The Center's corporate vision is the conceptualization, design, testing, and application of a commercially relevant Total Mapping System (TMS). TMS is a conceptual framework that supports comprehensive availability of geographic information, and combines three fundamental components: Spatial Data Acquisition Systems, Spatial Data Processing Systems, and Spatial Data Visualization and Distribution Systems.

In 1995, the Center for Mapping (CFM) proposed the development of a fully digital data acquisition system for large-scale mapping and other precise positioning applications. The proposed system, known as the Airborne Integrated Mapping System (AIMS^TM), will be installed in an aerial platform and will incorporate state-of-the-art GPS/INS positioning and imaging technologies (Charge-Coupled-Device (CCD), infrared, thermal). The planned performance parameters of the system are:

• accuracy of position and orientation of an aerial platform: 5-7 centimeters and ~30 arcsec, respectively in real-time
• essential processing of digital imagery such as histogram equalization and imprinting in real-time
• post-processing of digital imagery to extract feature coordinates at decimeter accuracies

The AIMS^TM initiative is based on experiences gained in the development of the Center's GPSVan^TM technology. The Center has successfully generated private sector interest in the AIMS^TM project by carefully evaluating both the commercial and technical aspects of the initiative. The goal of the CFM and its industry partners is to develop a low-cost (i.e., under $250,000), high-performance hardware prototype that acquires precise spatially referenced data from an aerial platform in real-time.The AIMS^TM project spans the Acquisition and Processing components of TMS. Clearly, AIMS^TM reflects the fundamental forces driving the Geographic Information Infrastructure (GII), and is symbiotic with respect to key developments in spatial data technology (e.g., providing data for Geographic Information Systems (GIS), applying advanced GPS technology, and complementing high-resolution satellite imagery). Although AIMS^TM will not replace conventional aerial mapping in the short run, industry partners have already prepared preliminary market assessments for the following applications emerging from the AIMS^TM project:

• transportation infrastructure
• electric and telephone utilities
• commercial aerial mapping companies
• defense and drug enforcement
• natural resources firms
• emergency response authorities

The aerial platform's orientation and position are provided by a tightly integrated Global Positioning System and Inertial Navigation System (GPS/INS) positioning module. A flexible architecture is being proposed that accommodates the integration of a variety of sensors beyond the CCD cameras, including infrared cameras and laser ranging devices. Off-the-shelf components will be used throughout the system. In addition to developing software to integrate the component technologies, Center personnel are developing advanced algorithms for GPS/INS integration, and image data management in real-time. Because AIMS^TM is an all digital, real-time data acquisition system, it will not require the ground control or film processing required in conventional aerial photography and map making projects. Moreover, AIMS^TM will provide a test environment for the development of advanced image processing techniques such as the automatic identification of conjugate points and the autonomous recognition of features on the ground. A thorough knowledge of state-of-the-art positioning and mapping technologies enables the Center to produce a tightly integrated system using the most advanced techniques and instrumentation. The AIMS^TM design will anticipate and accommodate new hardware and software as they become commercially available.

Image Acquisition

Electronic imaging technology has made great strides during the past two decades. Imaging systems have been widely used in applications ranging from space research to industrial manufacturing, and have also penetrated the consumer market. An important component of these systems is the Charge-Coupled Device (CCD), a solid-state image sensor. This technology has shown enormous progress since its invention and the first implementation of a 96-pixel CCD chip in 1971. The continuous increase in array size has been coupled with improvements in dynamic range and in reduction of readout noise.

Anticipated near-term improvements in digital camera imaging sensors include the following developments:

• We expect to increase our AIMS^TM digital camera to 9K X 9K this year using Lockheed Martin Fairchild's new camera system.
• The radiometric resolution will increase to 16 bits, and color is becoming standard. Experimental results report the availability of a 19 bit dynamic range, that approaches the capability of the human eye
• Further improvements in cameras with intelligent interface and control capabilities are expected. Built-in image compression will be available soon
• Interline, electronically shuttered, full-frame CCD sensors will dominate the digital video market with substantial benefits to mobile imaging technology
• HDTV will have a serious impact on the mobile image acquisition technology by offering quality, high-resolution images in a fully digital environment

The flexible architecture of AIMS^TM allows application customizing, including specialized post-processing tools, interfacing with multiple cameras of the same or different resolution, and different data capture resets, and real-time image processing capabilities. The potential applications include airborne stereo image capture, semiautomatic DEM extraction for orthophoto production, land use classification analysis with fast turnaround times based on georeferenced imagery, and high-accuracy land-based mobile mapping, where cm-level feature positioning is required. The proposed system, combined with servo-controlled cameras equipped with narrow viewing angle lens, can be used for target tracking in civilian and military applications.

AIMS^TM's modular architecture also allows the replacement of components as technology evolves, resulting in increased performance without affecting the basic system design. State-of-the-art technology, in particular, rapidly advancing processor and storage technology, can be introduced incrementally into the system as it becomes available. The prototype will be designed to employ commercial off-the-shelf products to ensure the most cost-effective implementation, provided this does not compromise the system's performance.

GPS and INS Positioning

In the last ten years the accuracy of dynamic differential GPS positioning has improved from several meters to a few centimeters for baselines in excess of 100 km. The accuracy of differential GPS positioning depends on the baseline length and the type of receiver (single or dual frequency). For instance, good quality single frequency GPS receivers are capable of providing positioning accuracy at the meter level over baselines of 20-40 km. Specially designed single frequency GPS receivers, employing narrow correlators and multipath estimating delay lock loops, provide accuracies at the 0.15 m level after 2-3 minutes of filtering for baselines of 10km. Ionosphere and ephemeris errors decrease this accuracy to the 0.50 m level as the baseline length increases to 40-50 km. High accuracy (cm-level) GPS dynamic positioning for baselines of 100 km or more requires dual frequency GPS receivers and ambiguity resolution with carrier phase positioning, augmented by the proper modeling of the atmospheric and the multipath effects. The differential troposphere and ionosphere, together with the double-differenced multipath, may have a significant effect on the positioning accuracy if not properly modeled. The GPS component of the integrated system consists of a stationary (base) GPS receiver and a rover GPS receiver, mounted on the moving platform. For high-accuracy positioning, both the base receiver and the rover receiver should be dual frequency.

On-The-Fly (OTF) ambiguity resolution is independent of receiver motion, and can be used when one or both base and rover receivers are moving. As a result, this technique is ideally suited for real-time cm-level positioning. With dual frequency GPS receivers, OTF ambiguity resolution is very fast (ambiguities are resolved within a few epochs of time). The OTF method establishes a search space using differential pseudorange positioning. The correct solution within the search space is identified using least-squares search or ambiguity covariance methods. Robust OTF ambiguity resolution techniques for long baselines are still under development.

Inertial Navigation Systems (INS) provide self-contained independent means for three-dimensional positioning with high short-term accuracy. The INS accuracy degrades over time, due to the unbounded positioning errors caused by the uncompensated gyro and accelerometer errors affecting the INS measurements. The degradation is much faster for low-cost INS systems. INS provides high-accuracy three-dimensional positioning when the GPS positionig is poor or unavailable over short periods of time (e.g., due to poor satellite geometry, high electromagnetic interference, high multipath environments, or obstructed satellite signals). In addition, the INS system provides much higher update positioning rates compared to the output rate conventionally available from GPS. High-accuracy INS units, e.g., Litton LN-100, provide attitude determination at the level of 1-3 arcmin. Updating INS with GPS positions will provide the System’s orientation at ~30 arcsec level. Integration of GPS with INS bounds the positioning errors of the inertial system with the uniform positioning errors affecting the GPS system. These errors depend on the systematic and random errors affecting the GPS measurements, as amplified by the satellite geometry. Using the GPS positions, the GPS/INS integration filter can estimate the error states affecting the INS measurements. These error state estimates are used to calibrate the INS system on a continuous basis. The high accuracy of the INS system over short periods of time allows correction of undetected cycle-slips affecting the GPS measurements and makes it possible to perform faster and more robust OTF ambiguity resolution.

There are two basic GPS/INS integration schemes: loosely and tightly coupled mode. In the loosely coupled mode, the GPS receiver and the INS are treated as separate navigation systems. The GPS receiver contains a filter, which processes the raw GPS observables and supplies a position, velocity, and time solution. The INS implements its navigation/attitude algorithms to give a position, velocity, and attitude. An integrated Kalman filter is then applied to combine the GPS and INS solutions. In the tightly coupled mode, however, only a single Kalman filter is applied to process both sets of sensor data: raw GPS code/phase observations and INS measurements. New tightly coupled GPS/INS integration architecture, where INS navigation computation module is implemented to process raw IMU and GPS measurements, is currently under development at the CFM. A Kalman filter, which uses a linear error model, is designed to combine the INS resolution with the raw GPS observations to estimate errors in INS position, velocity, and attitude data, as well as errors in inertial and GPS measurements. Estimated INS errors are fed back to the computation module to improve INS accuracy. The Kalman filter estimates directly position, velocity, and attitude parameters (instead of their errors), as well as errors in inertial and GPS measurements. The obvious advantages of tightly coupled GPS/INS systems include:

• Cycle-slip detection and OTF ambiguity resolution are supported by INS measurements and are performed on selected update steps
• Processing of raw GPS data makes it possible to optimally calibrate both INS and GPS errors through optimal estimation techniques

Airborne Integrated Mapping System Structure

The proposed Airborne Integrated Mapping System (AIMS^TM), is a hardware and software integration of GPS, INS, and Digital Imagery technologies in a mobile platform. The system can be augmented by the laser system to support the Image Acquisition module. The GPS time is used to synchronize the position information with the measurements from the other sensors. The proposed AIMS consists of three major components:

• The “Tightly Coupled GPS/INS Kalman Filter” uses the tightly coupled GPS/INS architecture to process both raw IMU and GPS measurements. The strapdown position, velocity, and attitude computation algorithms are embodied in the nonlinear state equations of the Kalman filter. GPS double-difference computation, cycle-slip detection, and OTF ambiguity resolution algorithms are also implemented in the filter. Numerically stable filtering algorithms will be applied. The tightly coupled GPS/INS integration architecture provides the best navigation accuracy and has good stability characteristics
• The “Image Acquisition & Control Component” manages acquisition and storage of the images, and controls operation of the cameras
• The “Control & Display Component” provides user interface and performs high level control of other components

Software Design and Development

The proposed system is divided into the following three modules, implemented as separate software components:

• GPS/INS Positioning Software Component
• Image Acquisition and Control Component
• Control and Display Component

The individual components will be designed and developed by systems/software engineers with specific expertise in the areas of GPS, INS, and Computer Vision and with significant C/C++ programming experience. The software will be designed as independent modules using an object-oriented design approach, and developed using C++, an object-oriented programming language. The target platform will be an Intel pentium computer running Windows NT^TM. A Graphical User Interface (GUI) will provide a user-friendly interface for control, and to display information for the system.

The software development will be in C++, an object-oriented programming language, in which software is developed in a modular fashion using software classes or objects. A class is essentially an independent software module with a clear, well-defined interface independent of the details of the implementation. Additional or changed functionality can be added to a class using inheritance, without having to alter the existing class. C++ will enhance the efficiency of the software development and will ensure the robustness of the software product. The use of object-oriented programming has been recognized as having significant benefits in the areas of software maintenance and reusability.

A commercially available version control system will be used to store the software components as they are developed and enhanced. Such a system is currently in use at the Center for Mapping and is an essential tool in the development of software involving more than one programmer. It allows the software developers to develop and debug different software modules of a single project while minimizing the effect on each other's efforts. In addition, this system allows previous versions of a project to be recreated and allows different configurations of modules to be updated whenever a common component is altered. Third-party software will be used whenever possible to ensure the most cost-effective product, providing this does not compromise the intended portability of the product. At this time, third party C++ software is used to facilitate the development of user interfaces to software products, and also for generic numerical operations such as matrix manipulation.