We discuss the optimization of components in a single-wavelength airborne laser bathymeter that is intended for a low-power unmanned aerial vehicle platform. The theoretical minimum energy requirement to detect the submerged sea floor in shallow (<; 5 m) water using a low signal-to-noise ratio (LSNR) detection methodology is calculated. Results are presented from tests of a prototype light detection and ranging (LiDAR) instrument that was developed by the University of Florida, Gainesville. A green wavelength (532 nm), 100-beamlet, low-energy (35-nJ/beamlet), short-pulse (480 ps) laser ranging system was operated from a low-altitude (500-m) aircraft, with a multichannel sensor that is capable of single photoelectron sensitivity and multiple stops. Data that were collected during tests display vertical structure in shallow-water areas based on fixed threshold crossings at a single-photon sensitivity level. A major concern for the binary detection strategy is the reliable identification and removal of noise events. Potential causes of ranging errors related to photomultiplier tube afterpulsing, impedance mismatching, and gain block overdrive are described. Data collection/processing solutions based on local density estimation are explored. Previous studies on LSNR performance metrics showed that short (15-cm) dead time could be expected in the case of multiple scattering objects, indicating the possibility of seamless topographic/bathymetric mapping with minimal discontinuity at the waterline. LiDAR depth estimates from airborne profiles are compared to on-site measurements, and near-shore submerged feature identification is presented.

Airborne light detection and ranging (LiDAR) surveys provide a rich data source of topographic information. However, beach monitoring with LiDAR data has been mostly limited to visualization and first-order measures derived from digital elevation models (DEMs). To exploit more information from multitemporal LiDAR data acquired over a beach, we extract surface features to detect morphologies that are indicative of shoreline change patterns. First, through cross-shore profile sampling of LiDAR-derived DEMs, the continuous 3-D beach surface is parameterized into several 1-D morphologic features progressing alongshore. Profiles are subsequently partitioned into binary erosion or accretion classes dependent on measured shoreline change between surveys. Then, a feature's class separability is quantified using information divergence measures nonparametrically constructed via Parzen windowing. The more interclass separation provided by a feature, the greater its discriminative potential, and the higher its ranking as a morphologic indicator. Rankings are computed across the survey epochs to evaluate performance stability of the features. Finally, the top-ranked features are implemented within a naive Bayes classifier to assess their ability to segment terrain more likely to erode. Results demonstrate the utility of the developed framework to systematically extract and incorporate useful morphologic information from LiDAR data for discerning patterns in beach change.

The Dynamic Tree (DT) Bayesian Network is a powerful analytical tool for image segmentation and object segmentation tasks. Its hierarchical nature makes it possible to analyze and incorporate information from different scales, which is desirable in many applications. Having a flexible structure enables model selection, concurrent with parameter inference. In this paper, we propose a novel framework, dynamic factor graphs (DFG), where data segmentation and fusion tasks are combined in the same framework. Factor graphs (FGs) enable us to have a broader range of modeling applications than Bayesian networks (BNs) since FGs include both directed acyclic and undirected graphs in the same setting. The example in this paper will focus on segmentation and fusion of 2D image features with a linear Gaussian model assumption.

Microwave emission and scattering models require the parametrization of surface roughness. Traditionally this has been achieved by sampling the surface in transects. In this work, roughness is characterized from 3D surface models derived from ground-based LiDAR. The dataset consist 18 surfaces with varying roughness characteristics. 2D profiles extracted from the surface model constitute the baseline to compare to traditional profiling methods. It was found that sampling using profiles produces an underestimation of the RMSh by 25–63% and an even more severe underestimation in the correlation length that can reach up to an order of magnitude difference. From the 17,178 2D extracted profiles it was determined a significant sensitivity of the roughness parameters to the detrending methods, as well as a poor fit between the experimental ACF and the exponential and Gaussian models. Finally, methodologies to detrend quasi-periodic surfaces and the decomposition of surface at different scales are proposed and illustrate the advantage of having a 3D representation.

Airborne laser swath mapping systems collect thousands of range measurements per second, enabling high-resolution surface mapping. Low-signal-to-noise-ratio (LSNR) detection techniques allow for implementation of light detection and range (LIDAR) instrumentation aboard platforms with stringent power, size, and weight restrictions. The University of Florida has developed the coastal area tactical-mapping system, which is a prototype LSNR LIDAR system that is capable of single-photon laser ranging. Laser pulses of 532 nm, 3.5 ¿J, and 480 ps are transmitted at 8 kHz, with the outgoing footprint split into a 10 × 10 array of elements. Backscattered photons are imaged back onto separate anodes of a multistop microchannel plate photomultiplier tube. A 2-GHz event timer processes returns over the array. Ranges for known targets are accumulated for repeated laser shots, and statistical analysis is applied to evaluate performance metrics that are related to minimum separation distance, bathymetric mapping depth, and atmospheric scattering. An LSNR detection paradigm based on Poisson statistics has enabled topographic target detection and shallow-water bathymetry. Quantitative results are used to draw conclusions about feature detection in topographic and bathymetric mapping scenarios.

Flow of water through stream networks directly impacts flooding and transport of sediments and pollutants in watershed systems. Hence, knowledge of streamflow is critical for water management and mitigation of flooding and drought events. Unfortunately, spatially dense networks of in situ streamflow measurements are generally unavailable and would be prohibitively expensive to deploy and maintain. Thus, a data fusion framework is needed that utilizes available data to predict streamflow. Observed data in spatial (e.g., topography and land cover), temporal (e.g., streamflow and groundwater levels), and spatiotemporal domains (e.g., rainfall) impact streamflow. Some of these quantities can be obtained from remote sensing imagery; however, combining such disparate data types using traditional data fusion methods is problematic. Physically based hydrologic models have been used to predict streamflow but often with significant uncertainty because numerous assumptions are made for many unmeasured input and parameter values. Traditional Bayesian inference approaches suffer from superlinear increases in computational complexity as the number of data sets to be fused grows. In this paper, a scalable spatiotemporal approach based on Bayesian networks (BNs) is presented for estimating streamflow. An information-theoretic methodology based on conditional entropy is employed to quantify the impact of adding nodes in the BN in terms of information gained. The framework offers the flexibility of embedding knowledge from hydrologic models calibrated for the study area by introducing them as additional nodes in the network, thereby improving prediction accuracy. Posterior probabilities of estimates and the associated entropy provide valuable information on the quality of predictions and also offer directions for future watershed instrumentation.

This paper presents an automated way of classifying mines in sidescan sonar imagery. A nonlinear extension to the matched filter is introduced using a new metric called correntropy. This method features high order moments in the decision statistic showing improvements in classification especially in the presence of noise. Templates have been designed using prior knowledge about the objects in the dataset. During classification, these templates are linearly transformed to accommodate for the shape variability in the observation. The template resulting in the largest correntropy cost function is chosen as the object category. The method is tested on real sonar images producing promising results considering the low number of images required to design the templates.

The dynamic tree (DT) graphical model is a popular analytical framework for image segmentation and object classification tasks. A DT is a useful model in this context because its hierarchical property encodes information in multiple scales and its flexible structure fits complex region boundaries better than rigid quadtree structures such as tree-structured Bayesian networks. This paper proposes a novel framework for data fusion by using a DT model to fuse measurements from multiple sensing platforms into a non-redundant representation. The structural flexibility of the DT will be used to combine common information across different sensor measurements of simulated objects of interest. The appropriate structure of the DT and its parameters for the data fusion application are presented and discussed along with fusion results from a simulated sonar survey mission.

Estimating and segmenting topography by fusing data acquired over multiple resolutions has been extensively studied over the years. The standard multiscale Kalman smoother estimator embedded with a single stochastic model parameterized using power-spectral matching methods has been found to give suboptimal performance in estimating nonstationary topographic variations. The modeling is based on the fact that topography and other geophysical phenomena exhibit a 1/fproperty. Although acceptable for data sets over large areas, this approximation is found to be poor for data sets over small areas. This letter employs multiple models regulated by a mixture-of-experts network to adaptively fuse the estimates. Alternate to power spectral methods, a fractal-based approach is used to segment data and parameterize the multiple models for better performance. Sensitivity and performance analyses are also performed on the parameters of the estimators, and ideal selection criteria are proposed.

Government and commercial airborne light detection and ranging (lidar) systems have enabled extensive measurements of the Earth's surface and land cover over the past decade. There is much interest, however, in employing smaller lidar systems that require less power to enable sensing from small unmanned aerial vehicles or satellites. Technological advances in the performance of small microlasers and photodetector sensitivity have recently enabled the development of experimental airborne lidar systems with low signal-to-noise ratios (LSNRs). Recent government and academic prototypes have indicated that LSNR airborne lidars could significantly increase the fidelity of terrain reconstruction over what is possible with existing conventional lidars. Thus, there is a need to build up a modeling capability for such systems in order to aid in future system and mission design. A numerical sensor simulator has been developed to model the expected returns from LSNR microlaser altimeter systems and predict their performance. Both optical and signal processing system components are considered, along with other factors, including atmospheric effects and surface conditions. Topographic (solid Earth) and bathymetric (littoral zone) measurement scenarios are considered. The analysis of topographic simulation data focuses on the effect of solar noise on SNR and elevation accuracy while bathymetric performance is evaluated with regard to water depth and scan angle for different water clarities. The mission conditions chiefly responsible for limiting the performance of LSNR lidar are discussed in detail, along with suggestions for further algorithm development and system performance evaluation.

This paper describes the design, development, and analysis of a scalable and portable architecture for multi-dimensional, non-parametric PDF estimation using Gaussian kernels on FPGAs.

Technological advances in the performance of small micro-lasers and photo-detector sensitivity have recently enabled the development of experimental airborne lidar systems with low signal-to-noise ratios (SNR). These systems show the potential to perform highly accurate topographic and bathymetric mapping. There is a need to build up a modeling capability in order to aid in future system and mission design. A numerical sensor model has been developed to predict measurement performance from low SNR lidar systems. Both optical and signal processing components are considered, along with other factors including atmospheric effects and surface conditions.

Complex hydrologic problems, such as estimating stream discharge (flow), require the utilization (fusion) of multiple sets of measurements (features). Due to the computational cost of incorporating all available features, it is desirable to use a reduced set of the most informative ones. A method is presented for determining the information gain of different feature subsets in a Bayesian network. The method is applied to the problem of estimating flow in a Spatio-Temporal Bayesian Network (STBN), under the constraint that the features retain their original physical meaning.

The L-band signals broadcast by GPS satellites are attenuated by vegetation, making it problematic, if not impossible, to predict the performance of the system in forested areas without some quantitative measure of the structure and density of the local forest canopy. Airborne laser swath mapping (ALSM) observations can be used to rapidly and remotely sample the structure and density of forested areas. We report here the results of a study performed to determine the attenuation of GPS signals in forests, by correlating changes in the signal-to-noise ratio (SNR) of the received GPS signals under different canopies, using three dimensional structure and density information about each canopy derived from ALSM observations. The results of this study verify that the loss of signal is strongly correlated with the local structure and density of the forest, and we demonstrate how the ALSM point cloud can be used to better predict the attenuation of the GPS signals. The results of this research also pertain to other modes of microwave transmission in forested areas, including satellite and cellular telephony, and the estimation of biomass from L-band radar.

Recent advances in technology have enabled the acquisition of high-resolution topographic data by means of airborne laser swath mapping (ALSM), which can yield digital elevation models (DEMs) with horizontal resolutions of 1m. A DEM is a grayscale image wherein pixel value corresponds to elevation. Using ALSM imaging systems over forested terrain, we filter out the laser returns from the occluding foliage and estimate bare-surface DEMs. Extracting stream networks from DEMs is important for modeling many hydrological processes. We apply a sequence of morphological operations to an ALSM DEM to detect stream channels in forested terrain. We verify the accuracy of the results using a set of error measures over simulated terrain and also using GPS ground truth over real terrain. For linking disconnected stream segments, a measure of pixel connectivity is used.

Modern airborne laser swath mapping (ALSM) systems measure both elevation and reflection intensity of the terrain. However, this intensity has been under utilized as a feature for image classification because it does not represent true terrain radiance. In areas with minimal topographic relief, such as beaches, we show that segmenting intensity images rather than elevation images has great potential for scene analysis. Several intensity-based features are extracted from ALSM data collected along a beach and partitioned into three classes to detect the water line. Class-conditional probability density functions are estimated for each feature to asses which are most informative. Results indicate significant class separation using centroidal features. Their classification performance is evaluated using a naive Bayes classifier and the area under receiver operating characteristic curves. The method presented provides a novel feature extraction and a systematic feature selection procedure for high-resolution ALSM intensity data.

Our paper proposes an approach for the extraction of stream channels from Airborne Laser Swath Mapping (ALSM) data. Recent advances in technology have led to high-resolution topographic data acquisition by means of airborne lidar (i.e. ALSM), which can yield Digital Elevation Model (DEM) datasets with horizontal resolutions of 1m and vertical rms errors in the range of 10 – 15 cm. The extraction of a stream network from a DEM plays a fundamental role in modeling spatially distributed hydrological processes and flow routing. We apply morphological filtering to an ALSM DEM to detect and characterize stream channels in forested terrain. Since the size and shape of morphological Structuring Elements (SEs) is known to strongly affect filtered results, we test for accuracy by developing a set of error measures over simulated terrain. We subsequently apply the filter to actual ALSM data. For linking disconnected stream segments, a measure of pixel connectedness known as the Connectivity Number is used. The method presented is shown to enable systematic characterization and comparisons of streams, even in heavily forested terrain.

Stream flow directly impacts flooding and transport of sediments and pollutants (e.g. NO3) in watershed systems, and hence knowledge of current and future flow is important. Unfortunately, spatially dense networks of in situ stream flow meters are not generally available and would be prohibitively expensive to deploy and maintain. It is known that surface morphology, land cover, and rainfall can impact stream flow through time, and these quantities can be measured using remote sensing techniques. Physically-based hydrologic models have been used to estimate flow using such data, but such approaches are limited because severe assumptions often have to be made for many input values. We present an alternative approach based upon a spatio-temporal Bayesian network (STBN) solution and compare its performance to methods based on physical models and more approximate probabilistic models. We show that the STBN method is able to capture the non-stationarity of hydrologic processes that contribute to streamflow in space and time while also limiting the computational requirements.

Airborne lidar data were acquired along St. Augustine Beach, Florida six times between August 2003 and June 2006. To identify sub-aerial morphologies indicative to beach erosion, the data sets were mined extensively by extracting several morphological features using cross-shore profile sampling. For each profile, the features were grouped into erosion or accretion classes and their class-conditional probability density functions (PDFs) estimated via Parzen windowing. PDF separability was ranked using symmetric and normalized measures of relative entropy (i.e. divergence). Results were compared to a simple median metric. The more interclass separation provided by a feature, the greater its potential as an indicator for erosion or accretion. Over short time periods (>1 month), beach slope and beach width ranked highest by providing the most separation and therefore high potential as indicators for erosion. Over longer time periods (>1 year), deviation-from-trend, which is the shoreline's deviation from the natural strike of the beach, ranked highest. This is significant in that the pier region's deviation from the natural trend is believed by coastal researchers to be a strong contributing factor to it being an erosion "hot spot". The method we have developed provides a systematic framework to mine high-resolution airborne lidar data over beaches, detect erosion-prone areas, and numerically rank a feature's potential as an indicator for erosion.

Recent advances in sensor technologies have resulted in tremendous increases in the amount of data collected for imaging applications such as airborne and space-based remote sensing of the Earth. Data acquisition and dissemination systems need to perform more processing than ever before to support real-time applications and reduce bandwidth demands on the downlink. FPGA-based reconfigurable computing systems are emerging as cost-effective solutions that offer enormous computation potential in the embedded systems arena. Research in this paper explores the potential capability offered by deploying reconfigurable computing systems in a remote sensing system by means of a commonly employed application. Multiple designs for a multiscale data-fusion algorithm were developed for an FPGA-based platform. These designs are used to demonstrate speedup over processor-based solutions and study the demands posed by such applications upon the system. Due to the vast number of sensors inputs, such applications pose high demands on the memory capacity and bandwidth which becomes a critical factor in determining the overall system performance. Results of our experiments depict that over an order of magnitude improvement can be obtained with efficient designs and appropriate hardware resources. Projections of enhanced performance with emerging system architectures are also presented.

Our paper proposes an approach for the extraction of stream channels from airborne laser swath mapping (ALSM) data. Recent advances in technology have led to high-resolution topographic data acquisition by means of ALSM yielding digital elevation model (DEM) datasets with horizontal resolutions of lm and a vertical accuracy of 0.15 m. We apply morphological operations on an ALSM DEM to detect stream channels. The results are compared with an existing terrain analysis tool known as TauDEM. Tools like TauDEM use morphology over large-scale areas but they often fail in detecting stream networks over small-scale areas. The proposed method uses small-scale morphology to provide complementary results for streamline locations over a small catchment area.

Accurate knowledge of soil characteristics is important for agricultural applications, as well as determining flood and contaminant infiltration rates and flow times. Current soil surveys often lack the necessary resolution to record these properties accurately. Remote sensing platforms, such as airborne based lidar, can easily measure topography at high resolution, yet it is difficult to measure soil properties remotely. It is known that topography gives information about certain soil characteristics, such as hydraulic conductivity. The relationship between topography and drainage is explored here and a probabilistic algorithm developed to predict soil drainage from surface morphology with measurable uncertainty.

Estimation of topography and fusion of data acquired over multiple resolutions have been extensively studied over the years. The standard MKS (multiscale Kalman smoother) algorithm embedded with a single stochastic model parameterized using power spectra matching methods have been found to give suboptimal performance in estimating nonstationary topographic variations. In this work, multiple models are regulated by a mixture-of-experts (MOE) network to adaptively fuse the estimates. A fractal based approach was employed to segment the data and parameterize the multiple models for better performance.

Sunlight flux in forest canopies is an important source of energy for the productivity of forests and is therefore an important input to biophysical process models. High-resolution three-dimensional estimates of sunlight flux are derived from directional foliage density measured from airborne laser swath mapping data. Forest edges are shown to exhibit measurable differences in flux. A method for determining appropriate in situ sampling intervals is also presented.

The multiscale Kalman smoother (MKS) is a globally optimal estimator for fusing remotely sensed data. The MKS algorithm can be readily parallelized because it operates on a Markov tree data structure. However, such an implementation requires a large amount of memory to store the parameters and estimates at each scale in the tree. This becomes particularly problematic in applications where the observations have very different resolutions and the finest scale data are sparse or aggregated. Such cases commonly arise when fusing data to capture both regional and local structure. In this work, we develop a reduced-complexity MKS algorithm and apply it to the fusion of topographic and bathymetric elevations on the Florida coast.

Several features extracted from airborne laser swath mapping (ALSM) data are examined to determine their effectiveness in separating buildings from trees across geographically and temporally diverse landscapes. These two classes are often spatially mixed in urban and suburban areas and can be quite difficult to separate based solely on geometric information due to the discrete sampling of ALSM. New median-based distance measures are used to quantify the separability of the classes using the different features. Information-based measures are also applied to the same data. For each of the test cases, it is possible to identify a common feature space in which the distance between the two classes is large. This distance information is an indication of the separability between classes and is therefore indicative of the potential success likely when trying to classify ALSM data. This analysis provides new insights into the richness of simple two-return ALSM data and to the spatial and temporal stability of ALSM features when discriminating between classes.

Multiscale Kalman smoothers (MKS) have been traditionally employed for data fusion applications and estimation of topography. The standard MKS algorithm embedded with a single stochastic model has been found to give suboptimal performance in estimating nonstationary topographic variations, particularly when there are sudden changes in the terrain. In this work, multiple models are regulated by a mixture-of-experts (MOE) network to adaptively fuse the estimates. Though MOE has been widely applied to one-dimensional data, its extension to multiscale estimation is new

A multiscale filter designed for segmenting ground from nonground points in ALSM data is extended to be adaptive by analytically determining admission thresholds based on prescribed error rates. The filter performs well on simulated and actual ALSM data

Several novel features are examined to determine their effectiveness in separating buildings from trees in airborne laser swath mapping (ALSM) data. New one- and two-dimensional distance measures are created to quantify the separability of the classes using the different features. Several features involving the intensity of the laser returns were found to be highly effective at separating the classes. The new distance measure provides insight into what makes a good/bad feature when discriminating between classes. It also lays the groundwork for future classification of ALSM data by providing a systematic method of ranking features to be used for classification.

INSAR data from the ERS-1/2 platforms are combined with multiple sets of data acquired by the NASA/JPL TOPSAR platform to obtain statistically optimal high-resolution estimates of topography over the Finke River Gorge in central Australia. The INSAR data are fused using a multiscale Kalman smoother. The estimated topography preserves the spatial resolution of the TOPSAR data while smoothing noise and providing estimates where there was no TOPSAR coverage. It is shown that the estimation error associated with the multiscale Kalman smoother is smaller than that obtained with a deterministic least squares approach.

Interferometric synthetic aperture radar (INSAR) and laser altimeter (LIDAR) systems are both widely used for mapping topography. INSAR can map extended areas but accuracies are limited over vegetated regions, primarily because the observations are not measurements of true surface topography. The measurements correspond to a height above the true surface that depends on both the sensor and the vegetation. Conversely, topography from LIDAR is very accurate, but coverage is limited to smaller regions. The authors demonstrate how these technologies can be used synergistically. First, the authors determine surface elevations and vegetation heights from dual-baseline INSAR data by inverting an INSAR scattering model. The authors then combine sparse LIDAR observations with the INSAR inversion results to improve the estimates of ground elevations and vegetation heights. This is accomplished via a multiresolution Kalman Filter that provides both the estimates and a measure of their uncertainty at each location. Combining data from the two sensors provides estimates that are more accurate than those obtained from INSAR alone yet have dense, extensive coverage, which is difficult to obtain with LIDAR. Contributions of this work include (1) combining physical modeling with multiscale estimation to accommodate nonlinear measurement-state relationships and (2) improving estimates of ground elevations and vegetation heights for remote sensing applications.