On the base of local criteria of processing quality, a class of local adaptive linear filters for image restoration and enhancement is introduced. The filters work in a running window in the domain of DFT of DCT and have O(Size of the window) computational complexity thanks to recursive algorithms of mnning DFT and DCT. The filter design and the recursive computation of mnmng DCT are outlined and filtering for edge preserved noise suppression, blind image restoration and enhancement is demonstrated.
Key words: Image restoration; Image enhancement, Adaptive filters
The operational limitations exposed during the Gulf War have led to the formulation of a requirement for anew generation of tactical reconnaissance pod for the Royal Air Force Tornado aircraft. The pod will contain a high resolution Electro-Optical sensor capable of day and night-time operations, digital recording of the imagery for airborne replay and ground exploitation, and a data-link for real time/near real time imagery transmission. The program requirement includes a deployable ground exploitation system to provide a comprehensive independent capability. The interoperability of the air and ground segments with other systems is addressed through NATO standardization agreements. This system will provide the Tornado with a highly flexible stand-off imaging system for day/night operations from a range of altitudes.
The Argus aircraft is a highly modified NC-135E fitted with an infrared and ultraviolet-visible sensor suite for radiometric and spectral data collection. Each suite is operated independently with its own separate gimbal for precision pointing, telescope, and relay optics. The system includes a silica window for the ultraviolet-visible, and a zinc selenide window for the infrared. The entire system was developed and fabricated in-house at the Phillips Laboratory. All sensors are calibrated as a system onboard the aircraft through a unique facility called the aircraft optical calibration facility. The data is all recorded digitally, and can be transferred to secure data reduction facilities via optical fiber. The system is modular, in that the ultraviolet-visible and infrared benches can be separated, or the entire system can be quickly removed to allow for the introduction of other sensor suites or systems. The gimbals and telescopes can be used independently of the rest of the system. The aircraft is also fitted with an anemometry system, which can be operated independently of the sensor systems. This aircraft is capable of many types of missions, and will soon be fitted with a LIDAR system for remote sensing. The philosophy in building the system is to make it capable of quick changes during mission.
The development of the new reconnaissance pod for the German Air Force is continuing according to schedule. A first flight is planned for the end of 1996. Carried on the centerline station of the IDS Tornado, the pod contains two daylight film cameras and one infra-red linescanner system. The infra-red image is recorded on a digital tape recorder and will also be displayed on the TV-Tabs. The modular structure of the pod exhibits a high flexibility for incorporation of various sensor systems on other payloads.
A number of commercial satellite systems have been proposed for operation in the 1997 time-frame with a capability to image at one-meter GSD. For reconnaissance applications, several key attributes determine system utility, including image interpretability, timeliness, area coverage capability, revisit frequency, satellite agility and metric accuracy. Interpretability, or the ability to extract intelligent information from an image, is driven by a variety of key parameters including edge sharpness, signal- to-noise ratio, spectral information and radiometric accuracy. Metric accuracy results from a system's ability to accurately reconstruct an image to a ground coordinate system--both with and without the use of ground control--to an accuracy required for mission planning, weapons delivery and mapping. Timeliness is driven primarily by the ability of the satellite constellation to access the same geographic location in short time intervals and at adequate GSD; sufficient agility is required to access multiple distributed point targets and collect large-area coverage for search applications. Unlike airborne reconnaissance, a satellite system's ability to deliver timely data can be adversely affected by cloud cover. Hence, frequent low-GSD revisit with dynamic tasking and collection capability are critical measures of effectiveness. Analytical descriptions and photographic simulations depict the interaction with and effect on these reconnaissance utility drivers as a function of the key system parameters.
The Ariel unmanned aerial vehicle (UAV) was designed for NASA Ames Research Center to satisfy emerging civil science needs for subsonic flight at altitudes on the order of 100,000 ft. These include atmospheric monitoring of chemical species and environmental conditions related to global climate change. Ariel may be useful for a variety of civil and military remote sensing applications since, at an altitude of 100,000 ft, the UAV wold fly above all manned aircraft. The Ariel has a gross weight of 6400 lb with a wing span of 105 ft, a little shorter than that of the manned ER-2. Ariel is powered by a new propulsion system called the Bipropellant Expansion Turbine (BET). With a 300 hp BET, Ariel can climb to an altitude of 100,000 ft and loiter at Mach 0.63 for two hours while carrying a 600 lb payload. During this loiter, the UAV travels about 750 nm at 100,000 ft. It is possible to trade payload weight for range or endurance. Further design optimization or use of more advanced technology can result in substantially improved performance. With adequate funding, a proof of concept version of Ariel could be developed for initial flights by the year 2000.
The Pioneer UAV system has seen operational use in every U.S. contingency operation since the system's original fielding in 1986. Originally procured as a non-developmental item, the Pioneer was selected for purchase after a successful fly-off competition which was conducted in late 1985. The Pioneer system is a Department of Defense joint system, having been flown by the U.S. Navy, U.S. Marine Corps, and U.S. Army. The system received extensive acclaim for outstanding performance in Operational Desert Shield and Desert Storm. Pioneers are currently being flown by the U.S. Navy from LPD class naval vessels and the U.S. Marine Corps from land based operations. Both services are currently supporting the NATO Joint Task Force in Bosnia.
This paper describes the development of an IBSM (Image Based System Model) using a UAV (Unmanned Air Vehicle) reconnaissance suite consisting of a visible sensor, IR sensor, and data link as an example. Included in the data chain are target disposition and geometry, atmospheric effects, sensor image chain OTFs (Optical Transfer Functions), data compression and transmission interaction and effects, and data display and/or recording. The presentation will include a description of the process of writing custom Khoros-callable functions expressly for this application of the 4 GL. Interesting behaviors of Khoros discovered during development and the transition from version 1 to version 2 of Khoros will be covered. Emphasis will be on the application of the resulting images and data to the system acquisition process from source selection through test. Growth applications of the techniques developed beyond sensor systems will also be discussed.
In 1994, wildfires burned over 117,360 hectares (290,000 acres) of forest land on the Payette National Forest in central Idaho. To assess the impact of the fires on the timber resource, a special mission was planned for NASA's Earth Resources-2 high altitude reconnaissance aircraft. For this project, the aircraft carried a 12 inch lens Wild- Heerbrugg RC-10 aerial mapping camera loaded with Kodak SO- 060 Aerochrome Infrared film. The 1:60,000 scale photography was used to map burn intensity within the gross fire perimeter in high, medium, low and not-burned classes. Polygons were digitized and the digital information was used in the ARC-INFO environment to determine the gross effects of the wildfires on the timber resource. An accuracy assessment was performed by comparing the interpretation to data collected from plots established within the burned area. In this paper, we review the need for the information, the mission parameters, the image exploitation phase, the results of the accuracy assessment and the application of the data to forest management.
The use of three off-the-shelf tactical reconnaissance imaging sensors has been approved for Open Skies Treaty assigned aircraft overflights. A fourth sensor, a video camera has also been approved. The quality of the collected imagery is limited by placing restrictions on the spatial resolutions of the systems, ignoring the additional spectral data that will be acquired. All weather, day and night coverage is authorized over any target of interest. This paper illustrates the kinds of information (intelligence) that will be acquired over three types of targets of interest to Treaty participants. The data developed for this paper is based on analyses of Treaty training overflights and historical tactical missions flown using the same sensors. The data was then correlated with criteria and ratings in the recently published Civil National Imagery Interpretability Rating Scale Reference Guide.
This paper presents an overview of a new camera product for manned tactical reconnaissance applications where optimum ground coverage is desired. A review of Electro-Optical (E- O) framing technology is presented. Camera design and major hardware elements are discussed, and camera performance is compared to that of pushbroom and panoramic E-O cameras. The paper also describes the modes of operation and shows examples of step frame imagery.
Low noise images are contract-limited, and image restoration techniques can improve resolution significantly. However, as noise level increases, resolution improvements via image processing become more limited because image restoration increases noise. This research attempts to construct a reliable quantitative means of characterizing the perceptual difference between target and background. A method is suggested for evaluating the extent to which it is possible to discriminate an object which has merged with its surroundings, in noise-limited and contrast limited images, i.e., how hard it would be for an observer to recognize the object against various backgrounds as a function of noise level. The suggested model will be a first order model to begin with, using a regular bar-chart with additive uncorrelated Gaussian noise degraded by standard atmospheric blurring filters. The second phase will comprise a model dealing with higher-order images. This computational model relates the detectability or distinctness of the object to measurable parameters. It also must characterize human perceptual response, i.e. the model must develop metrics which are highly correlated to the ease or difficulty which the human observer experiences in discerning the target from its background. This requirement can be fulfilled only by conducting psychophysical experiments quantitatively comparing the perceptual evaluations of the observers with the results of the mathematical model.
The assessment of imaging system performance is critical to ensuring a system is delivering the highest quality products possible. The principle two methods for evaluation of reconnaissance system performance are resolution and the National Imagery Interpretability Rating Scale (NIIRS). This paper describes both methods, and presents benefits and limitations of each. Use of the NIIRS is shown to be on the increase as it and an associated prediction model have been relatively recently released unclassified.
There are numerous technological challenges in the Tactical Reconnaissance (Tac Recce) arena as the digital imagery
era dawns. Foremost among them are the problems of imagery transmission bandwidth and the storage of the collected
In this paper I seek to address these problems in an interrelated manner. I do not propose any new technological
innovation, but rather a fundamental change in the philosophy of the collection, transmission, and storage of tactical imagery.
The core of the approach requires that the area being imaged has already been imaged before (old imagery). This is
reasonable given satellite, long range, UAV, and tactical imagery collection systems presently planned for, anticipated data
collection rates, and how hot spots are repeatedly imaged. In addition, the Defense Airborne Reconnaissance Office
(DARO) expects to be imaging tens of thousands of square kilometers each day within the next decade.
When new tasking to collect imaging is received, imagery collected before by some imagery collection system must be
taken with the aircraft (A/C) or person sent out to collect new imagery. As the new imagery is collected, the old and new
imagery of the same area would be automatically registered. The old imagery can be pre-scaled, pre-warped, pre-rotated,
etc., in order to maximize the efficiency of this process. The registered images can be spatially and spectrally thresholded in
order to isolate significant deltas. Automatic target cueing (ATC)/automatic target recognition (ATR) could be used on both
images for comparison to further isolate new objects of interest. Segmentation techniques could then be used to extract
objects or regions of interest from the new image and only these objects or regions would be transmitted to the ground, a
relay aircraft, or a satellite.
Once at the ground station or long-term storage site, the new information could be inserted into the original image, thus
minimizing the amount of storage space required as areas are repeatedly imaged.
This paper assumes satellite communications is a preferred means of imagery transmission due to its real-time imagery
transmission and dissemination advantages.
Exploitation of remotely sensed and aerially derived imagery has, in the past, been primarily performed through the use of
analog light tables, by displaying individual pieces or rolls of imagery over a brightly lit surface to allow light through the nonopaque
surface of the film medium. The interpreter would then peer through optical viewing scopes allowing him (or her) to
analyze the imagery. Over the course of the last two decades, digital data, or as it is better known, "softcopy imagery," has for
many become the desired path which technology has dictated. Softcopy imagery offers many benefits, such as the ability to
manipulate imagery in ways analog workstations cannot and were never designed to do. Functions which can be performed on
softcopy imagery are endless and growing constantly: image spatial rectification, pixel manipulation, image contrast, and brightness
enhancements. All are performed by the running of algorithmic equations to manipulate the digital data. It has become evident
that in the future a large portion of imagery analysis will be performed by softcopy. However, studies indicate that aerial
imagery will continue to be acquired via hardcopy means for many civil, educational, and commercial applications in the
foreseeable future, making it clear that any large scale transformation from hardcopy to softcopy will not be feasible for a long
time to come. A major issue dictating the slow-down in this transition is the over 35 years of hardcopy imagery archived and
housed in facilities throughout the world, including the recently declassified "Corona" satellite imagery which will provide a
wealth of hardcopy data for use by ecologists and conservationists. Yes, the technology to transfer hardcopy to softcopy exists,
but the time and cost required to complete this task would be phenomenal and, in many cases, when digitization and storage
become affordable, it still may prove beneficial to retain the imagery in a hardcopy form for retention of the highest quality
resolution. An analogy which I feel best portrays this dilemma is the automobile-eventually all automobiles will be electric or
hydrogen driven but the time and cost involved in the transformation predicts a slow progression. Since a predominate amount
of imagery analysis, especially in the intelligence community, is the comparison ofnew imagery data to that of archived imagery
in order to detect changes or to monitor progressions, it is conceivable that the majority of imagery analysts will be using a
combination of hardcopy and softcopy workstations in order to facilitate analysis. The incorporation of hardcopy and softcopy
functions into one workstation is the most cost effective and time essential means in which in-depth analysis can be performed.
In this time of declining Department of Defense budgets, it is imperative that more cost effective solutions are identified for meeting present requirements. Lockheed Martin Tactical Communication Systems is under contract to develop a Commercial Off-The-Shelf-based communication ground terminal that is Common Data Link (CDL) compatible. This system is called the Tactical Interoperable Ground Data Link or TIGDL. The TIGDL is made up of more than 50% COTS, and provides the flexibility and scaleability to add and/or delete functions depending on user requirements and constraints. The TIGDL includes an antenna trailer that is capable of being towed by a High Mobility Multi-Wheeled Vehicle. The TIGDL will also provide an Asynchronous Transfer Mode fiber optics interface to the user that meets the Common Image Ground/Surface Station requirements.
Over ten years ago, when ATARS architecture and components were selected to operate at 200 feet at 600 knots, high-end digital tape recorders were the only solution for recording digital imagery for various tactical airborne reconnaissance platforms. But now an all solid-state solution for high speed recording has evolved. High-speed Solid State Recorder (HSSR), which is described in this paper. Trade-offs made during HSSR development are discussed showing how this new product's parameters were selected. Convergence of existing, Fairchild-proven, solid-state mass memory product technologies with availability of commercially driven, high- density FLASH memory chips produced this needed COTS solution to recording imagery or other high-speed digital data in harsh environments, typical of military fighter aircraft.
To conduct observation flights under the terms of the Open Skies Treaty, Germany is using the surveillance system installed in a Tupolev 154 M aircraft. The equipment provides the mission crew with autonomous long range mission capability, totally independent of ground support. The mission equipment provides the necessary functions requires to perform the observation tasks like: mission planning; terrain mapping; video supervision of the overflown terrain; aircraft position reference; and generating, and storing of ground image data. The system described in this paper defines a Video Optical System for high resolution color images. Images can be displayed on board in real time, annotated with mission relevant data, like date, time, location, mission data and ground coordinates. The high resolution images will be recorded on a digital tape recorder. The all digital data handling guarantees optimized results for ground evaluations.
Area array sensors operating in the visible portion of the spectrum can provide imagery suitable for real time and near real time tactical reconnaissance applications. Recent CCD sensor developments, image compression developments and developments in imaging system hardware are enabling greater volumes of imagery to be collected, stored, processed and transmitted to support tactical applications. Area array CCDs with 6000 X 6000 pixels or larger can provide synoptic coverage with good resolution. ITO clock electrode architectures enable extended spectral range and higher responsivity while real time compression permits data transfer of image files in near real time. Flight testing of evolving hardware systems is under way to demonstrate overall system capabilities. The Tier II+ visible EO camera is an example of evolving capability.
In the era of Declining Defense Dollars, the cost of sophisticated aircraft and highly trained personnel has heightened interest in Unmanned Air Vehicles (UAVs). The obvious lure is the lower vehicle cost (no crew station and crew support systems) and reduced needs for highly skilled air crews. Reconnaissance (commonly called recce) aircraft and their missions are among the commonly sighted applications for UAVs. Today's UAV recce aircraft (such as the Predator) are the genesis of much more sophisticated UAVs of the future. The evolution of the UAV will not be constrained to recce aircraft, but the recce mission will be significant for UAVs. The recce hole has historically been that of a battlefield data collector for post mission review and planning. In the electronic battlefield of the future, that role will be expanded. Envisioned mission for future recce aircraft include real-time scout, target location and fire coordination, battle damage assessment, and large area surveillance. Associated with many of these new roles is the need to store or assess much higher volumes of data. The higher volume data requirements are the result of higher resolution sensors (the Advanced Helicopter Pilotage infrared sensor has a data rate of near 1.2 Gigabits per second) and multi-sensor applications (the Multi-Sensor Aided Targeting program considered infrared, TV, and radar). The evolution of the UAV recce role, and associated increased data storage needs (from higher data rates and increased coverage requirements), requires the development of new data storage equipment. One solution to the increased storage needs is solid-state memory. As solid-state memories become faster, smaller, and cheaper they will enable the UAV recce mission capability to expand. Because of the speed of the memory, it will be possible to buffer and assess (identify the existence of targets or other points of interest) data before committing to consumption of limited storage assets. Faster memory search times and random memory access will permit comparisons of iterative imagery to identify changes (moving-target-indicator or new/missing objects) in real-time. This has three positive effects: (1) Real-time assessment of enemy troops and weapon strengths (coupled with directional data) for planning or fire coordination. (2) Real-time battle damage assessment to accelerate determination of the allocation of friendly assets. (3) Selective storage of imagery (only that which contains points-of-interest) reduces the amount of data storage leading to increased area coverage and mission times. Orbital Sciences Corporation/Fairchild Defense is completing the development of a High-Speed Solid State Recorder which is a critical element of the Enhanced Recce Management System we are developing. The High-Speed Solid State Recorder will be a major part of defining the mission for future UAV recce aircraft.
Modern High Density Digital Recorders are ideal devices for the storage of large amounts of digital and/or wideband analog data. Ruggedized versions of these recorders are currently available and are supporting many military and commercial flight test applications. However, in certain cases, the storage format becomes very critical, e.g., when a large number of data types are involved, or when channel- to-channel correlation is critical, or when the original data source must be accurately recreated during post mission analysis. A properly designed storage format will not only preserve data quality, but will yield the maximum storage capacity and record time for any given recorder family or data type. This paper describes a multiplex/demultiplex technique that formats multiple high speed data sources into a single, common format for recording. The method is compatible with many popular commercial recorder standards such as DCRsi, VLDS, and DLT. Types of input data typically include PCM, wideband analog data, video, aircraft data buses, avionics, voice, time code, and many others. The described method preserves tight data correlation with minimal data overhead. The described technique supports full reconstruction of the original input signals during data playback. Output data correlation across channels is preserved for all types of data inputs. Simultaneous real- time data recording and reconstruction are also supported.
In 1995 the F/A-18 TAC RECCE Program was expanded beyond the initial electro-optic and infrared image recording capability of the Advanced Tactical Air Reconnaissance System (ATARS). The program now also includes integration of new high-resolution Synthetic Aperture Radar (SAR) reconnaissance modes and air-to-ground data link of ATARS and SAR image data. Delivery of the first reconnaissance equipment production units and fleet release is scheduled for 1998. F/A-18D two-seat aircraft will be retrofit with the RECCE equipment and designated F/A-18D(RC) (Reconnaissance Capable). This presentation describes recent F/A-18D(RC) operational assessment results, tactical reconnaissance equipment, functions, and interfaces for the 1998 fleet release. The equipment consists of the ATARS, the aircraft RECCE kit (access door, sensor windows, fairings), the upgraded APG-73 radar, and the data link pod. Functions include mission planning, automatic and manual acquisition of RECCE targets, image data recording, crew-station image review and edit, and data link. Interfaces include those with the Tactical Automated Mission Planning System, ground exploitation stations, and the aircraft carrier environment.
Aerial reconnaissance, defined herein as the exploration of an area with air or space borne imaging systems for intelligence purposes, has a long and interesting history, and continues to thrive today. It has become a critical tool both in establishing military superiority and in keeping world peace. The reason aerial reconnaissance remains so important to the nations of the world is that the means by which it is performed and its information products continue to advance. The ever increasing capabilities of the imaging systems and usage of their products is driven by the extent to which their performance can be characterized and the product information communicated. This paper presents a historical review of the methods used for the evaluation of aerial reconnaissance system performance. The progression from long-used resolution methods, to today's use of imagery interpretability rating scales is described. Particular emphasis is paid to the CORONA years, a time in which much of the metric development work was accomplished.