The technologies utilized in force protection in the Air Force military environment have focused on applying commercial off-the-shelf technology to the maximum extent possible. This change from uniquely developed systems presents an opportunity where Air Force and Department of Defense technology efforts can be more widely applied in law enforcement contexts. The organizational involvement and extensive dedicated test resources for Air Force Security Force are explained. Providing details on product evaluations already conducted and the methodology employed can benefit a segment of the law enforcement community.

A scanning magneto-resistance microscope was developed that allows for high resolution imaging of magnetic tapes and digital media. By using second harmonic detection to remove thermal anomalies we are able to image sufficient lengths of tape for authentication purposes and for data recovery from damaged samples. This allows for high contrast images and direct conversion of the scanned information into originally recorded analog audio waveforms or digital data.

Automatic face recognition algorithms have the potential to impact several applications including surveillance, human augmentation, multimedia indexing and retrieval, and authentication. In this paper, a technique to retrieve images by visual appearance similarity is applied to the problem of face recognition. The framework for representing and computing similarity is based on the design and use of multi-scale Gaussian differential features (MGDFs) as appearance features. In the first part of this paper, the relevance of MGDFs as appearance features and an algorithm to deduce global similarity is developed. In the second part of this paper, multi-scale representations are applied to face recognition. Results from experiments on standard test collections tested in this paper indicate that at least 96% recognition accuracy is obtained, and when compared with other techniques, the MGDF based representation yields comparable or better results. The MGDF based technique is very general; it was originally developed for global appearance similarity retrieval in heterogeneous images, and has been applied to retrieve similar textures, trademarks, binary shapes and heterogeneous gray-level collections.

A growing number of law enforcement applications, especially in the areas of border security, drug enforcement and anti- terrorism require high-resolution wide area surveillance from unmanned air vehicles. At the University of Massachusetts we are developing an aerial reconnaissance system capable of generating high resolution, geographically registered terrain models (in the form of a seamless mosaic) in real-time from a single down-looking digital video camera. The efficiency of the processing algorithms, as well as the simplicity of the hardware, will provide the user with the ability to produce and roam through stereoscopic geo-referenced mosaic images in real-time, and to automatically generate highly accurate 3D terrain models offline in a fraction of the time currently required by softcopy conventional photogrammetry systems. The system is organized around a set of integrated sensor and software components. The instrumentation package is comprised of several inexpensive commercial-off-the-shelf components, including a digital video camera, a differential GPS, and a 3-axis heading and reference system. At the heart of the system is a set of software tools for image registration, mosaic generation, geo-location and aircraft state vector recovery. Each process is designed to efficiently handle the data collected by the instrument package. Particular attention is given to minimizing geospatial errors at each stage, as well as modeling propagation of errors through the system. Preliminary results for an urban and forested scene are discussed in detail.

This paper introduces our Dynamically Reconfigurable Vector technology, and discusses the significant benefits afforded by this technology to law enforcement applications employing unattended sensor networks.

We have developed and tested an algorithm that can compare striation marks that are acquired with a standard camera and sidelight as well as 3D-information acquired with structured light.

This document describes the law enforcement charter and activities of the United States Customs Service and the internal technology organization that supports it, the Applied Technology Division. The enforcement activities of Customs include interdiction, outbound anti-smuggling, investigation and surveillance, processing of documentation and data, and detection of drugs and other contraband. An overview of the various technologies applied in support of these activities over the past 25 years is provided. Additionally, technologies proposed for implementation in the future are discussed.

The amazing growth of the Internet, both in size and in influence on our society has lead to increased risks of its exploitation by criminal and terrorist groups. As of now, this exploitation has been relatively limited, at least with respect to the likely activity in the years to come. There is a need to act now to develop and put into place intelligence methodologies to aid analysis of Internet-based national security and criminal threats and to augment existing Internet security practices. These methodologies cannot be a purely technically based, or the true societal significant of Internet activity will be overlooked. They cannot be a purely localized activity, or the divergent needs of various regions and organizations will not be represented. They cannot be simply responsive to incidents, such as viruses or system attacks, or the advantage will remain with the intruders. They cannot be centrally controlled or performed, or the need for rapid `Internet- speed' response will not be met. Internet security threats are distributed, ongoing and multifaceted, so the strategy for dealing with them must be distributed ongoing and multifaceted.

This paper describes the Advanced Packet Vault, a technology for creating such a record by collecting and securely storing all packets observed on a network, with a scalable architecture intended to support network speeds in excess of 100 Mbps. Encryption is used to preserve users' security and privacy, permitting selected traffic to be made available without revealing other traffic. The Vault implementation, based on Linux and OpenBSD, is open-source.

Many private and public networks are based on network switching technologies. However, switched networks present a number of challenges to intrusion detection equipment. These challenges include limited visibility of network flows at the edges of the network, high-speed packet processing, and highly-aggregated flows in the core. In addition, switched networks typically implement protocols specific for Layer 2 functions, such as connection establishment and connection routing, which can be attacked to deny service to higher layer protocols and applications. Since these attacks cannot be detected by Internet Protocol intrusion detection equipment. Layer 2 intrusion detection is required. This paper describes an approach for performing intrusion monitoring in switched, Layer 2 networks, specifically, Asynchronous Transfer Mode networks.

The U.S. Public Telephone Network (PTN) is a massively connected distributed information systems, much like the Internet. PTN signaling, transmission and operations functions must be protected from physical and cyber attacks to ensure the reliable delivery of telecommunications services. The increasing convergence of PTNs with wireless communications systems, computer networks and the Internet itself poses serious threats to our nation's telecommunications infrastructure. Legacy technologies and advanced services encumber well-known and as of yet undiscovered vulnerabilities that render them susceptible to cyber attacks. This paper presents a taxonomy of cyber attacks on PTNs in converged environments that synthesizes exploits in computer and communications network domains. The taxonomy provides an opportunity for the systematic exploration of mitigative and preventive strategies, as well as for the identification and classification of emerging threats.

National-scale critical infrastructure protection depends on many processes: intelligence gathering, analysis, interdiction, detection, response and recovery, to name a few. These processes are typically carried out by different individuals, agencies and industry sectors. Many new threats to national infrastructure are arising from the complex couplings that exist between advanced information technologies (telecommunications and internet), physical components (utilities), human services (health, law enforcement, emergency management) and commerce (financial services, logistics). Those threats arise and evolve at a rate governed by human intelligence and innovation, on `internet time' so to speak. The processes for infrastructure protection must operate on the same time scale to be effective. To achieve this, a new approach to integrating, coordinating and managing infrastructure protection must be deployed. To this end, we have designed an underlying web-like architecture that will serve as a platform for the decentralized monitoring and management of national critical infrastructures.

Over the last year, network-based intrusion have increased exponentially, due to the popularity of scripted or automated attack tools. This increase in intrusions has rekindled interest in honeypot systems, which can be used to trap and decode the attack methods used by the black hat community. This paper will review the current state of honeypot technology as well as describe methods of deploying entire honeypot networks.

A safe communication infrastructure is critical to maintaining the prosperity and pre-eminence of the United States in the 21st century. Intrusion detection systems (IDS) help identify and respond to malicious activity targeted at computing and networking resources. In this paper, we shall first give an overview of intrusion detection concepts and taxonomy. Next, we introduce and discuss several commercial and public-domain IDS's available today. Then, we describe recent developments in conventional intrusion detection: distributed, modular system which include both anomaly and misuse detection. Lastly, we give a peek at the new breed of pro-active, preventative tools. These so-called Delphic tools identify the threats and risks in the very early attack stages; ideally, even before an attack takes place. Examples mentioned are quantitative security risk analysis and Bayesian multiple hypothesis tracking.

Unobtrusive detection of concealed weapons on persons or in abandoned bags would provide law enforcement a powerful tool to focus resources and increase traffic throughput in high- risk situations. We have developed a fast image scanning 94 GHz radar system that is suitable for portable operation and remote viewing of radar data. This system includes a novel fast image-scanning antenna that allows for the acquisition of medium resolution 3D millimeter wave images of stationary targets with frame times on order of one second. The 3D radar data allows for potential isolation of concealed weapons from body and environmental clutter such as nearby furniture or other people. The radar is an active system so image quality is not affected indoors, emitted power is however very low so there are no health concerns for operator or targets. The low power operation is still sufficient to penetrate heavy clothing or material. Small system size allows for easy transport and rapid deployment of the system as well as an easy migration path to future hand held systems.

A method of detecting weapons concealed under clothing using passive millimeter wave imaging is described. The optical properties of clothing are discussed and examples given of the spectral reflectivity and transmission. The transmission tends to be constant from 60 to 150 GHz above which it decreases for some clothing materials. The transmission of a cotton T-shirt is typically 95% and of a leather jacket up to 85% at lower frequencies. A model is presented for calculating the contrast of a metallic concealed weapon when hidden under clothing and it indicates contrasts as large as 200 K can be realized outdoors. The advantages of real time over static frame imagery are discussed. It is concluded that real time imagery offers considerable advantages as weapons can be very varied in size, position and orientation and movement offers vital clues to the human observer which aid the recognition process.

A handheld, battery-operated prototype of a concealed weapon detector has been built and tested. Designed to detect both metallic and non-metallic weapons, the sensor utilizes focused ultrasound (40 kHz frequency) to remotely detect concealed objects from beyond arm's length out to a range of about 12 feet (4 meters). The detector can be used in prison settings, by officers in the field to allow for stand-off frisking of suspects, and to supplement security at courthouse entrances and other monitored portals. The detector emits an audible alarm (with provision for an earphone jack) as well as a visible light-bar indicator when an object is detected. A high intensity aiming light, with momentary switch, allows the user to accurately determine the location of the concealed object. Current efforts are aimed at increasing the probability of detection, reducing the false-alarm rate, and extending the range of detectability out to 20 feet. Plans for accomplishing these tasks will be presented together with data showing the effective range and probability of detection for the present system.

A MMW sensor developed by Trex Enterprises generates image data of a person hiding a gun under his clothing at a distance of 27 feet. The goal of this research was to develop an algorithm that would automatically recognize the weapon. Tracking, segmentation, and recognition procedures were designed and successfully applied to the data.

A handheld through-the-wall surveillance system is being developed for use by law enforcement and military personnel. The system utilizes high-power ultrasonic transducers to detect and locate stationary or moving persons inside metallic and non-metallic walled enclosures. Design details are presented with proof-of-concept data and analyses. The sensor system is being designed to operate as a handheld monitor with a near real-time user display of the location, in range and azimuth, of each detected individual. Preliminary test data include wall penetration/sensitivity, locating accuracy, and probability of detection. Applications of this technology include detecting and locating unconscious, sleeping, tightly bound, or otherwise stationary persons, as well as moving persons, inside a closed room. The sensor should also prove useful in border patrol applications for inspecting truck trailers and shipping containers at points of entry.

Raytheon Electronics Systems, under contract from the DARPA Advanced Technology Office, has designed, fabricated and delivered the Modular Miniature Imaging Sensor (M²IS). M²IS is a rifle- or tripod-mountable system that integrates a high-performance multispectral sensor with an eyesafe laser rangefinder and a digital compass. A cooled 480 X 640 InSb focal plane array and dual-FOV reflective optics provide capability to acquire and identify targets at ranges of several kilometers. The LRF and compass facilitate reporting target location. M²IP provides the law enforcement officer an integrated surveillance and targeting system that consumes less than 6.5 W and weighs less than 7.5 lbs. This paper describes measured performance and capabilities of the system.

Low power and low cost are primary requirements for an imaging infrared camera serving law enforcement applications. These include handheld, vehicle and helmet mounted systems for search and surveillance applications. In this paper, a 120 X 160 amorphous silicon (a-Si) microbolometer-based uncooled infrared camera technology offering a low cost, low power solution to infrared surveillance for UGS applications is presented. A 120 X 160 micro infrared camera has been demonstrated which exhibits a noise equivalent temperature difference sensitivity approximately 50 mK using f/1 optics and approximately 80 mK using f/1.2 optics. This sensitivity has been achieved without the use of a thermoelectric cooler for array temperature stabilization thereby significantly reducing the power requirements.

In this paper we will describe advances in microbolometer uncooled thermal imaging sensor technology as they apply to law enforcement applications. Improvements in sensor performance that will be described include: (1) reduced pixel pitch, (2) increased spatial resolution, (3) increased thermal sensitivity, (4) reduced electrical power, and (5) reduced size. Since cost considerations dominate many, if not most, potential law enforcement applications, microbolometer sensor cost issues will be addressed in terms of current and projected cost trends. In addition to the use of theoretical considerations in describing microbolometer technology advancements currently being made or planned, examples of actual improvements, in the form of real imagery and/or actual performance measurements, will be provided in the paper. Finally, we will look at those areas of law enforcement that are most likely to benefit from the application of microbolometer uncooled thermal imaging sensor technology. These include: (1) surveillance sensor systems, (2) unattended sensor systems, (3) mobile sensor systems and platforms, and (4) gunfire localization and counter sniper systems.

An increasingly important need today is to guard against terrorist attacks at key locations such as airports and public buildings. Liquid explosives can avoid detection at security checkpoints by being concealed as beverages or other benign liquids. Magnetic resonance (MR) offers a safe, non-invasive technology for probing and classifying the liquid contents inside sealed non-metallic containers or packages. Quantum Magnetics has developed a Liquid Explosives Screening System or `Bottle Scanner' to screen for liquid explosives and flammables, described at an earlier SPIE conference in 1996. Since then, the Bottle Scanner's performance has been significantly improved by the incorporation of neural network-based liquid classification. Recently we have shown that the incorporation of additional discrimination parameters can further enhance liquid classification. In addition to screening for explosives and flammables, the Bottle Scanner can be effective against chemical agents, many of which contain fluorine or phosphorous, both of which have MR signatures. Finally, we have evidence that the Bottle Scanner may also be able to detect narcotics dissolved in beverages, one of the methods used to smuggle narcotics across international borders. The development of the Bottle Scanner has been funded by the Federal Aviation Administration.

This paper describes a prototype sensor system for detection and 3D tracking of bullets and other small projectiles. The intended purpose of the system is to rapidly locate a sniper to a few meters accuracy at ranges to 1 km in three dimensions. The system detects and tracks a single bullet, and based on the measured 3D trajectory, backtracks to the sniper location. Details of the system are describe including optics, infrared camera, scanning system, laser ranging system, computer control and electronics, and data reduction algorithm. The system has been field tested against bullets, and has been shown to locate a sniper to a few meters accuracy at 500 meters range. Plans for improving tracking performance are also described.

This paper describes the work related to the development of a maneuverable ballistically deployed sensor system. These maneuverable sensors provide the user with remote surveillance of building interiors. The system is designed to be compatible with deployment using an M-203 grenade launcher and to survive the anticipated 10,000 Gs that occur during impact with, (and simultaneous attachment to) the wall. Subsequent to deployment/attachment, an imaging sensor is maneuvered to an unobtrusive location at the upper edge of the target window and the video signal is transmitted to the remote user.

Accurate Automation Corporation has completed the conceptual design of a mortar launched air vehicle system to perform close range or over-the-horizon surveillance missions. Law enforcement and military units require an organic capability to obtain real time intelligence information of time critical targets. Our design will permit law enforcement to detect, classify, locate and track these time critical targets. The surveillance system is a simple, unmanned fixed-winged aircraft deployed via a conventional mortar tube. The aircraft's flight surfaces are deployed following mortar launch to permit maximum range and time over target. The aircraft and sensor system are field retrievable. The aircraft can be configured with an engine to permit extended time over target or range. The aircraft has an integrated surveillance sensor system; a programmable CMOS sensor array. The integrated RF transmitted to capable of down- linking real-time video over line-of-sight distances exceeding 10 kilometers. The major benefit of the modular design is the ability to provide surveillance or tracking quickly at a low cost. Vehicle operational radius and sensor field coverage as well as design trade results of vehicle range and endurance performance and payload capacity at operational range are presented for various mortar configurations.

As the size, weight, and cost of miniature electronics have continued to decrease, the feasibility of developing portable, low-cost, unmanned surveillance systems has increased tremendously in recent years. Today, manufacturers are developing small, unmanned surveillance aircraft for both military and civilian applications. However, many military, paramilitary, and civilian applications require the surveillance platform to remain stationary over a given area for extended periods. Furthermore, most current unmanned surveillance aircraft are too large to be carried by the operators in the field and require special support equipment. This paper describes the preliminary design and analysis of a lighter-than-air vehicle based surveillance system that is lightweight, portable, and will allow extended station keeping over a desired position. This new system can be applied to military surveillance and reconnaissance, border patrol/law enforcement, search and rescue, surveying, and photographic mapping.

A system has been developed for delivering and attaching a sensor payload to a target using a standard 40-mm grenade launcher. The GLIMPS projectile is intended to be a general purpose delivery system for a variety of sensor payloads including visual, acoustic, and chemical sensors. The GLIMPS projectile flight characteristics are similar to existing 40-mm rounds, with a useful range of up to 300 m. The projectile incorporates an attachment mechanism, a shock mitigation system, a power source, and a telemetry system for transmission of sensor data at up to 1/4 mile range. A second design is also being considered. It is a small taggant projectile that uses an adhesive to attach a tracking transmitter or other small payload to a vehicle at up to 50 m range. While initially developed as a military system, both projectiles can be used to enhance law enforcement operations.

Past studies have shown that speaker identification (SID) algorithms that utilized LPC cepstral feature and a vector quantization classifier can be sensitive to changes in environmental conditions. Many experiments have examined the effects of noise on the LPC cepstral feature. This work studies the effects of co-channel speech on a SID system. It has been found that co-channel interference will degrade the performance of a speaker identification system, but not significantly when compared to the effects of wideband noise on an SID system. Our results show that when the interfering speaker is modeled as one of the speakers within he training set, it has less of an effect on the performance of an SID system than when the interfering speaker is outside the set of modeled speakers.

This paper presents a secure, Internet-enabled, third wave knowledge management system. TheResearchPlace^TM, that will facilitate a collaborative, strategic approach to analyzing public safety problems and developing interventions to reduce them. TheResearchPlace, currently being developed under Government and private funding for use by the National Cancer Institute, Federal agencies, and the Defense Advanced Research Project Agency, will augment Geographic Information Systems and analytical tool capabilities by providing a synergistic workspace where teams of multidisciplinary professions can manage portfolios of existing knowledge resources, locate and create new knowledge resources that are added to portfolios, and collaborate with colleagues to leverage evolving portfolios' capabilities on team missions. TheResearchPlace is currently in use by selected alpha users at selected federal sites, and by the faculty of Howard University.

This research began due to federal inmates abusing their telephone privileges by committing serious offenses such as murder, drug dealing, and fraud. On average, about 1000 calls are made per day at each federal prison with a peak of over 4000. Current monitoring capabilities are very man- intensive and only allow for about 2-3% monitoring of inmate telephone conversations. One of the main deficiencies identified by prison officials is the need to flag phone conversations pertaining to criminal activity. This research looks at two unique voice-processing methods to detect phone conversion pertaining to criminal activity. These two methods are digit string detection and whisper detection.

Voice Stress Analysis (VSA) systems are marketed as computer-based systems capable of measuring stress in a person's voice as an indicator of deception. They are advertised as being less expensive, easier to use, less invasive in use, and less constrained in their operation then polygraph technology. The National Institute of Justice have asked the Air Force Research Laboratory for assistance in evaluating voice stress analysis technology. Law enforcement officials have also been asking questions about this technology. If VSA technology proves to be effective, its value for military and law enforcement application is tremendous.

The problem of information and knowledge management in the knowledge intensive and time critical environment of law enforcement has posed an interesting problem for information technology professionals in the field. Coupled with this challenging environment are issues relating to the integration of multiple systems, each having different functionalities resulting in difficulty for the end user. COPLINK offers a cost-efficient way of web enabling stovepipe law enforcement information sharing systems by employing a model for allowing different police departments to more easily share data amongst themselves through an easy-to-use interface that integrates different data sources. The COPLINK project has two major components: COPLINK Database Application and COPLINK Concept Space Application.

Key advances in microelectronics have enabled the development of an entirely new class of microsensor devices for surveillance and monitoring. These sensor devices are both ground based and airborne and feature large scale cooperative networking, geo-location and tracking algorithms, low power for long unattended operation, and robust communications for data retrieval.

The current trend to develop low cost, miniature unattended ground sensors will enable a cost-effective, covert means for surveillance in both urban and remote border areas. Whereas the functionality (e.g., sensing range and life in the field) of these smaller UGS (i.e., acoustic, seismic, magnetic, chemical or biological) may be limited due to size and cost constraints, a network of these sensors working cooperatively together can provide an effective surveillance capability. A key factor is the ability of these sensors to work cooperatively to achieve a `collective' functionality that can meet the surveillance objective. This paper describes results of using target identification (ID) features (i.e., the ID feature space of the target) to improve the tracking of closely spaced targets (i.e., the kinematic space of the targets). A Multiple Level Identification (MLID) approach was used to determine and maintain confidences for multiple target identifications for each target. These confidences were incorporated into the processing of kinematic data (i.e., target bearing reports) to improve the tracker's estimated position of the target's location. Results describing the effectiveness of using MLID on target tracking performance are reported using simulated target trajectory and ID data.

This paper describes an unattended surveillance system that is used to monitor border crossings, remote airfields, choke points, and any other remote area not suitable for continuous human presence. The system employs intelligent acoustic/seismic unattended ground sensors, two-way radio communication, and a low light video camera. The acoustic/seismic sensors provide long-range target detection, target classification, and angular bearing to the target. The camera is cued by the acoustic/seismic sensors and takes a picture of the target for visual confirmation. All information, including the picture, is transmitted via radio to a monitoring post.

In Carnegie Mellon University's CyberScout project, we are developing a network of mobile and stationary sentries capable of autonomous reconnaissance and surveillance. In this paper, we describe the cooperative perception algorithms and mission planning necessary to achieve this task, including sensor-to-sensor target handoff methods and an efficient decentralized path-planning algorithm. These methods are applied to a typical law enforcement application, a building stakeout scenario.

In this paper, we present our system infrastructure design and a cost function based geographical self-routing algorithm for networking reconfigurable smart sensors. The algorithm allows for the sensors to automatically negotiate in a geographical radial topology relative to a central location, utilizing other sensors as routes or hops for forwarding information to this central location. A number of these sensors are deployed in the field and performance measurements for routing times are analyzed and presented.

We consider data management on ad hoc networks of sensing and processing nodes. We describe the construction of simple nodes from off the shelf components (PC 104 single board computers with flash memory, video capture cards and 802.1 lb wireless interfaces). We describe a Java interface to controlling these nodes and accessing images and image processing algorithms. We demonstrate target tracking across nodes and the potential for heterogeneous sensor types.

The Internetted Unattended Ground Sensor system being developed for DARPA is a distributed ground sensor system intended for autonomous surveillance activities, consisting of a multitude of sensor types, giving it all-weather detection, tracking, and classification capabilities. In addition to its military applications, it would seem to have potential law enforcement uses as well, such as border surveillance and facility monitoring. This paper provides a description of this system so that law enforcement agencies can determine whether it might help solve certain surveillance problems.

This paper describes the design of a small sensor that can detect and track different targets, namely vehicles, personnel and sniper fire. Building on previous work with portable sensors using both seismic and acoustic transducers, the goal was to design a sensor with similar functions that fits in a small projectile deployed from a standard M203 grenade launcher. We discuss methods to reduce weight, size, and power consumption. We use a shell-within- shell design in which the instrument separates from the outer body at the apex of its flight. After the separation, spring loaded arms unfold from the inner body. The unfolding arms serve multiple purposes: to hold the acoustic transducers on the periphery of a small disk with a measurement aperture larger than the shell (about 5 times the shell diameter), to stabilize the sensor in flight, and to act as a ground plane for radio transmission. An example of a hand-emplaced version using the same processor is also discussed.

Consistent with the underlying long-term objectives of the development of the Unattended Ground Sensors program, we are developing a new planted ground sensor platform based on Micro-Electro-Mechanical Systems which offers magnetic, chemical and acoustic detection. The envisioned micro-system will be low-power and low-cost, and it will be built around a single type of microstructure element integrating a monolithic optical system and electronics package.

Quantum Magnetics (QM) has developed a sensing array using small and lightweight magnetoresistive (MR) sensors. These sensors, which operate at room temperature with high sensitivity and wide bandwidth, provide new operational performance capabilities. The wide bandwidth makes them ideal for both passive and active detection techniques. Using a DSP-based electronics developed by QM, we have been able to operate these sensors with an unprecedented noise performance at low frequencies. Recent tests using an MR room temperature gradiometer show that its resolution equals that of a fluxgate room-temperature gradiometer we have previously developed. These results represent an important development for both attended and unattended ground sensor applications since MR sensors cost about ten times less than fluxgate sensors.

Recent advances in electronic miniaturization have facilitated the design and development of a deployable, stand-alone sensor for battlefield RF (radio frequency) MASINT (measurement and signature intelligence). Recent results of a Phase I effort to assess battlefield RF signatures and compare sensor sensitivity, size, and number for optimal coverage will be presented. An RF sensor suite is being designed that will be networked for robust and redundant data gathering and wireless, stand-alone operation. The RF sensor will essentially function as a swept-frequency spectrum analyzer, measuring frequency content, amplitude and modulation characteristics over predetermined (or user programmable) bandwidths of interest. Recommendations and preliminary schematics for a compact, ruggedized, low-power RF sensor will be presented together with a design for a low-power centralized wireless network for data transfer and processing. The RF sensor technology developed in this effort will have a predominantly military application but should also find use in security and surveillance applications.

The Acoustic Signal Processing branch of the U.S. Army Research Laboratory has demonstrated, using the output of microphone arrays of battlefield acoustic sensors, that harmonic lines are effective spectral features for the target identification of acoustic signatures from ground vehicles such as tanks and trucks. However, battlefield acoustic target identification is extremely challenging because of the variations in the signature caused by the environmental effects on the acoustic propagation, vehicle operating characteristics (e.g. gear, load) and target dynamics such as range, velocity, and aspect. The utilization of air acoustic propagation models to normalize the spectral features can be used to mitigate a significant portion of this variability for robust battlefield acoustic target identification. The integration of acoustic propagation models is a new approach, and potentially represents a very significant improvement to current target recognition methods.

Battelle was tasked by the National Institute of Justice to both assess the state-of-the-art in bomb disposal robot technologies and to identify research and development needs. A well-documented market-driven approach was taken to define the bomb disposal technician needs along with the shortcomings of existing systems. This paper will summarize the findings of that report.

DARPA's Tactical Mobile Robot program includes a Throwable Robot (Throwbot) designed to be thrown into buildings, then teleoperated for surveillance purposes. Use by ground troops imposes significant size and weight limits, as does the requirement that it survive ballistic delivery. The current program stresses the state of the art in robotics and packaging, but further challenges exist. Future Throwbots would benefit from significant increases in autonomy, to deal with RF communication difficulties in buildings and to allow simultaneous operation of multiple vehicles by one person. This paper describes both currently planned and advanced autonomous capabilities.

Marsupial robots are a type of heterogeneous mobile robot team. A mother robot transports, supports, and recovers one or more daughter robots. This paper will cover the marsupial robot concept, the application of law enforcement, and recent results in collaborative teleoperation for the related task of urban search and rescue.

We conducted a web-based survey to establish law enforcement robotics needs for applications that extend beyond explosive ordnance disposal. The survey addressed scenarios and tasks where a robot would be used if available, and the tools, features and parameters deemed most important to carry out those tasks. We present in this paper the results of the survey and summarize current robotics research and development efforts by various segments of the Department of Defense that could potentially help meet those law enforcement needs. We also provide a recommended course of action to the Department of Justice for the development of these robotics capabilities.

Video generates a rich set of image information and often the useful information is only a very limited set of the available information. Another well-known fact is that visually reviewing of long video recording is a time demanding task. In combination with the continuously increasing number of video surveillance systems, this leads to an increasing need for automated analysis of long image sequences. The goal for this work is to develop and evaluate a method for automatic detection and tracking of events recorded onto a surveillance video, such as appearance of persons or vehicles in a surveyed area, to evaluate the usefulness for forensic applications and real time applications.

Observation of long sequences of video images in surveillance applications may encounter several problems due to camera motion or rotation, unexpected size and speed of objects, variation of color due to sunshine and shadowy areas. Robust tracking algorithms are needed to compensate for the variations of different recording conditions. In this paper we evaluate the detection probability of our tracking algorithm with ROC curves and with synthetic degradation methods. Recorded experimental multi-sensor data is used to compare the accuracy in different spectral bands. Moving object detection in a guarded area can produce many false alarms due to the moving environment such as trees and bushes, birds and animals. By applying tracking and classification, false alarms can be reduced avoiding unnecessary recordings and preventing the displacement of guards. Track speed, size, direction and range (distance to camera) are calculated. The objects are classified roughly into classes as person, vehicle, and fast moving object or simply as moving object. The results of the algorithm applied to the experimental data and the algorithm evaluation are presented.

In this paper we describe an Advanced Digital Video Surveillance system based on TASC-developed object behavior based video analysis and indexing prototype. The advantages of using video analysis in surveillance and physical security applications are twofold. Firstly, the ability to automatically analyze the surveillance video contents facilitates timely detection of events that require immediate attention. Secondly, the amount of video to be archived can be reduced considerably by recording only the portions of video that include behaviors and events of interest from vast amount of surveillance data being collected everyday. Our object-behavior and event based indexing paradigm for video data treats an identifiable object behavior, action or event as the basic indexing unit facilitating efficient querying and report generation as well as derivation of statistical information about the behavior patterns over periods of time. We describe our methodology and present preliminary results in near real time behavior and event detection.

In order to clarify detail of objects in shadows, we present a combination of frame averaging and homomorphic filtering. Multi-frame averaging allows exploitation of temporal redundancy in video to reduce noise present in dark scenes. One common method to enhance images suffering from low contrast is gain. Gain has the side effect of saturating brighter areas of the scene. Homomorphic filtering allows the effects of differing illumination conditions to be equalized across a scene. This is accomplished by applying a spatial high-pass filter to the natural log of pixel values. By following frame averaging with homomorphic filtering, details obscured by shadows may be clarified without losing important foreground information. Noise reduction through frame averaging greatly improves the results of homomorphic filtering, which usually amplifies Gaussian white noise.

Time-multiplexed video recordings must be demultiplexed for effective analysis by investigators: a task often made difficult by the proprietary nature of the recording equipment. In addition to a lack of standards, some systems may record scenes from different cameras with a time-varying sequence. To address this issue, a content-based image analysis technique for extracting a specified camera view from a multiplexed video sequence is presented. A single frame from the specified view is used as a reference to which subsequent frames are compared. Frames matching the reference within a specified similarity threshold are passed on to the demultiplexed output. An optional region of interest may be used to select an area of the scene that is unique to that camera view but relatively consistent over time. The software implementation of this technique allows multiple camera views to be demultiplexed simultaneously from a digitized stream of multiplexed video. Similarity metrics and adaptive techniques are discussed.

The purpose of this study is visualization of concealed light elements materials in metallic environments using a cooled CCD camera with an image intensifier. In this study, gamma-rays were used for visualization. As a gamma-ray source, radio-isotopes such as ⁵⁷Co, ¹³³Ba and ¹³⁷Cs were used. Gamma-rays penetrate through sample materials and are firstly enhanced by an image intensifier and then detected by a cooled CCD camera. Materials such as explosives or organic matters were observed through various metal plates.

From the court we were asked whether it is possible to determine if an image has been made with a specific digital camera. This question has to be answered in child pornography cases, where evidence is needed that a certain picture has been made with a specific camera. We have looked into different methods of examining the cameras to determine if a specific image has been made with a camera: defects in CCDs, file formats that are used, noise introduced by the pixel arrays and watermarking in images used by the camera manufacturer.

Advances in the processing power of computers and mathematical algorithms over the last decade have given forensic scientists greater image analysis capabilities. It is now possible to extract a wealth of information from a digitized image. It is therefore the forensic scientist's task to determine the relevance and accuracy of this information in the context of the customer's request and the circumstances of the case.

We present a new image processing strategy that enables an automated extraction of signatures from striation patterns. To this end, a signal model is proposed that allows a suitable description of the interesting features of forensically relevant striation marks. To provide for a high image quality, several images of the same surface area are recorded under systematically varying conditions. The images obtained are then combined to an improved result by means of appropriate sensor fusion techniques. Based upon the signal model, the signal of interest is concentrated, and a compact representation of the grooves is obtained. To enable an efficient description of the relevant features even in the cases of deformed surfaces or curved striation marks, a straightening of the grooves is performed before. In the following, a meaningful signature describing the information of interest is extracted using the whole length of the grooves. This signature can be used for an objective evaluation of similarity of striation patterns.

On the market several systems exist for collecting spent ammunition data for forensic investigations. These databases store images of cartridge cases and the marks on them. Image matching is used to create hit lists that show those cartridges in the database which have marks that are most similar to the marks of the cartridge case under investigation. The research in this paper focuses on the different methods of feature selection and pattern recognition that can be used for optimizing the results of image matching. A fast pre-selection method based on signatures is applied that is based on the Kanade Lucas Tomasi equation. The positions of the points compared with this method are compared. In this way 11 of the 49 images were in the top position in combination with the third scale of the a trous wavelet. Light conditions and the prominence of the marks determines to a large extent whether correct matches are found in the top ranked position. All images were retrieved in the top give percent of the complete database. This method takes only a few minutes, which can be structured for comparisons to be carried out in seconds.

In this research an evaluation has been made of the different ways of contents based image retrieval of logos of drug tablets. On a database of 432 illicitly produced tablets (mostly containing MDMA), we have compared different retrieval methods. Two of these methods were available from commercial packages, QBIC and Imatch, where the implementation of the contents based image retrieval methods are not exactly known. We compared the results for this database with the MPEG-7 shape comparison methods, which are the contour-shape, bounding box and region-based shape methods. In addition, we have tested the log polar method that is available from our own research.

Software-based grouping of multiplexed video based on video content, as opposed to the signal generated by multiplexers, is described. The method is based on energy minimization approach. The algorithm automatically determines the amount of multiplexed camera views, and then the frames are grouped with respect to camera views. The algorithm is free of any threshold differences between camera views, and does not depend on the presence of quiet zones. The method also compensates for interference noise, local and global motion, are contrast changes.