Scene Reflectance Research Articles

Nonstationary natural media analysis from polarimetric SAR data using a two-dimensional time-frequency decomposition approach

In synthetic aperture radar (SAR) polarimetry, it is generally assumed that the sensor has a fixed orientation with respect to objects and illuminates a scene with monochromatic radiations. Modern high-resolution SAR sensors have a wide azimuth beam width, however, and a large bandwidth in range. During SAR image formation, multiple squint angles and radar wavelengths are integrated to synthesize the full-resolution SAR image, and variations in the polarimetric signatures due to changes in the azimuthal look angle and in the wavelength commonly remain unconsidered. In this paper, a fully polarimetric two-dimensional (2D) time–frequency analysis method is introduced to decompose processed polarimetric SAR images into range–frequency and azimuth–frequency domains. This 2D representation permits characterization of the frequency response of the scene reflectivity, observed under different azimuth look angles and wavelengths. For the case of Bragg resonance in agricultural areas, the influence of anisotropic scattering and frequency selectivity on the polarimetric descriptors is pointed out in detail and compared with theoretical predictions from a quasi-periodic surface model. Lastly, a statistical analysis of polarimetric parameters is presented, which permits clear discrimination of media showing nonstationary behavior during the SAR integration.

The Effect of Scene Elevation on the Coherence of Wide-Angle Crossing-Node SAR Pairs

The interference between two radar images, acquired from approximately the same location, has a number of uses. In particular, the literature abounds with examples of coherent change detection (CCD) and interferometric synthetic aperture radar (InSAR) uses that, respectively, enable the detection of small changes and the extraction of terrain elevation. In almost all of these examples, collections have been deliberately constrained so that they have almost identical collection geometries. Details of the variation in geometry have been analyzed extensively but generally only with respect to variation in the mean grazing angle. Much of this variation is due to the angular variation in the scene reflectivity; however, when collection geometries differ in other respects, the reduction in performance cannot be entirely explained by angular variations alone. This paper identifies a source of performance degradation when the collection geometries are different for scenes containing local relief. The error is due to approximations during image formation that both the collection geometry and the terrain are flat. The effect of these errors is derived as a blurring kernel to the true scene. The impact of the blurring is then assessed in the interferometric case and broken into common and differential blurring terms to facilitate the impact of some special cases and to explain why the parallel-pass geometry works as well as it does even though the true three-dimensional nature of the collection and the ground are ignored.

A survey on image-based rendering—representation, sampling and compression

Image-based rendering (IBR) has attracted a lot of research interest recently. In this paper, we survey the various techniques developed for IBR, including representation, sampling and compression. The goal is to provide an overview of research for IBR in a complete and systematic manner. We observe that essentially all the IBR representations are derived from the plenoptic function, which is seven dimensional and difficult to handle. We classify various IBR representations into two categories based on how the plenoptic function is simplified, namely restraining the viewing space and introducing source descriptions. In the former category, we summarize six common assumptions that were often made in various approaches and discuss how the dimension of the plenoptic function can be reduced based on these assumptions. In the latter category, we further categorize the methods based on what kind of source description was introduced, such as scene geometry, texture map or reflection model. Sampling and compression are also discussed respectively for both categories.

Spectral Detection of Emergence in Corn and Cotton

Multispectral reflectance of emerging cotton (Gossypium hirsutum) and corn (Zeamays) seedlings was measured during the 2000 and 2001 growing seasons. Reflectance in blue, green, red, and near infrared (NIR) wave lengths was used to detect seedling emergence, to monitor leaf area growth, and to measure the effect of bare soil reflectance on scene (bare soil and seedlings) reflectance. Cotton and corn seedlings were detected 1 day after initial emergence (1 DAE) in 2000 by the red band. The red band detected seedlings in 2001 at 9 and 8 DAE in early and late planted corn, respectively, and on 0 DAE for cotton. The normalized difference vegetative index (NDVI) detected seedlings at 1 DAE or 2 DAE in both years. Seedling ground cover in 2000 on the initial detection date in the target areas averaged 1.3% and 0.9%, respectively, for cotton and corn; comparable values in 2001 for cotton, early planted corn, and late planted corn, were 1.4%, 0.4%, and 0.8%, respectively. The red wave band was the most sensitive single band for detecting the presence of seedlings, but NDVI was the most sensitive spectral indicator, which was apparently due to the red band since the NIR band did not always detect seedlings. Seedling leaf area was linearly correlated with NDVI values beginning at 1 or 2 DAE. Bare soil was the major component of the scene during stand establishment and dominated single band reflectance and NDVI values. A dry soil surface that was smoothed and sealed by rain usually caused single band reflectance to increase. The high variability in spectral characteristics of bare soil restricted the interpretation of the spectral data to concluding whether or not seedlings were emerging, but without estimating numbers and seedling size.

Remote estimation of vine canopy density in vertically shoot-positioned vineyards: determining optimal vegetation indices

Three spectral vegetation indices were examined for their correlation with vertically shoot-positioned canopy density (leaf area per metre of canopy) using both field spectroscopy data at the vine scale, and aerial image analysis at the vineyard scale. Vine canopy density accounted for a significant amount of the variability in all three vegetation indices (r2 = 0.65 to 0.92). The perpendicular vegetation index (PVI) was found to have the greatest information cost as well as the poorest coefficient of determination value, and thus was the least desirable of the indices examined. The ratio-based indices (RVI and NDVI) were shown to have similar information contents; however, the RVI was found to be more linearly related to canopy density over a broad range of values, and thus more desirable for vineyard remote sensing applications. Results from this analysis corroborate with findings from investigators in woodland and forest environments, and provide evidence of the complex nature of vineyard scene reflectance properties.

Data fusion techniques for super-resolution imaging

In this paper, we present two new techniques of using data fusion, based on the modality of the data generation process, to generate a super-resolved image from a sequence of low-resolution image intensity data. First, we develop a generalized interpolation scheme wherein an image is decomposed into appropriate subspaces, interpolation is carried out in individual subspaces and subsequently the interpolated values are transformed back to the image domain. Various structural properties of the image, such as 3D shape of an object, regional homogeneity, local variations in scene reflectivity, etc., can be better preserved during the interpolation process. The issue of preserving the structure in an image has received very little attention in the information fusion literature and we concentrate on this issue. In the second method, the data to be fused consist of a sequence of decimated, blurred and noisy versions of the high-resolution image. The high-resolution image is modeled as a Markov random field (MRF) and a maximum a posteriori (MAP) estimation technique is used to fuse the data to obtain a super-resolved image. The proposed technique does not require sub-pixel registration of given observations as most other techniques in the literature do. A simple gradient descent method is used to optimize the functional. The discontinuities in the intensity process can be preserved by introducing suitable line processes for improved result, albeit at the cost of computation.

Generalized interpolation and its application in super-resolution imaging

In this paper, we present a generalized interpolation scheme for image expansion and generation of super-resolution images. This is done by decomposing the image into appropriate subspaces, carrying out interpolation in individual subspaces and subsequently transforming the interpolated values back to the image domain. Various optical and structural properties of the image, such as 3-D shape of an object, regional homogeneity, local variations in scene reflectivity, etc., can be better preserved during the interpolation process. The motivation for doing so has also been explained theoretically. The generalized interpolation scheme is also shown to be useful in perceptually based high-resolution representation of images where interpolation is done on individual groups as per the perceptual necessity. Further, this scheme is also applied to generation of high-resolution transparencies from low resolution transparencies.

Monitoring regional vegetation change using reflectance measurements from multiple solar zenith angles

Many traditional models of vegetation canopy reflectance have commonly used one of two approaches. Either the canopy is assumed to consist of discrete objects of known reflectance and geometric-optics are then used to calculate shading effects, or, as in the turbid medium approach, the canopy is treated as a horizontally homogeneous layer of small elements of known optical properties and radiative transfer theory is used to calculate canopy reflectance. This paper examines the effect of solar zenith angle on the reflectance of red and near-infrared radiation from forests using a combination of these modelling approaches. Forests are first modelled as randomly spaced eucalypt crowns over a homogeneous understorey and the fractional coverage of four components: shaded and sunlit canopy and shaded and sunlit understorey are calculated. Reflectance from each fraction is then modelled for a range of solar zenith angles using the Verhoef SAIL model. The overall scene reflection as seen by a nadir viewing satellite sensor is compared for three forest types representing a gradient of crown density from open dry grassy woodlands to dense wetter closed forest with an understorey of mesophytic plants. Modelled trends in scene reflectance change are consistent with aircraft measurements carried out at three different solar zenith angles. Results indicate that an increase in both tree density and solar zenith angle leads to an increase in the dominance of shaded components. In the visible band, both the sparsely treed woodland and the medium density dry forest show similar trends to that predicted by a turbid medium model, however, the wet forest shows a less rapid decrease in reflectance with solar zenith angle. In the near-infrared band, as tree density increases from woodland to wet forest, overall scene reflectance shows increased departure from that modelled using the traditional assumption of smooth homogeneous canopies, changing from an increase with solar zenith angle for the woodland to a decrease with solar zenith angle for the forest types.

O3 profiles retrieved from limb scatter measurements: Theory

An algorithm is presented for retrieving vertical profiles of O3 concentration using measurements of UV and visible light scattered from the limb of the atmosphere. The UV measurements provide information about the O3 profile in the upper and middle stratosphere, while only visible wavelengths are capable of probing the lower stratospheric O3 profile. Sensitivity to the underlying scene reflectance is greatly reduced by normalizing measurements at a tangent height high in the atmosphere (∼55 km), and relating measurements taken at lower altitudes to this normalization point. To decrease the effect of scattering by optically thin aerosols/clouds that may be present in the field of view, these normalized measurements are then combined by pairing wavelengths with strong and weak O3 absorption. Excluding pointing errors, we conclude that limb scatter can be used to measure O3 between 15 km and 50 km with 2–3 km vertical resolution and better than 10% accuracy.

3-D Scene Modeling of Semidesert Vegetation Cover and its Radiation Regime

To explore the potential of multiangle remote sensing for estimating biophysical or ecological parameters over a variety of landscapes, a modeling tool that is capable of handling three-dimensional (3-D) heterogeneous structures, deriving ecological parameters from the vegetation structure, and effectively working on different scene scales is very desirable. A 3-D scene modeling approach for these purposes is presented in this paper. This 3-D model fulfills its goal by taking advantage of radiosity theory and computer graphics techniques. It consists of two major modules: a modified extended L-systems (MELS) method to generate a 3-D realistic scene and a radiosity-graphics combined method (RGM) to calculate the radiation regime based on the 3-D structures rendered with MELS. The 3-D simulation tool is then evaluated using field measurements of both plant structure and spectra collected during the NASA Earth Observing Satellite Prototype Validation Exercise Jornada field campaign near Las Cruces, NM. The modeled scene reflectance is compared with measurements from three platforms (ground, tower, and satellite) at various scales (from the size of individual shrub component to satellite pixels of kilometers). The agreement with measured reflectances is excellent at all sampling scales tested. As an example of the model's application, we use the model output to examine the validity of a linear mixture scheme over the Jornada semidesert scene. The result shows that the larger the sampling size (at least larger than the size of the shrub component), the better the hypothesis is satisfied because of the unique structure of the Jornada scene: dense plant clumps (shrub component) sparsely scattered on a predominantly bare soil background. A range of possible applications of this 3-D scene model is highlighted, and further work needed for 3-D modeling is also discussed.

Development of a New Underwater Bathymetric Laser Imaging System: L-Bath

The design, construction, and performance of a new high-resolution underwater bathymetric prototype system (L-Bath) with extended imaging capability is presented. The design offers simultaneous reflectance and depth information on a pixel-by-pixel basis so that high-resolution reflectance and bathymetric maps of underwater targets can be provided with exact registration. The design supports operation in shallow coastal waters under daylight conditions where high turbidity and the influence of ambient backscatter are particularly limiting for underwater imaging systems. Its configuration is similar to existing laser line scanning systems but uses a pulsed laser for the source and a fixed field-of-view high-resolution linear charge-coupled device (CCD) as receiver. The pulsed laser allows short camera integration times, thereby reducing the influence of the ambient daylight signal, and the fixed field of view of the detector provides a precision nonmoving multielement receiver with imaging capability. As the laser sweeps across the field of view of the CCD, the position and signal strength of each laser target spot is imaged, permitting a measure of bathymetry and reflectance. Using the CCD, a high-resolution slice through the reflected target spot radiance distribution is imaged so that system resolution can exceed the target spot size. The image of the target spot radiance distribution, modified by in-water scattering and target reflectance, provides new opportunities for image manipulation compared to typical underwater laser line scanning based systems. The simultaneous acquisition of reflectance and bathymetric maps permits discrimination capability between real objects of relief from scene reflectance variations.

Linear Mixture Modeling Approach for Estimating Cotton Canopy Ground Cover using Satellite Multispectral Imagery

Researchers have sought techniques for estimating plant canopy characteristics from remote sensing data that are not site-specific. In this study, a procedure for estimating cotton ( Gossypium spp.) canopy ground cover based on a linear mixture modeling approach was tested using Landsat multispectral imagery. Ground cover estimated from scene reflectance in the red and near-infrared wavebands was compared with corresponding field measurements. Statistical analysis of the results suggest that the procedure produced accurate, unbiased estimates of cotton canopy ground cover. The procedure appeared to implicitly compensate for the effects of shadows in the scene. These results agree with the findings of earlier studies and support the proposition that this procedure should be more universally applicable than commonly used empirical curve-fit techniques.

Satellite-sensor calibration verification with the cloud-shadow method

An atmospheric-correction method appropriate for high-spatial-resolution sensors that uses cloud-shaded pixels together with pixels in a neighboring region of similar optical properties is described. This cloud-shadow method uses the difference between the total radiance values observed at the sensor for these two regions, thus removing the nearly identical atmospheric radiance contributions to the two signals (e.g., path radiance and Fresnel-reflected skylight). What remains is largely due to solar photons backscattered from beneath the sea to dominate the residual signal. Normalization by the direct solar irradiance reaching the sea surface and correction for some second-order effects provides the remote-sensing reflectance of the ocean at the location of the neighbor region, providing a known ground target spectrum for use in testing the calibration of the sensor. A similar approach may be useful for land targets if horizontal homogeneity of scene reflectance exists about the shadow. Monte Carlo calculations have been used to correct for adjacency effects and to estimate the differences in the skylight reaching the shadowed and neighbor pixels.

Estimating Cotton Canopy Ground Cover from Remotely Sensed Scene Reflectance

AbstractMany agricultural applications require spatially distributed information on growth‐related crop characteristics that could be supplied through aircraft or satellite remote sensing. A study was conducted to develop and test a methodology for estimating plant canopy ground cover for cotton (Gossypium hirsutum L.) from scene reflectance. Previous studies indicated that a relatively simple relationship between ground cover and scene reflectance could be developed based on linear mixture modeling. Theoretical analysis indicated that the effects of shadows in the scene could be compensated for by averaging the results obtained using scene reflectance in the red and near‐infrared wavelengths. The methodology was tested using field data collected over several years from cotton test plots in Texas and California. Results of the study appear to verify the utility of this approach. Since the methodology relies on information that can be obtained solely through remote sensing, it would be particularly useful in applications where other field information, such as plant size, row spacing, and row orientation, is unavailable.

Two models for rapidly calculating bidirectional reflectance of complex vegetation scenes: Photon spread (PS) model and statistical photon spread (SPS) model

Two computationally fast Monte Carlo algorithms (models) are described for calculating the bidirectional reflectance factors (BRFs) from architecturally realistic canopies and from statistically characterized canopies. Both methods use a novel concept, called photon spreading. In this technique, the photon is treated as a particle when it travels through the canopy, but with each collision it is also treated like a wave, contributing to the canopy reflectance. This concept allows one to obtain canopy reflectance of the same accuracy with 30,000 photons that is obtained with 10 to 50 million photons in the traditional Monte Carlo method, leading to computational speed. The PS model takes about 10 min on a Pentium based PC and the SPS model a few minutes, for calculating BRF for 81 viewing directions. The two algorithms are compared against the SAIL model, another architecturally realistic model (for pure Lambertian scattering), Diana, and an analytical model for horizontal layers. The algorithms show good agreement. The potential of the statistical algorithm (SPS) for calculating canopy reflectance for large‐scale scenes is demonstrated by calculating the reflectance of a “forest scene” which consists of statistically defined vegetation extending over several acres.

Reflectance relations in row crop scenes

Abstract Models that relate composite reflectance to its components are useful for inferring crop growth information from measured scene reflectance. Radiation measurements in Thematic Mapper bands (TM1, TM2, TM3, and TM4) were made from cotton, soybean, sunflower and grain sorghum at three stages of growth and used to evaluate three reflectance models. Two models, AIRM1 and AIRM2, assumed that scene components contribute in an additive independent manner to composite reflectance. The third model, TRIM, assumes that radiation transmitted through the canopy interacts with bare soil in two scene components. The AIRM2 and TRIM models divide the composite reflectance into canopy, bare soil, and shadow components, but AIRM1 considers only canopy and bare soil. Ranking of models in order of decreasing accuracy for predicting composite reflectance in bands TM3 and TM4 was AIRM2, TRIM, and AIRM1. The AIRM1 and AIRM2 models estimated average TM3 reflectance at full plant cover between 1 and 4 per cent for all crop...

A simple geometrical model for analysing the spectral response of a citrus canopy using satellite images

Abstract A simple geometrical model has been proposed for a citrus canopy. We assume the citrus orchard to be a lattice structure, with the trees positioned at its points and where the composite-scene reflectance is the sum of the reflectance of its individual components as weighted by their respective surfaces within a unit area. The model has been used to analyse the citrus spectral response obtained from Landsat-5 TM images for winter and summer, where the status of the orchard is different. The correlations between spectral and geometrical data show the influence of per cent crop cover, shadows and background in the composite scene reflectance. We conclude that the summer images could be more useful than the winter ones for parcel classification according to per cent crop cover.

Water turbidity and perpendicular vegetation indices for paddy rice flood damage analyses

Reflectance factors were obtained for paddy rice at two sites in calibrated TM scenes acquired on two dates for the Kanto district of Japan—1 day after a typhoon inundated the paddies (6 August 1986, when the rice was at boot stage of development) and a month later (7 September 1986, during grain filling). A water turbidity index (WTI), analogous to soil brightness, and the perpendicular vegetation index (PVI) were computed for these data as WTI = 0.91 RED + 0.43 NIR and PVI = − 0.43 RED + 0.91 NIR , respectively. Rice yield measured for the inundation-damaged paddies and experimental nondamaged paddies at Tsukuba, Japan, were pooled and related to the PVI calculated from the respective TM scene and ground-measured reflectance factors. The ground and TM data coincided well, and the regression relation YIELD (t/ha) = 0.26 PVI − 3.0 for the pooled data. The relation between flood water turbidity at boot stage and yield was also examined for the two test sites; the more turbid the flood water, the greater the damage to the crop. The results demonstrate the possibility of spectrally assessing flood damage to rice and other crops.

Plant canopy information extraction from composite scene reflectance of row crops

As an aid in the interpretation of remotely sensed data from row crops with incomplete canopies, a model was developed that allowed the calculation of the fractions of sunlit soil, shaded soil, sunlit vegetation, and shaded vegetation for each resolution element in a scan of a remote sensor for a given set of conditions (plant cover, plant height/width ratio, row spacing, row orientation, time of day, day of year, latitude, and size of resolution element). Using measured representative reflectances of the four surfaces, composite reflectances were calculated as a function of view angle. Also, representative temperatures for each surface were used to simulate composite temperatures viewed by an IR scanner. With composite reflectances and temperatures known as a function of view angle, ways were explored to extract plant cover and plant temperature data from the composite data.

Coherent Optical Simulation of Radar Imagery

An analytical model of the synthetic aperture radar system is outlined. This model describes the radar image as a spatial two-dimensional bandpass filtering of the radar scene reflectivity density. The offset of the passband along the range frequency axis determines the major wavelength and geometrical properties of the simulated radar image. Through the use of a Fourier optical system, the photographs of radar target models and an aerial scene were bandpass filtered. The resulting images exhibit the characteristics of radar imagery, including texture (fading) and image dependence on object orientation. Use of the model and optical processing to simulate radar imagery from aerial photographs is discussed.