Growing commitments in international operations gives rise to increasing demands on reconnaissance and surveillance systems, precision targeting, E/O warfare and stealth technologies, sensor networks and information processing. Emerging enabling technologies and capabilities using near-invariance of spectral signatures combined with spatial and dynamic target properties is presented. Particular applications are real-time detection of low contrast targets employing camouflage, concealment and deception. Modelling, simulation and validation capabilities are also discussed.

The ultimate performance of QWIP implies hard requirements on the response-to-dark-current ratio for both high operating temperature and low background, e.g. space, applications. A way to improve this ratio by finding the optimal combination of band structure and material parameters is suggested. Experiments have been conducted on GaAs/AlGaAs structures optimised for 8.5 to 16 μm with similar types of band profile. The doping concentration in the quantum well (QW) is the principal parameter in such optimisation because it affects linearly the photocurrent and exponentially the dark current. As a result of the first experiment series we found an optimal QW doping concentration corresponding to the maximum response-to-dark-current-ratio, thus verifying the validity of the widely used hydrodynamic model. Experiments with a varying number of quantum wells for a constant total thickness were also carried out and analyzed. The resulting variation in barrier thickness changes the balance between the quantum efficiency and photoconductive gain. A critical thickness was found, where the temperature-independent component of the dark current increases drastically. For low background applications, especially in combination with long wavelength detection, it is not enough to only reduce the thermally-assisted and sequential tunnelling components of the dark current. Other sources of the dark current usually neglected at high temperature start to play a role. Interface shape and background doping in the barriers are examples of increasingly important factors. We discuss the contribution of these factors to the dark current.

The objective of this paper is to present the Swedish land mine and UXO detection project "Multi Optical Mine Detection System", MOMS. The goal for MOMS is to provide knowledge and competence for fast detection of mines, especially surface laid mines. The first phase, with duration 2005-2009, is essentially a feasibility study which focuses on the possibilities and limitations of a multi-sensor system with both active and passive EO-sensors. Sensor concepts used, in different combinations or single, includes 3-D imaging, retro reflection detection, multi-spectral imaging, thermal imaging, polarization and fluorescence. The aim of the MOMS project is presented and research and investigations carried out during the first years will be described.

In this work are shown the principle, first experimental results and a model design of a new type of multi channel Fourier transform (FT) spectrometer for visible (VIS) and infrared (IR) region operating in real time. The main principle of this spectrometer is that measured collected and collimated optical radiation passes through a linear array or matrix of optical Fabry-Perot interferometers. Each interferometer is placed in front of and close to each element of the array detector. By processing the signal the spectrum of the optical radiation can be extracted. This design does not require intermediate optics between interferometer and array detector and allows for a reliable and extremely compact construction. Production cost can be low when a simple wedge type interferometer is integrated with existing array or matrix detectors, e.g. CCD camera. One other benefit is that the shape of the interferometer determines whether the spectrometer is suitable for measuring wide spectra radiation or has the ability to discriminate optical coherent radiation. Experimental results achieved for VIS and NIR range of spectra are promising. The principals of this design can be used for a variety of applications besides as a spectrometer. For example warning systems for lasers and restricted coherency sources and also filtering of optical signals and for measuring the spectral content working in a wide spectral range.

Spectroscopic and polarimetric imaging have an increasing range of applications in remote sensing as well as inspection systems. It is shown how a limited polarimetric imaging capability can be added to a conventional hyperspectral camera based on a transmission grating imaging spectrometer. This is done by utilizing the undiffracted part of the light and separating its focus at the detector into two components using a simple walkoff plate composite. The resulting camera has full hyperspectral capability in the visible and near infrared spectral range, and in addition it forms broadband images for two orthogonal linear polarizations. Example imaging results are given and it is shown how polarimetric information can be used to detect manmade objects in a natural scene. A discussion of the limitations of the system is given.

In this work we evaluate the imaging performance of a range-gated underwater system in natural waters. Trials have been performed in both turbid and clear water. The field trials show that images can be acquired at significantly longer distances with the gated camera, compared to a conventional video camera. The distance where a target can be detected is increased by a factor of 2. For images suitable for object identification, the range improvement factor is typically 1.5. We also show examples of image processing of the range-gated images, which increases the image quality significantly.

We present results from imaging experiments performed in Norway during the 2005/06 winter season. Pairs of infrared sources with different temperatures are placed at different distances, ranging from 50 to 1200 m, from two focal plane array infrared cameras. One of the cameras is sensitive in the 3-5 μm wavelength range and the other in the 8-10 μm wavelength range. During the winter months digital sequences of the IR-sources were recorded, under different meteorological conditions. These conditions ranged from perfectly clear, cloudless weather to heavy snowfall. Analysis consists of comparing the perceived contrast, as measured with the cameras, with the "real" contrast as defined by the temperatures of the IR-sources. It is assumed that the transmission coefficient is the product of the atmospheric transmission (without snow) and a transmission factor associated with the falling snow. FASCODE simulations, using the pertinent temperature and humidity data that were measured during the recordings, are performed to characterize the atmospheric transmission coefficient (without snow). A comparison of the experimental results and the simulation results allows one then to estimate the effect of the falling snow on the extinction coefficient or visibility range. We observed a strong negative correlation between visibility and precipitation rate and better visibility in the MWIR range than in the LWIR range.

Computer programs for prediction of optical signatures of targets and backgrounds are valuable tools for signature assessment and signature management. Simulations make it possible to study optical signatures from targets and backgrounds under conditions where measured signatures are missing or incomplete. Several applications may be identified: Increase understanding, Design and assessment of low signature concepts, Assessment of tactics, Design and assessment of sensor systems, Duel simulations of EW, and Signature awareness. FOI (the Swedish Defence Research Agency) study several methods and modeling programs for detailed physically based prediction of the optical signature of targets in backgrounds. The most important commercial optical signature prediction programs available at FOI are CAMEO-SIM, RadThermIR, and McCavity. The main tasks of the work have been: Assembly of a database of input data, Gain experience of different computer programs, In-house development of complementary algorithms and programs, and Validation and assessment of the simulation results. This paper summarizes the activities and the results obtained. Some application examples will be given as well as results from validations. The test object chosen is the MTLB which is a tracked armored vehicle. It has been used previously at FOI for research purposes and therefore measurement data is available.

GSIM is a simulator of infrared missile seekers. It simulates a missile flight scenario in a 3D environment from launch to arrival at the target. The radiation from objects and background is calculated using weather parameters, e. g. solar radiation and air temperature. After adding atmospheric transmission and sensor properties, the resulting seeker image is created. The primary output from GSIM is a sequence of images, generated by the seeker during the flight. In April 2005 helicopter trials were performed, where missile-like image sequences were recorded. The helicopter followed a slant path towards a number of ground vehicles, standing in a field. The recorded data has been used for evaluation of GSIM. At the evaluation, modelled temperature on background and targets was compared with radiometrically measured temperature. Simulations in GSIM were made of a selection of the recorded sequences. The difference between modelled and measured temperature is about 1 - 2 K, for both targets and background. When comparing the simulated and the original sequence, the visual impression is that there is an apparent similarity between the images. Long distance images show the strongest resemblance to reality, due to the limited level of detail in the background model.

As a one-dimensional nanostructural material, carbon nanotube (CNT) has been used to build different nanoelectronic devices due to its unique electrical properties. In this paper, the infrared (IR) responses of individual single-wall carbon nanotube (SWNT) and SWNT film are studied. A single-wall carbon nanotube is assembled onto a pair of electrodes to form Schottky contacts. The photongenerated electron-hole pairs within the carbon nanotube are seperated by an external electric field between the two electrodes. The separated carriers contribute to the current flowing through the carbon nanotube and form photocurrent. By monitoring the photocurrent, the incident infrared can be detected and quantitated. The single-wall carbon nanotube based infrared sensor is designed and a series of efficient and reliable fabrication and assembly processes are developed for the sensor fabrication. With an atomic force microscope based nanomanipulation system as the assembly tool, a single carbon nanotube can be easily assembled onto the electrodes. Since the assembly process is controllable and reliable, it becomes possible to fabricate an individual carbon nanotubes based infrared sensor array, which was difficult to fabricate with other fabrication method. The photocurrent responses of individual SWNT IR sensor and SWNT film IR sensor are measured and analyzed. Experimental results show the good sensitivity of SWNTs to the infrared light. Our results shows a three orders higher photocurrent than the previous reported results. It has also been shown that an individual SWNT IR sensor is more sensitive than a SWNT film IR sensor.

Here we present an engineering study showing how altering various aspects of the growth parameters of an InAs dot within an InGaAs well (DWELL) QDIP affects its performance. Amongst our findings, we show capability to control the absorption wavelength both during and after growth by altering the size of the dots and via the quantum confined Stark effect respectively. The addition of AlGaAs current blocking layers is shown to reduce deleterious dark current by over two orders of magnitude.

This paper reviews characteristics and performance of the amorphous silicon microbolometers with a pixel-pitch of 25 μm. We first present the advantages of amorphous silicon uncooled microbolometer technology which enables the production of high volume and low cost uncooled IRFPA. The 25 μm pixel architecture profits from the low thermal time constant of 45 μm pixel detector, to design higher thermal insulated pixel. It enables the development of a 25 μm pixel-pitch detector which has high performance despite the pixel pitch reduction. Thanks to this new pixel design and by pushing the design rules even further, high fill factor has been kept, without the use of complex, as well as expensive, two-level structure. IRFPAs are then described in terms of ROIC architecture, packaging, operability and electro-optical performances. New readout integrated circuit structure has been specially developed for this pixel pitch. High level functions like gain, offset correction, image flip and windowing could be operated through a serial link to simplify the electrical interface. At a 60Hz frame rate, focal planes with less than 50mK (f/1) NETD are now achieved with high stability regarding environmental temperature.

Infrared (IR) applications are more and more demanding regarding reliability and cost. Moreover contracts are getting further than simple cost acquisition considerations, asking for life cycle cost requirements. Life cycle cost approach is the way to combine increase of reliability with reduction of costs. Starting from the design of a IR detector and taking in consideration some specific tactical system constraints, the whole cost and reliability analyses have to be made including maintenance approaches. The key parameters for the detectors are the thermal cycles behaviour, the cooler reliability and the vacuum behaviour. Based on Sofradir experience of life cycle cost contracts and on IR staring arrays results, this paper discuss the optimization of life cycle cost and reliability as well as future trends regarding IR detectors and reduction of life cycle cost.

HgCdTe (Mercury Cadmium Telluride / MCT) staring arrays for infrared detection do show constant improvements regarding their compactness and performances. New detectors are now proposed offering system solutions in the different IR wavebands and profiting of the latest technology improvements as well as MCT performance advantages and cost reduction. As a matter of fact, the size of MCT wafer has grown to 4", the pixel pitch was lowered to 15μm while maintaining outstanding results on the Focal Plane Arrays (FPA) uniformity. New functions as the Analog to Digital Conversion (ADC) are added to read-out circuits. Results are presented concerning Non Uniformity Corrections (NUC) stabilities for two Sofradir products. Then results from developments of integrated ADC are addressed and finally, the Jupiter 1280x1024 mid-wave (MWIR) MCT detector performance results are presented.

Full replacement of scanning type thermal imaging (TI) equipment by starring type TI is on agenda of XXI century. Realization requires development and launching production of affordable large format high performance photovoltaic (PV) infrared (IR) focal plane arrays (FPA) covering spectral range from 1.3 to 14 μm. Advanced Infrared Focal Plane Arrays include Short-Wave (SWIR) 1.3-2.5 μm operating at near room temperature T_op=300 K, Mid-Wave (MWIR) 3- 5.5 μm operating at higher temperatures (HOT) T_op=200-240 K, extended Long-Wave (LWIR) 8-14 μm operating at temperatures T_op=80-100 K and multi-color arrays. Challenge is serious and adequate research and development works have to be done. One of major approach is fabrication of megapixel IRFPA based on Hg_1-xCd_xTe epitaxial multi-layer structures grown by molecular beam epitaxy (MBE). To be successful with approach it is necessary to optimize design of pixel and develop reproducible MBE growth technology issuing epitaxial structures with perfect layers and interfaces. Perhaps novel FPA will be based on photodiodes (PD) with p-n junction opposite to usually used n⁺-p junction. PD with optimal p-n junction could have lower dark current value than same size n⁺-p junction. It is desirable for proper multiplexing of PD arrays to Silicon Read-out Integrated Circuits (ROICs). Objective of the present work was to examine Hg_1-xCd_xTe LWIR PV FPA (λp ranged from 10 to 11 μm at T_op=80-100 K) performance variation with doping level, absorber thickness, interface shunting and operating temperature.

The first generation of high performance thermal imaging sensors in the UK was based on two axis opto-mechanical scanning systems and small (4-16 element) arrays of the SPRITE detector, developed during the 1970s. Almost two decades later, a 2nd Generation system, STAIRS C was introduced, based on single axis scanning and a long linear array of approximately 3000 elements. This paper addresses the development of the UK's 3rd Generation High Performance Thermal Imaging sensor systems, under a programme known as "Albion". Three new high performance detectors, manufactured in cadmium mercury telluride, operating in both MWIR and LWIR, providing high resolution and sensitivities without need for opto-mechanical scanning systems will be described. The CMT material is grown by MOVPE on low cost substrates and bump bonded to the silicon read out circuit (ROIC). All three detectors are designed to fit with existing standard Integrated Detector Cooling Assemblies (IDCAs). The two largest detectors will be integrated with field demonstrator cameras providing MWIR and LWIR solutions that can rapidly be tailored to specific military requirements. The remaining detector will be a LWIR device with a smart ROIC, facilitating integration times much longer than can typically be achieved with focal plane arrays and consequently yield very high thermal sensitivity. This device will be demonstrated in a lab based camera system.

The characteristics of a multifunctional two-colour-avalanche gain Focal Plane Array, FPA, in which one of the bands can be used in avalanche mode to produce current gain with low excess noise and low dark current, are reported. The multifunctional FPA is based on a bi-colour pseudo-planar MCT detector structure, developed at the CEA-LETI, which superposes two planar type diodes with different composition of Cd. The electro-optic characteristics of the multifunctional LW-MW-avalanche gain detectors are reported for 256x256 30μm pitch arrays hybridised on a bi-colour read-out circuit, and for direct measurements on 30μm pitch test arrays. An avalanche gain of M=5300 at an inverse bias of V_pol=-12.5V and a noise factor close to F=1, is reported for MW wavelength diode, characterised by a cut-off wavelength of λ_C=5.03μm. A new measure of the sensitivity limit of the APD, the equivalent shot noise limited dark current, i_{eq_in,} was defined and estimated from dark current noise measurements i_{eq_in}=3.0 10^-10A at gain M<300 and i_{eq_in}=1.0 10^-10A at M=5300. The results will be discussed in view of the wide scope of applications which are enabled by the multifunctional FPAs.

Various experiments have been carried out recently in Middle East desert regions for prediction of aerosol particle concentration and size distribution. The differences between urban and non-urban desert coastal environments, and urban and non-urban desert continental environments, are of interest. During these experiments aerosol particle concentrations for similar weather but different environmental conditions were measured and analyzed. The well-known MODTRAN urban, desert, maritime, and continental aerosol models were tested for aerosol particle distribution prediction. Unfortunately, many of those models did not predict correctly the aerosol distributions in different types of arid environments. Comparison of results leads to the following interesting conclusions: the aerosol distribution for each of our environments, such as urban-coastal-desert, non-urban-coastal-desert, urban-continental-desert, and non-urban-continental-desert, can be predicted by multiplying the Mediterranean coastal aerosol model parameters with semi-empirical correction functions or factors. The functions are different for coastal and continental environments, and for urban and non-urban environments. The models proposed in this work better describe effects of different atmospheric conditions for Middle East desert aerosol modeling.

Various experiments have been carried out recently in the Middle East urban (Beer Sheva, Israel) environment for prediction of aerosol particle concentration and size distribution. During these experiments aerosol particle concentrations for different weather conditions were measured and analyzed. New aerosol size distributions are described in this work, based on an extensive series of measurements. This measurement data show the processes of absorption and scattering by aerosol particles in urban inhomogeneous areas for rainy atmospheric conditions. Several parts of the results are compared with those obtained through measurements in different geographic and climatic environments, as well as with different aerosol distribution models. Effects of different atmospheric conditions during rain on aerosol modeling are better described in this work.

Thermal infrared (IR) lenses require efficient anti-reflection coating. Moth-eye (or egg-box) 2D subwavelength gratings have demonstrated their ability to reach a very high transmission for a wide wavelength and angular range. The use in thermal IR is simplified by the lower resolution for lithographic technology, compared to visible waveband. However, deeper structures must be engraved and lithography must be adapted to IR materials. In order to be cost-effective, the patterning must be produced by replication techniques, such as embossing. Our laboratory is now experimenting hot embossing of moth-eye patterns in chalcogenide substrates. In this paper, theoretical analysis, micro-lithographic technology and manufacturing processes are detailed.

Pupil plane encoding has shown to be a useful technique to extend the depth of field of optical systems. Recently, further studies have demonstrated its potential in reducing the impact of other common focus-related aberrations (such as thermally induced defocus, field curvature, etc) which enables to employ simple and low-cost optical systems while maintaining good optical performance. In this paper, we present for the first time an experimental application where pupil plane encoding alleviates aberrations across the field of view of an uncooled LWIR optical system formed by F/1, 75mm focal length germanium singlet and a 320x240 detector array with 38-micron pixel. The singlet was corrected from coma and spherical aberration but exhibited large amounts of astigmatism and field curvature even for small fields of view. A manufactured asymmetrical germanium phase mask was placed at the front of the singlet, which in combination with digital image processing enabled to increase significantly the performance across the entire field of view. This improvement is subject to the exceptionally challenging manufacturing of the asymmetrical phase mask and noise amplification in the digitally restored image. Future research will consider manufacturing of the phase mask in the front surface of the singlet and a real-time implementation of the image processing algorithms.

Large-aperture, high-resolution medium-wave IR sensors are applied for oblique imaging from a variety of airborne platforms. Applicable optical configurations include both reflective (aplanatic Cassegrain) and refractive devices. In the former case the MWIR typically complements a primary EO band, with a common front objective and a dichroic beam splitter. ELOP's successful Condor2 LOROP camera is an example of such configuration. The camera provides consistent EO and IR performance at long stand-off ranges from a high-altitude fast jet platform. The long range oblique IR imaging represents a particularly complex case for system analysis and performance prediction. Atmospheric attenuation changes significantly as a function of altitude and line-of-sight depression angle. Temperature of the camera optics may vary in wide bounds and could differ substantially from the target temperature. Pixel pitch, optical transmittance and (in case of Cassegrain optics) a ratio of central obscuration all have a strong effect onto the resolution performance. The paper presents a normalized multi-parametric performance study where the resolved spatial frequency is expressed as a fraction of the optical cutoff frequency. The major dependencies of the resolution on the key parameters are illustrated. The analysis includes a relatively wide range of variation of the design parameters, covering most of the feasible reflective and refractive optical configurations.

We present a computationally inexpensive method for multi-modal image registration. Our approach employs a joint gradient similarity function that is applied only to a set high spatial gradient pixels. We obtain motion parameters by maximizing the similarity function by gradient ascent method, which secures a fast convergence. We apply our technique to the task of affine model based registration of 2D images which undergo large rigid motion, and show promising results.

We point out a problem that exists when designing new computer vision systems that have image acquisition components: the need to estimate the capabilities of the system in real-world scenarios, before the system is ready for image acquisition. Also in ready-made instruments, it is sometimes preferable to avoid operation unless a successful outcome is likely, when the price of failure is high, as in the use of space probes. Thus the end result should be established before actual acquisition by the system. Our proposed solution to the problem is based on real images of real scenes, acquired by a pre-made high quality imaging device, whose operation is established. These images then undergo processing, which follows a meticulous understanding of the inner workings of cameras. It results in images that would be as the output result of the system about to be used. This enables estimation of subsequent computer vision tasks.

The main image degradation occuring in long distance ground-to-ground infrared video acquisition is due to atmospheric turbulence. The turbulence strength essentially depends on climatic conditions and on the distance between the scene and the camera. Atmospheric turbulence can show dramatically different effects, but in the case of horizontal observations in the troposphere, at a distance over a couple of kilometers, it can be efficiently simulated by local blurring and warping. Some additive noise may be detected depending on atmospheric conditions and on the acquisition system. In the acquisition conditions, the degraded images can be split into areas degraded by the same perturbation, which is called local isoplanatism.The goal of this paper is to test locally the most classical restoration methods on real images, in order to deduce some criterion allowing selection of the most suited method. The first part of the paper is devoted to the physical explanation of local isoplanatism (LI). In the second part, once the case study has been shown to fit with LI assumptions, we show that global restoration techniques do not work properly compared to local ones. Then local restoration results are analysed within uniform areas and transition areas so as to find the best restoration technique. Several examples are shown.

The SeaWolf Mid-Life Update (SWMLU) programme is a major upgrade to the UK Royal Navy's principal point defence weapon system. The update includes the addition of an Electro-Optic (EO) sensor to upgraded 'I' and 'K' band radars. The update presents a significant engineering challenge both in terms of hardware integration and software processing. The processing of sensor data into a coherent fused picture is a key element of the overall system design, and is critical to achieving the required system performance. Further to the fusion of object locations, derived object properties from both the spatial and temporal domains are also incorporated to create a highly detailed picture. Core functionality of the data fusion process is the association of objects between sensors and the labelling of objects into targets and own missiles. The data association results have a direct influence on overall system performance and labelling accuracy of objects is crucial to satisfy the system performance requirements. This paper discusses the data association and object labelling process followed in the SWMLU system and highlights sources of error and confusion for the EO sensor case. The effects of incorrect data associations are presented at the system-level. A number of software test environments for the EO sensor subsystem are introduced and analysed with a focus on data association.

The new emphasis on Anti-Terrorism and Force Protection (AT/FP), for both shore and sea platform protection, has resulted in a need for infrared imager design and evaluation tools which demonstrate field performance against U.S. Navy AT/FP requirements. In the design of infrared imaging systems for target acquisition, a discrimination criterion is required for successful sensor realization. It characterizes the difficulty of the task being performed by the observer and varies for different target sets. This criterion is used in both assessment of existing infrared sensor and in the design of new conceptual sensors. In this experiment, we collected 12 small craft signatures (military and civilian) in the visible band during the day and the LWIR and MWIR spectra in both the day and the night environments. These signatures were processed to determine the targets' characteristic dimension and contrast. They were also processed to bandlimit the signature's spatial information content (simulating longer range) and a perception experiment was performed to determine the task difficulty (N₅₀ and V₅₀). The results are presented in this paper and can be used for Navy and Coast Guard imaging infrared sensor design and evaluation.

In this paper, a novel tracking system based on the Track Before Detect Approach (TBD) is introduced. Each IR sequence is preprocessed first by using a whitening algorithm. This stage is used to reject clutter and emphasize targets. Afterwards, a Dynamic Programming Algorithm (DPA) is used for numerous frames in the sequence. The algorithm, a derivative of the well known Viterbi Algorithm, gives each pixel in the image a score based on the current frame and the previous one. By doing so, the temporal behavior difference between targets, clutter and noise is utilized to distinguish between them; we give scores accordingly. At the end of this stage, after the last frame of the IR sequence has been processed, the pixel with the highest accumulated score is chosen as the Target, and its path is found. The paper deals with the different issues characterizing the system, enabling it to have versatility over a wide range of scenes. Future work will involve the use of the system for tracking of targets in hyperspectral cubes.

Rotorcraft flying in low-altitude is endangered by power lines or telephone wires. The development of automated tools that can detect obstacles in the flight path and warn the crew would significantly reduce the workload of pilot and increase the safety. Detection and warning are rudimental demand and desire for Helicopter Avoidance Obstacle System (HAOS). And that, An advanced HAOS may be capable of classifying thin obstacles and enhanced vision with distances of obstacles. A laser 3D imaging system for helicopter avoidance obstacle (HAO) had been developed successfully. The laser 3D imaging helicopter avoidance obstacle system can not only detect thin obstacles but also catch more information of all objects of the area in front of the helicopter as possible. Then the information is transformed into intuitionist 3D image modality. In this paper, special features and characteristic of the laser imaging system for HAO are analyzed and discussed. Several design gist for this system are proposed. Especially, the developed zero backlash imaging technology and real-time dynamic imaging synchronizing with radar space scanning are described. The technique implementation problem and the system structure are given as well. Finally, the results of system ground test are presented. The ground test of the developed laser imaging system has demonstrated that the developed imaging system performance can achieve and satisfy commendably the requirements of the mission to prevent "wire strike".

Using a set of radiometric thermal hyperspectral data cubes, we developed an algorithm which detects the formation of an anomalous gas cloud. Once we've established the presence of the cloud in the latter images, we determine the origin of the cloud in the earlier ones and track its propagation. Gas usually expands from point sources and it is difficult to know whether it is significant when it occupies merely a few pixels in the image. After the gas size expands, it is easier to analyze as an interesting anomalous feature. Our algorithm includes techniques such as the improved K-Means classification, Spectral Angle Mapper (SAM), match filter and tracking; in the paper we will show results based on real data taken by the "FIRST" camera (Field-portable Imaging Radiometric Spectrometer Technology).

In the first Gulf War, unmoored floating mines proved to be a real hazard for shipping traffic. An automated system capable of detecting these and other free-floating small objects, using readily available sensors such as infra-red cameras, would prove to be a valuable mine-warfare asset, and could double as a collision avoidance mechanism, and a search-and-rescue aid. The noisy background provided by the sea surface, and occlusion by waves make it difficult to detect small floating objects using only algorithms based upon the intensity, size or shape of the target. This leads us to look at the sequence of images for temporal detection characteristics. The target's apparent motion is such a determinant, given the contrast between the bobbing motion of the floating object and the strong horizontal component present in the propagation of the wavefronts. We have applied the Proesmans optical flow algorithm to IR video footage of practice mines, in order to extract the motion characteristic and a threshold on the vertical motion characteristic is then imposed to detect the floating targets.

This paper describes 3D measuring methods of head and eye positions for development of a tracking system. The authors prove that the 3D information about the eye position can be estimated from the camera image of a single camera. A 3D position is generally detected and measured using stereoscopic images of multi-lens camera system. This pair of stereo images allows us to obtain the 3D information about the object. However, it is possible to estimate the position of eyes using a single image without stereo images. We found the simple method for getting the 3D position of eyes using the interest points of both ends of eyes. In this method, the interest points on the strait line give us the estimates of the 3D position of both eyes. Moreover, we propose the measuring method using the result of a geometric analysis.

The authors have developed a 3D workspace system using collaborative imaging devices. A stereoscopic display enables this system to project 3D information. In this paper, we describe the position detecting system for a see-through 3D viewer. A 3D display system is useful technology for virtual reality, mixed reality and augmented reality. We have researched spatial imaging and interaction system. We have ever proposed 3D displays using the slit as a parallax barrier, the lenticular screen and the holographic optical elements(HOEs) for displaying active image^1)2)3)4). The purpose of this paper is to propose the interactive system using these 3D imaging technologies. The observer can view virtual images in the real world when the user watches the screen of a see-through 3D viewer. The goal of our research is to build the display system as follows; when users see the real world through the mobile viewer, the display system gives users virtual 3D images, which is floating in the air, and the observers can touch these floating images and interact them such that kids can make play clay. The key technologies of this system are the position recognition system and the spatial imaging display. The 3D images are presented by the improved parallax barrier 3D display. Here the authors discuss the measuring method of the mobile viewer using infrared LED point markers and a camera in the 3D workspace (augmented reality world). The authors show the geometric analysis of the proposed measuring method, which is the simplest method using a single camera not the stereo camera, and the results of our viewer system.

A 3D mouse system is described that can specify the position and movement of the fingertip in a 3D space. When a mouse-moved event occurs to control the 2D object on the screen and the 3D object within the display, a touch panel on the screen can detect the movement of the fingertip. On the other hand, it is necessary to track the fingertip in a space for interaction to the 3D floating object as the touching image appears out in front of the display, serving as a virtual touch panel in the air. A 3D position is generally detected and measured using a stereo viewing. A series of two or more stereo images allows us to obtain the 3D information about the object. However, it is possible to estimate the position of a fingertip using a single image without stereo images. This paper proposes the measuring method of the position of a fingertip using a single camera image. This measuring system can provide 3D information to control 3D virtual objects. The system finds out the tip point and the width of a finger on the image, and approximately calculates the 3D position of a fingertip.

Image preliminary processing (pre-processing) of remote sensing satellites is considered as a vital process for the received data. The received video data is converted into an easy understandable raster images and any distortions that results during the satellite manufacturing process or through data acquisition and transmission are eliminated. This paper proposes the pre-processing algorithms that should be performed to the received data from remotely sensing satellites with Visible and Infra-Red (VIR) sensors beginning from reception of data by antenna at the Ground Data Reception Station (GDRS) until delivery of level 1A product (radiometrically corrected image). Briefly, the received data should be processed in two levels; level 0 to produce the raw image, and level 1 to produce radiometrically and geometrically corrected image.