Surface Reflectance Function Research Articles

Contributed Session III: Disentangling object color from illuminant color: The role of color shifts

Research has shown that disentangling surface and illuminant colors was possible based on various scene statistics. This study investigates the statistical cues induced by the chromatic effects of interreflections. We present a numerical analysis of ambiguous spectral pairs, in which the spectral power distribution of the illuminant in one scene matched the surface reflectance function in the other scene and vice versa. If the scenes are flat or convex and perfectly matte (Lambertian), the reflected light spectra of both cases are identical. However, the incident light undergoes interreflections for concave scenes. The spectral power of interreflections will be absorbed spectrally in an exponential way, dependent on the number of interreflections. We found that this causes systematic shifts towards the spectral reflectance peaks, resulting in brightness, saturation and hue shifts. Those paired cases' color differences (CIEDE2000) are so large that humans would be able to observe them if viewed simultaneously. In addition, we find that the color shifts cause qualitatively different gradients for chromatic materials and achromatic light and vice versa. Further psychophysical testing is necessary to see whether the different color shifts for the two cases can be recognized in isolation due to material or light properties. Moreover, the light densities and light vectors are spectrally different for these cases, creating different appearances of 3D objects in non-empty rooms.

Theoretical and Experimental Research on Laser-Induced Damage of Fused Silica Optics Due to Stimulated Brillouin Scattering

The damage behavior of fused silica optics induced by stimulated Brillouin scattering (SBS) was studied theoretically and experimentally based on a coupled-wave theory. The SBS-induced damage mechanism was systematically investigated by analyzing the evolutions of the Stokes-light intensity and induced stress as a function of laser pulse shape, thickness, and surface reflectance of the optics. The results show that the SBS-induced damage often occurs on the front surface of the fused silica optics. The Tophat beam has the most effect on an SBS-induced damage, followed by a focused Gaussian pulse. SBS effect is strongly thickness dependent, which is associated with the gain length. The SBS-induced damage can be efficiently suppressed by decreasing the surface reflection of the optics. Good agreements are obtained between the experimental and theoretical results, which strengthen our understanding of the damage mechanism of fused silica and provide a favorable reference for developing high-power laser systems.

Modeling surface appearance from a single photograph using self-augmented convolutional neural networks

We present a convolutional neural network (CNN) based solution for modeling physically plausible spatially varying surface reflectance functions (SVBRDF) from a single photograph of a planar material sample under unknown natural illumination. Gathering a sufficiently large set of labeled training pairs consisting of photographs of SVBRDF samples and corresponding reflectance parameters, is a difficult and arduous process. To reduce the amount of required labeled training data, we propose to leverage the appearance information embedded in unlabeled images of spatially varying materials to self-augment the training process. Starting from an initial approximative network obtained from a small set of labeled training pairs, we estimate provisional model parameters for each unlabeled training exemplar. Given this provisional reflectance estimate, we then synthesize a novel temporary labeled training pair by rendering the exact corresponding image under a new lighting condition. After refining the network using these additional training samples, we re-estimate the provisional model parameters for the unlabeled data and repeat the self-augmentation process until convergence. We demonstrate the efficacy of the proposed network structure on spatially varying wood, metals, and plastics, as well as thoroughly validate the effectiveness of the self-augmentation training process.

Generative constraints on image cues for perceived gloss

Image structure is generated by distinct physical sources that include 3-D surface geometry, surface reflectance, and the light field. A fundamental problem in midlevel vision involves understanding how, and how well, the visual system separates images into their underlying sources. A large body of recent work has shown that the perception of gloss exhibits significant dependencies on both 3-D surface geometry and the light field in which the surface is embedded. We recently proposed that these dependencies arise from the visual system relying on a restricted set of dimensions of specular image structure that are heuristically used to estimate surface gloss (Marlow, Kim, & Anderson, 2012). Here, we report a series of studies that manipulate surface geometry and the structure of the light field to prospectively test whether these image cues predict perceived gloss for a broad range of surface geometries and light fields. We identify generative constraints on each of our proposed cues that allow us to predict which cue will vary most for a given surface geometry, reflectance function, and light field. Our psychophysical data reveal that our generative manipulations succeed in promoting the significance of the intended cue, which in turn increases their predictive value in observers' judgments of gloss. Our results suggest that the perception of gloss in static, monocular images relies on a heuristic weighting of cues to specular image structure, and offer a unified explanation of why surface geometry and the light field modulate perceived gloss.

An Anisotropic Flat Index (AFX) to derive BRDF archetypes from MODIS

Spectral vegetation indices can be generalized as a function of surface reflectance with respect to wavelength. However, there is significant information on vegetation structure embedded in the anisotropic effects of the target. In this study, we describe and characterize a new vegetation index, the Anisotropic Flat Index (AFX) that captures this anisotropic scattering information and can be derived from the Moderate Resolution Imaging Spectroradiometer (MODIS) bidirectional reflectance distribution function (BRDF) product suite (MOD43A1). The AFX is created by normalization of net scattering magnitude (obtained from volumetric and geometric-optical scattering) with the isotropic scattering. The AFX summarizes the variability of basic dome-bowl anisotropic reflectance patterns of the terrestrial surface. A classification scheme for BRDF typology is created based on AFX archetypes that capture characteristic BRDF shape types. This study fully characterizes AFX in a number of steps. First, sensitivity to random noise and observation geometries is explored by comparing the AFX with other variables derived from field measurements that comprehensively sample the viewing hemisphere. Second, AFX is compared with normalized difference vegetation index (NDVI) values using field measurements from many ground campaigns, as well as global MODIS observations from EOS Land Validation Core Sites (LVCS). Third, a BRDF typology is developed by classification of an a priori database of BRDF archetypes from field measurements, and from MODIS observations that cover the full range of vegetation types from grasslands to closed forest (MCD43A). Fourth, the response of AFX to the parameter variability of canopy architectures and background optical properties for three vegetation types with discontinuous woody canopies is investigated through the use of a 5-Scale BRDF model simulation. Finally, global BRDF archetypes are mapped and discussed through the use of a global high-quality MODIS BRDF/albedo gap filled product (MCD43GF). The results show that the AFX summarizes BRDF archetypes and provides additional information on vegetation structure and other anisotropic reflectance characteristics of the land surface.

The Impact of Subscale Inhomogeneity on Oxygen A Band Cloud-Top Pressure Estimates: Using ESA’s MERIS as a Proxy for DSCOVR-EPIC

Medium Spectral Resolution Imaging Spectrometer (MERIS) oxygen A band measurements were used as a proxy for the Earth Polychromatic Imaging Camera (EPIC),to be launched on NASA’s Deep Space Climate Observatory (DSCOVR). The high spatial resolution of MERIS (1 × 1 km2) is exploited to study the effects of subscale spatialheterogeneity of clouds on the cloud-top pressure retrieved at the coarser spatial resolutionof EPIC (10 × 10 km2). In general, for a sub-scale cloud fraction less than 1, a shift of cloud-top pressure toward the middle atmosphere is found, with a low-bias for highclouds and a high-bias for low clouds. In addition, the deviation is found to be a function of surface reflectance. The subscale variability of fully clouded EPIC pixels causes a weak underestimation of cloud-top pressure, when compared to averaged high-resolution retrievals.

표면반사율 모델링을 위한 새로운 N차원 기저함수

일반적으로 표면반사율과 분광반사율을 N차원의 칼라 코드로부터 정확히 복원하기 위해서는 N개의 기저함수가 필요하다. 전형적인 렌더링 응용에서 벡터의 덧셈, 스칼라 곱셈 및 성분별 곱셈에 대한 벡터 연산이 이질동형이라고 가정하고 광원의 중첩, 광원-표면간 상호간섭 및 상호반사와 같은 물리적인 연산을 모델링하지만 벡터 연산이 물리적인 현상을 그대로 반영하는 것은 아니다. 그러나 만약 기저함수가 특성함수로써 제한된다면 표면반사율과 분광반사율의 사상 결과 및 벡터들은 렌더링에서 물리적인 연산인 이질이형을 유지하게 된다. 본 논문은 새로운 N차원의 특성함수를 제안하고 N차원의 기저함수로 근사화된 먼셀 칼라 칩에 대하여 제안한 알고리즘의 정확성을 평가할 것이다. The N basis functions are typically chosen so that Surface reflectance functions(SRFs) and spectral power distributions (SPDs) can be accurately reconstructed from their N-dimensional vector codes. Typical rendering applications assume that the resulting mapping is an isomorphism where vector operations of addition, scalar multiplication, component-wise multiplication on the N-vectors can be used to model physical operations such as superposition of lights, light-surface interactions and inter-reflection. The vector operations do not mirror the physical. However, if the choice of basis functions is restricted to characteristic functions then the resulting map between SPDs/SRFs and N-vectors is anisomorphism that preserves the physical operations needed in rendering. This paper will show how to select optimal characteristic function bases of any dimension N (number of basis functions) and also evaluate how accurately a large set of Munsell color chips can approximated as basis functions of dimension N.

An introduction to image-based 3D surface reconstruction and a survey of photometric stereo methods

This paper provides an introduction to photometric methods for image-based 3D shape reconstruction and a survey of photometric stereo techniques. We begin with taxonomy of active and passive shape acquisition techniques. Then we describe the methodical background of photometric 3D reconstruction, define the canonical setting of photometric stereo (Lambertian surface reflectance, parallel incident light, known illumination direction, known surface albedo, absence of cast shadows), discuss the 3D reconstruction of surfaces from local gradients, summarize the concept of the bidirectional reflectance distribution function (BRDF), and outline several important empirically and physically motivated reflectance models. We provide a detailed treatment of several generalizations of the canonical setting of photometric stereo, namely non-distant light sources, unknown illumination directions, and, in some detail, non-Lambertian surface reflectance functions. An important special case is purely specular reflections, where an extended light source allows capturing a surface that consists of perfectly specular surface patches. Linear combinations of purely Lambertian and purely specular reflectance components are favorably used for reconstructing smooth surfaces and also human skin. Nonuniform surface reflectance properties are estimated based on a simultaneous 3D reconstruction and determination of the locally variable parameters of the reflectance function based on a multitude of images. Assuming faceted surfaces, the effective resolution of the 3D reconstruction result can be increased to some extent beyond that of the underlying images. Other approaches separate specular and diffuse reflectance components based on polarization data or color information. The specular reflections can be used additionally to estimate the direction from which the surface is illuminated. Finally, we describe methods to combine photometric 3D reconstruction techniques with active and passive triangulation-based approaches.

From Color Sensor Space to Feasible Reflectance Spectra

The interaction of light and object surfaces generates color signals in the visible band that are responsible for digital acquisition system outputs. Inverting this mapping from the sensor space back to the wavelength domain is of great interest for many applications. Since 1964, with the idea of Cohen to exploit the characteristic of smoothness of surface reflectance functions, a lot of work has been done in the analysis, synthesis and recovering of spectral information using linear models. The general use of such models is for the establishment of a one-to-one relationship between sensor's data and reflectance spectrum, with the requirement of ensuring the quality of the recovered spectrum in terms of physical feasibility and naturalness. In this paper, we propose a solution to correct the outcome of a generic recovery method, in order to take into account quality constrains. Our strategy assumes the smoothness of the solution of the recovery method, an assumption implicitly satisfied from the adoption of linear models to represent reflectance functions.

Reflecting on a room of one reflectance

We present a numerical analysis of rendered pairs of rooms, in which the spectral power distribution of the illuminant in one room matched the surface reflectance function in the other room, and vice versa. We ask whether distinction between the rooms is possible and on what cues this discrimination is based. Using accurately rendered three-dimensional (3D) scenes, we found that room pairs can be distinguished based on indirect illumination, as suggested by A. L. Gilchrist and A. Jacobsen (1984). In a simulated color constancy scenario, we show that indirect illumination plays a pivotal role as areas of indirect illumination undergo a smaller appearance change than areas of direct illumination. Our study confirms that indirect illumination can play a critical role in surface color recovery and shows how computer rendering programs, which model the light-object interaction according to the laws of physics, are valuable tools that can be used to analyze and explore what image information is available to the visual system from 3D scenes.

A computational strategy exploiting genetic algorithms to recover color surface reflectance functions

Information about the spectral reflectance of a color surface is useful in many applications. Assuming that reflectance functions can be adequately approximated by a linear combination of a small number of basis functions, we address here the recovery of a surface reflectance function, given the tristimulus values under one or more illuminants. Basis functions presenting different characteristics and cardinalities are investigated, and genetic algorithms are employed to optimize the estimation. Our analysis of a variety of standard datasets provides information about the ability of each set of basis functions we used to model generic reflectance spectra.

Multispectral Imaging: How Many Sensors Do We Need?

The surface reflectance functions of natural and manmade surfaces are invariably smooth. It is desirable to exploit this smoothness in a multispectral imaging system by using as few sensors as possible to capture and reconstruct the data. In this paper we investigate the minimum number of sensors to use, while also minimizing reconstruction error. We do this by deriving different numbers of optimized sensors, constructed by transforming the characteristic vectors of the data, and simulating reflectance recovery with these sensors in the presence of noise. We find an upper limit to the number of optimized sensors one should use, above which the noise prevents decreases in error. For a set of Munsell reflectances, captured under educated levels of noise, we find that this limit occurs at approximately nine sensors. We also demonstrate that this level is both noise and dataset dependent, by providing results for different magnitudes of noise and different reflectance datasets.

Plane‐parallel biases computed from inhomogeneous Arctic clouds and sea ice

Monte Carlo simulations of the expected influence of nonuniformity in cloud structure and surface albedo on shortwave radiative fluxes in the Arctic atmosphere are presented. In particular, plane‐parallel biases in cloud albedo and transmittance are studied for nonabsorbing, low‐level, all‐liquid stratus clouds over sea ice. The “absolute bias” is defined as the difference between the cloud albedo or transmittance for the uniform or plane‐parallel case, and the albedo or transmittance for nonuniform conditions with the same mean cloud optical thickness and the same mean surface albedo, averaged over a given area (i.e., bias > 0 means plane‐parallel overestimates). Ranges of means and standard deviations of input parameters typical of Arctic conditions are determined from the First International Satellite Cloud Climatology Project (ISCCP) Regional Experiment Artic Cloud Experiment (FIRE/ACE)/Surface Heat Budget of the Arctic Ocean (SHEBA)/Atmospheric Radiation Measurement Program (ARM) experiment, a cooperative effort of the Department of Energy, NASA, NSF, the National Oceanic and Atmospheric Administration, the Office of Naval Research, and the Atmospheric Environment Service. We determine the sensitivity of the bias with respect to the following: domain averaged means and spatial variances of cloud optical thickness and surface albedo, shape of the surface reflectance function, presence of a scattering layer under the clouds, and solar zenith angle. The simulations show that the biases in Arctic conditions are generally lower than in subtropical stratocumulus. The magnitudes of the absolute biases are unlikely to exceed 0.02 for albedo and 0.05 for transmittance. The “relative bias” expresses the absolute bias as a percentage of the actual cloud albedo or transmittance. The magnitude of the relative bias in albedo is typically below 2% over the reflective Arctic surface, while the magnitude of the relative bias in transmittance can exceed 10%.

Daylight, biochrome surfaces, and human chromatic response in the Fourier domain.

We first report Fourier analyses of a collection of 348 daylight spectral power distributions and 1,695 biochrome surface reflectance functions. The power spectra of the daylights are low pass with more than 99% of spectral power below 1 cycle/300 nm and 99.9% below 3 cycles/300 nm. The power spectra of reflectance functions are also low pass with more than 99% of spectral power below 4 cycles/300 nm and 99.9% below 11 cycles/300 nm. Consequently, the resulting color signals are typically low pass with, for our samples, an estimated frequency cutoff of 5 cycles/300 nm. Theoretical and experimental data concerning human chromatic response in the frequency domain show that this limit corresponds to the highest frequency that the color system can resolve. The implications for normal and abnormal human color vision are discussed.

Computational theory of color transparency: recovery of spectral properties for overlapping surfaces

A computational theory of color transparency for color images containing X junctions is described. This theory is based on physical models of color transparency under conditions of additive or subtractive color mixture that describe the relationship among four colors at an X junction. Algorithms are derived for recovering transmittance and surface reflectance functions of a transparent medium from a set of sensor responses at an X junction. The algorithms are based on the ability to describe surface reflectance and transmittance functions by using a linear combination of orthogonal basis functions. We also address algorithms for determination of depth ordering of overlapping surfaces and the type of color mixture by checking the physical realizability of recovered functions.

Reflectance and Shape from Images Using a Collinear Light Source

In this paper, a novel technique, called the photogeometric technique, is presented for surface reflectance extraction and surface recovery from an image sequence of a rotating object illuminated under a collinear light source (where the illuminant direction of the light source lies on or near the viewing direction of the camera). The rotation of the object is precisely controlled. The object surface is assumed to be smooth and uniform. The technique first computes the 3D locations of some surface points which give singular brightness values and builds the surface reflectance function by extracting the brightness values of these surface points from the image sequence. Then the technique uses the surface reflectance function and two images of the surface to recover surface depth and orientation simultaneously. The technique has been tested on real images of surfaces with different reflectance properties and geometric structures. The experimental results and comprehensive analysis show that the proposed technique is efficient and robust.

Sampling procedural shaders using affine arithmetic

Procedural shaders have become popular tools for describing surface reflectance functions and other material properties. In comparison to fixed resolution textures, they have the advantage of being resolution-independent and storage-efficient. While procedural shaders provide an interface for evaluating the shader at a single point, it is not easily possible to obtain an average value of the shader together with accurate error bounds over a finite area. Yet the ability to compute such error bounds is crucial for several interesting applications, most notably heirarchical area sampling for global illumination, using the finite element approach, and for generation of textures used in interactive computer graphics. Using affine arithmetic for evaluating the shader over a finite area yields a tight, conservative error interval for the shader function. Compilers can automatically generate code for utilizing affine arithmetic from within shaders implemented in a dedicated language such as the RenderMann shading language.

Color constancy. III. General linear recovery of spectral descriptions for lights and surfaces.

We present a color-constancy algorithm that uses quantum-catch data from reflected lights to recover surface reflectance functions and illuminant spectral power distributions. The algorithm recovers both surface and light-source spectral properties simultaneously. The method works in all situations that were handled by the earlier two-stage algorithms of Maloney and Wandell [J. Opt. Soc. Am. A 3, 29 (1986)] and D'Zmura and Iverson [J. Opt. Soc. Am. A 9, 490 (1992); 10, 2148, 2166 (1993); 11, 1970 (1994)]. In addition, the method handles problems that lie outside the scope of earlier algorithms. Using this method, a trichromatic visual system can recover, when provided adequate information, spectral descriptions of arbitrarily high accuracy for lights and surfaces. We determine conditions under which bilinear models can be used to recover color properties uniquely with the new procedure, and we formulate an algorithm for checking whether a particular bilinear model provides perfect color constancy. This research extends our analysis of linear methods for color constancy begun earlier [J. Opt. Soc. Am. A 10, 2148, 2166 (1993)].

Color constancy II Results for two-stage linear recovery of spectral descriptions for lights and surfaces

Our analysis of color constancy in a companion paper [J. Opt. Soc. Am A 10, 2148 (1993)] provided an algorithm that lets one test how well linear color constancy schemes work. Here we present the results of applying the algorithm to a large parametric class of color constancy problems involving bilinear models that relate photoreceptoral spectral sensitivities, surface reflectance functions, and illuminant spectral power distributions. These results, supported by simulation and further analysis, provide a detailed classification of two-stage linear methods for recovering the spectral properties of reflectances and illuminants from reflected lights.

Variational approach to interreflection in color images

Interreflections affect the colors of surfaces as they appear in images. The light reflected by one surface then impinges upon a second surface changes the color of the overall illumination that it receives and that hence the color of the light that it reflects. Both the relative colors and positions of the two surfaces affect the result. We analyze the physics of the interreflection process and extract constraints on the possible surface reflectances, ambient illumination, and geometric configuration of the surfaces. By using the calculus of variations, a finite-dimensional model of reflectance, and a one-bounce model of interreflection, we express these constraints as a set of equations that are then solved for the surface spectral reflectance functions of the surfaces, the spectrum of the ambient illumination, and local interreflection factors related to the scene geometry. The interreflection factors express how the image is altered by interreflection effects and can be used to produce an image shaded as it would appear had there been no interreflection; the surface reflectance functions provide color constancy. Although it is more complex than some previous analyses of interreflection, the variational approach is more general and relaxes some restrictive assumptions concerning the type of illumination and the number of surfaces.