James R. Buss

Early tumor detection by multiple infrared unsupervised neural nets fusion

The unsupervised classification algorithm called Lagrange constraint neural network (LCNN) has been successfully applied to the sub-pixel multispectral remote sensing, [25]. Here, we apply the LCNN to the early breast cancer detection using two-color mid and long infrared images of the breast. This could be a new paradigm shift that enabled smart neural network algorithm to sort out the underlying malignant heat sources for physician diagnoses. The nonintrusive 2-color passive infrared imaging that could be repeated for record track with no radiation hazard seems to be alternative paradigm shift for the first-line screening against breast cancer. The sub-pixel super-resolution capability of the remote sensing is equivalent to the sub-milimeter scaling of the close-up breast imaging for the vascular and the angiogenesis effects. We demonstrate the potential benefit of the multicolor mid & long infrared imaging capable for detecting the abnormal under-skin thermal textures as well as stage-zero detection of the ductal carcinoma in situ.

Carbon nanotube based spectrum infrared detectors

Carbon nanotubes (CNT) have a potential to be efficient infrared (IR) detection materials due to their unique electronic properties. The ballistic electronic transport property makes the noise equivalent temperature difference smaller compared to other semi-conducting materials. By overlaying CNT-based mid-IR (3-5μ) detectors on a long-wave IR (8-15μ) focal plane array, the mid-IR detector causes no filters loss. In order to verify this approach, a single pixel CNT- based infrared photodetector is fabricated by depositing the CNTs on the substrate surface and then aligning them using the atomic force microscopy (AFM)-based nanomanipulation system. Functionality of the single pixel CNT infrared detector is then verified and dark current is analyzed experimentally.

Subpixel jitter video restoration on board of micro-UAV

We review various image processing algorithms for micro-UAV EO/IR sub-pixel jitter restoration. Since the micro-UAV, Silver Fox, cannot afford isolation coupling mounting from the turbulent aerodynamics of the airframe, we explore smart real-time software to mitigate the sub-pixel jitter effect. We define jitter to be sub-pixel or small-amplitude vibrations up to one pixel, as opposed to motion blur over several pixels for which there already exists real time correction algorithms used on other platforms. We divide the set of jitter correction algorithms into several categories: They are real time, pseudo-real time, or non-real-time, but they are all standalone, i.e. without relying on a library storage or flight data basis on-board the UAV. The top of the list is demonstrated and reported here using real-world data and a truly unsupervised, real-time algorithm.

TG16 point target detection experiment POLLEX, Livorno 2001

NATO Task group TG16 is cooperating on topics related to ship self-defence. One of these topics is related to IR Search and Track sensors, which are in development for detection of low altitude air targets. In particular the group is working on models to predict the range performance of these sensors. Newly developed models include marine boundary layer effects such as refraction due to temperature gradients, scintillation due to turbulence and particle size distributions. TG16 organized in May 2001a trial in the Mediterranean Sea near Livorno, Italy, called POLLEX to further validate these models. Seven nations particulated with complementary instruments for measurements of the target signatures and environmental characteristics. Three targets were provided, a series of small visual/IR sources at a fixed distance, visual/IR soruces on a ship moving in and out up-to and beyond the horizon and a helicopter. The weather conditions during the measurement period showed interesting variations in Air to Sea Temperature DIfference and atmospheric turbulence. Data have been analzyed and samples of the results, as collected and/or analyzed by the participants, are discussed in this paper.

Independent component analysis for remotely sensed image classification with limited data dimensionality

The application of independent component analysis (ICA) to remotely sensed image classification has been studied recently. It is particularly useful for classifying objects with unknown spectral signatures in an unknown image scene, i.e., unsupervised classification. Since the weight matrix in ICA is a square matrix for the purpose of mathematical tractability, the number of objects that can be classified is equal to the data dimensionality, i.e., the number of spectral bands. When the number of spectral bands is very small (e.g., 3-band CIR photograph and 6-band Landsat image), it is impossible to classify all the different objects present in an image scene with the original data. In order to solve this problem, we present a data dimensionality expansion technique to generate artificial bands. Its basic idea is to use nonlinear functions to capture the second and high order correlations between original bands, which can provide additional information for detecting and classifying more objects. The results from such nonlinear band generation approach are compared with a linear band generation method using cubic spline interpolation of pixel spectral signatures. The experiments demonstrate that nonlinear band generation approach can significantly improve unsupervised classification accuracy, while linear band generation method cannot since no new information can be provided.

Lagrange constraint neural network for fully constrained subpixel classification in hyperspectral imagery

Linear unmixing approaches are used to estimate the abundance fractions of the endmembers resident in each pixel. Generally, two constraints will be applied. First, the abundance fractions of each endmembers should be nonnegative, which is called nonnegativity constraint. The second constraint, called sum-to-one constraint, says the sum of all abundance fractions should be one. One great challenge is to include the nonnegativity constraint while solving linear mixture model. In this paper, we propose a Lagrange constraint neural network (LCNN) approach to linearly unmix the spectrum with both sum-to-one and nonnegativity constraints.

Fully digital foliage penetrating synthetic aperture radar processor

A high performance, fully digital Foliage Penetrating Synthetic Aperture Radar (FOPEN SAR) system is described. The FOPEN SAR algorithm is illustrated using Matlab. Digital implementation is derived and simulated using VHDL. The complex mathematical functions required by the algorithm have been demonstrated. Simulations have achieved an SNR equals 290 dB when compared to the baseline results from Matlab. The accuracy of the simulation was limited by the resolution of certain trigonometric and exponential functions implemented using VHDL, and thus can be improved upon. This would allow greater flexibility between speed/area considerations without degradation of the target resolution (100dB-signal accuracy).

Infrared sensor system (IRSS) laboratory and field test results

The U.S. Navy Office of Naval Research (ONR) has developed an infrared search and track (IRST) demonstrator system named the infrared sensor system (IRSS). This technology-base sensor was successfully developed and tested both in the laboratory and at-sea. IRSS now is being transitioned to the Naval Sea Systems Command (NAUSEA) IRST Engineering and Manufacturing Development (E&MD) Program, where it will serve, with appropriate modifications, as the engineering development model (EDM) and will be fielded aboard a U.S. Navy ship. This paper summarizes the process of developing and fielding IRSS, describes test results accomplished at sea during 1996, and discusses the technical and engineering lessons associated with design, development and testing of IRSS. Results are presented covering the areas of sensor component and overall system radiometrics (e.g., sensitivity and dynamic range), channel uniformity, stabilization, and optical, electrical and information (i.e., signal processing/track) resolution.

Navy EO sensor test-bed development for infrared search and track

The U.S. Navy Office of Naval Research (ONR) is developing electro-optical (EO) sensor testbeds based on second-generation infrared (IR) focal plane array (FPA) technology, to be used for research and system demonstration in both shipboard and airborne applications. These technology-base sensors will demonstrate the functionality as well as many of the operational characteristics necessary ultimately in fielded infrared search and track (IRST) systems. The shipboard application involves the augmentation of existing and developmental radar assets in countering the advanced, sea-skimming, anti-ship missile (ASM) threat for mast-mounted use aboard high-value surface combatants. The ONR shipboard IRST testbed, the focus of this paper, currently is completing development, integration, and testing activities; BDM Federal (Arlington, Virginia) has supported these activities under contract to ONR over the past three years. The shipboard IRST testbed employs two separate, scanning FPAs covering the long wavelength IR (LWIR) and medium wavelength IR (MWIR). Preliminary test results show both a good correlation to predicted performance and a high level of absolute performance in terms of sensitivity and stabilization.