured at very short intervals, reducing the effects of spacecraft motion while retaining spatial relationships by capturing the same image parts in adjacent frames. This provides an image of the full earth disc composed of about 1000 frames having very small error in mutual spatial relationships among frames. The frames themselves capture image areas with recognizable landmark detail used for location reference. Sufficient overlap is included to permit spacecraft motion during imaging operations to expand and contract the overlap without losing it completely, and the frame image exposure time is short enough so that spacecraft motion has minimal effect on image quality. Through these devices, the FPA imager offers opportunities to meet and exceed current INR performance requirements.
Achieving required INR performance with a current-technology instrument requires a complex spacecraft to meet very demanding instrument pointing requirements. Onboard facilities, complex interaction with the spacecraft attitude and orbit control system, and complex ground system processing are required to deal with INR. The FPA instrument requires no complex interaction with the spacecraft other than to receive occasional commands, and to provide an output data stream and housekeeping information for transmission to earth. Furthermore, it appears technically feasible to carry an FPA instrument on a commercial communications spacecraft and produce image results that would, at a minimum, be comparable with predicted GOES I results and better with regard to the quality of image spatial relationships.
GROUND PROCESSING OF FPA INSTRUMENT OUTPUT DATA
Ground processing of FPA instrument output data is needed to provide means for assigning each individual pixel a longitude and latitude, adding geopolitical markings to the images, and providing radiometric calibration. In processing the instrument data stream, pixel values are not changed. Rather, each frame of pixels is annotated with the information needed for frame and pixel location, inclusion of geopolitical boundaries, and radiometric calibration.
Calibrated and annotated frames of FPA instrument data would be suitable for broadcast transmission to the NWS' Advanced Weather Interactive Processing System (AWIPS), currently under development, or to other systems that select satellite imagery data and resample using specialized coordinate systems. From the same broadcast transmission, image frames can be selected to form either a full earth image or a multiplicity of images for regions of interest. This means the FPA instrument can be operated in a free-running mode, with no special scanning modes or schedule interruptions for special events.
Based on the NASA Management Instruction, "Planning and Approval of Major Research and Development Projects" (NMI 7121.1), and a review of comparable instrument development projects, an availability date of 2003 is projected for a flight-ready FPA imaging instrument.
US domestic commercial geostationary communications satellite programs will present launch opportunities for such an instrument in the year 2003 time frame and beyond. This projection stems from the requirement to replace second-generation commercial satellites, having a 10- to 12-year mission life, in the same time frame in which an FPA instrument is anticipated.
A range of development costs for an FPA instrument and unit costs for other deliverable items have been projected based on recent contracting actions. The result, a projected range for a pro-rata share of spacecraft and launch costs, is based on the ratio of the FPA instrument mass to the in-orbit mass of the communications payload of the commercial operator's spacecraft. Since the spacecraft operator would provide all satellite communications, command, and control services, and the government would operate the instrument payload, an estimate of annual costs for telemetry and command and data communications services also has been projected.
SUMMARY OF CONCLUSIONS
An objective of this project was to determine if an imager using FPA detector technology could be designed to provide performance that would meet the NWS requirements for imaging from geostationary orbit for the year 2000 and beyond. We also examined the potential for including such an imager in the payload of a commercial geostationary communications satellite. The limited pointing accuracy of communications satellites compared with that for satellites such as the GOES I, which are dedicated to the imaging mission, led us to explore means to compensate for pointing inaccuracies and for spacecraft motion by exploiting the imaging capabilities of the FPA instrument. We postulated ground processing concepts to achieve required INR performance given the 3 minute full-earth images provided by the FPA instrument. While many tradeoffs and options still exist, we can present our conclusions based on our work to-date. We summarize the detailed conclusions presented in sections 2 through 5 of this report as follows:
* Second-generation FPA detector technology is the key to improved imaging from nondedicated satellites. We believe that a rapid imager is feasible and can provide the resolution and radiometric performance sought by the NWS as expressed in the NWS' GOES-N requirements [GOESN89].
* A number of technologies now under development are important to FPA imager development. For example: IR detector arrays should be cooled to 70 degrees K or less. A passive radiative cooler to do this would be too large for an imager on a nondedicated satellite. The reverse turbo Brayton mechanical refrigerator is preferred. Mechanical refrigerator developments are well underway with lifetime goals of 10 years.
* For step-stare scanning, the most demanding NEdT requirement is to be able to detect a 0.1 degree K temperature change in a 300 degree K scene in channels 4 and 5. This puts a 0.1 percent limit on residual FPA detector nonuniformities, which is within the state-of-the-art for FPA technology. TDI scanning offers superior radiometric performance and reduced requirements for detector production yield compared with step-stare scanning.
* The use of FPA detectors precludes the use of image motion compensation and mirror motion compensation as defined for the GOES I system. This is due to the rigid spatial relationships among the pixels within a single frame of an image, and the consequent inability to alter the positions of individual pixels during image creation.
* The use of overlapping image frames in the step-stare scanning method offers deterministic spatial relationships within and among image frames to a greater degree than TDI scanning. This provides well-coregistered images using ground-based processing for assembling the frames into a whole image. While the predictor-corrector and feedback approaches to INR place the burden of image spatial relationships on the spacecraft and the imager scan system, the overlap approach uses data redundancy and ground-based image processing and results in the simplest space segment. Whole-image assembly from overlapped frames permits compensation for all spacecraft-earth relative motions.
* The proposed instrument can be operated in a free-running mode thereby precluding the need for scheduling special scanning modes or scenarios. The high resolution, improved image quality, and rapid 3 minute update rate for full-earth coverage make the resulting imagery suitable for AWIPS use and for the mesoscale observation objectives of the NWS Modernization program.
* It appears programmatically and legally feasible to deploy FPA instruments on nondedicated commercial communications spacecraft. Costs for such deployment may be less than deployment of a dedicated spacecraft. Our analysis projects cost for an operational instrument plus deployment to be in the range of $50 to $75 million which is at the low end of the cost range for a launch vehicle alone.
* Deployment of an instrument on a non-dedicated satellite could be done as a primary mission or as an adjunct to a primary mission. It also is possible to deploy a developmental instrument in this way to avoid risk to an operational system. Opportunities for deployment on commercial geostationary communications satellites including several Direct Broadcast Service and Mobile Subscriber Service satellites will be available in the post-2000 time frame. There appears to be little opportunity for sharing satellites with currently programmed government geostationary systems.
In the post-2000 time frame, NOAA will need advanced remote sensing technology that includes imaging from geostationary orbit. FPA technology can meet and, in some cases, exceed the performance anticipated in the GOES-N requirements for that era. Technical, programmatic, and legal issues do not appear to present barriers to lower-cost, high-quality imaging instruments to satisfy anticipated needs.