Images of outdoor scenes captured in bad weather suffer from poor contrast. Under bad weather conditions, the light reaching a camera is severely scattered by the atmosphere. The resulting decay in contrast varies across the scene and is exponential in the depths of scene points. Therefore, traditional space invariant image processing techniques are not sufficient to remove weather effects from images. We present a physics-based model that describes the appearances of scenes in uniform bad weather conditions. Changes in intensities of scene points under different weather conditions provide simple constraints to detect depth discontinuities in the scene and also to compute scene structure. Then, a fast algorithm to restore scene contrast is presented. In contrast to previous techniques, our weather removal algorithm does not require any a priori scene structure, distributions of scene reflectances, or detailed knowledge about the particular weather condition. All the methods described in this paper are effective under a wide range of weather conditions including haze, mist, fog, and conditions arising due to other aerosols. Further, our methods can be applied to gray scale, RGB color, multispectral and even IR images. We also extend our techniques to restore contrast of scenes with moving objects, captured using a video camera.

/pdf/contrast-restoration-of-weather-degraded-images-orn61jh9qu.pdf

Contrast restoration of weather degraded images

Imaging polarimetry has emerged over the past three decades as a powerful tool to enhance the information available in a variety of remote sensing applications. We discuss the foundations of passive imaging polarimetry, the phenomenological reasons for designing a polarimetric sensor, and the primary architectures that have been exploited for developing imaging polarimeters. Considerations on imaging polarimeters such as calibration, optimization, and error performance are also discussed. We review many important sources and examples from the scientific literature.

/pdf/review-of-passive-imaging-polarimetry-for-remote-sensing-3kk5ut05ha.pdf

Review of passive imaging polarimetry for remote sensing applications

We present an approach for easily removing the effects of haze from passively acquired images. Our approach is based on the fact that usually the natural illuminating light scattered by atmospheric particles (airlight) is partially polarized. Optical filtering alone cannot remove the haze effects, except in restricted situations. Our method, however, stems from physics-based analysis that works under a wide range of atmospheric and viewing conditions, even if the polarization is low. The approach does not rely on specific scattering models such as Rayleigh scattering and does not rely on the knowledge of illumination directions. It can be used with as few as two images taken through a polarizer at different orientations. As a byproduct, the method yields a range map of the scene, which enables scene rendering as if imaged from different viewpoints. It also yields information about the atmospheric particles. We present experimental results of complete dehazing of outdoor scenes, in far-from-ideal conditions for polarization filtering. We obtain a great improvement of scene contrast and correction of color.

/pdf/polarization-based-vision-through-haze-5cx4tcr4pn.pdf

Polarization-based vision through haze

Colour cameras mimic the human eye and record only a small part of the information contained in the incoming light. Modern image sensing techniques, which subdivide the light spectrally or record information about the polarization of the incoming light, can extract much more information for applications ranging from biological studies to remote sensing1,2,3,4,5. Spectral imaging techniques6 typically rely on filters or interferometers combined with scanning or subsampling to record a spectral image ‘cube’ (which has wavelength as a third dimension). This leads to inefficient use of the incoming light and/or long recording times. Here, we show that surface plasmons enable direct recording of spectral image cubes in a single exposure. By texturing metal surfaces at the nanometre scale, incoming light is converted to surface plasmons and can then be separated according to wavelength and polarization, before being recoupled to light through subwavelength apertures that illuminate individual photodetector elements. This photon-sorting capability provides a new approach for spectral and polarimetric imaging with extremely compact device archictures.

Plasmonic photon sorters for spectral and polarimetric imaging

Haze removal or dehazing is a challenging ill-posed problem that has drawn a significant attention in the last few years. Despite this growing interest, the scientific community is still lacking a reference dataset to evaluate objectively and quantitatively the performance of proposed dehazing methods. The few datasets that are currently considered, both for assessment and training of learning-based dehazing techniques, exclusively rely on synthetic hazy images. To address this limitation, we introduce the first outdoor scenes database (named O-HAZE) composed of pairs of real hazy and corresponding haze-free images. In practice, hazy images have been captured in presence of real haze, generated by professional haze machines, and O-HAZE contains 45 different outdoor scenes depicting the same visual content recorded in haze-free and hazy conditions, under the same illumination parameters. To illustrate its usefulness, O-HAZE is used to compare a representative set of state-of-the-art dehazing techniques, using traditional image quality metrics such as PSNR, SSIM and CIEDE2000. This reveals the limitations of current techniques, and questions some of their underlying assumptions.

/pdf/o-haze-a-dehazing-benchmark-with-real-hazy-and-haze-free-2pqbysbrgn.pdf

O-HAZE: A Dehazing Benchmark with Real Hazy and Haze-Free Outdoor Images

Imaging polarimetry has the potential to be a key sensor technology in a number of target detection applications. Imaging polarimeters measure the polarization state of light emitted from and/or reflected from scenes. The light is polarized because of the geometry, roughness and material properties of the objects embedded in the scene. This added information enhances conventional intensity and color imagery, potentially surpassing its performance in low contrast situations. In this paper, we describe a divided aperture MWIR imaging polarimeter which acquires multiple polarization images simultaneously. At the heart of the polarimeter is a relay lens set that produces four identical images on a single focal plane array from a single aperture. Each of the four images measures a different orientation of linear polarization, 0, 45, 90 and 135 degrees. The relay lens set operates inside of a pour fill Dewar of a InSb MWIR camera. The design and calibration method for the polarimeter are given along with example data sets taken from the air over Huntsville, AL.

A division of aperture MWIR imaging polarimeter

A method for the correction of systematic errors generated by large orientational and retardance errors in the polarization optics in the dual rotating retarder polarimeter is presented. Small orientational and retardance errors ( 10% in some matrix elements). We incorporate correction terms for large orientation and retardance errors into the dual rotating retarder data reduction algorithm. Using these data reduction algorithms and a calibration step, the associated systematic errors are calculated and removed from the measured Mueller matrix. This procedure is especially useful for spectral and multi-wavelength systems in which the retardance and often the orientation of the retarders are wavelength dependent. The equations, the procedure to calculate the orientations of the polarization elements and the retardances of the retardation elements, and the method to correct for any errors are presented here. The effect of these errors on the calculated Mueller matrix elements and their correction is shown analytically and through experimental data taken on an infrared spectropolarimeter.

Mueller matrix algorithms

In recent years there has been an increasing interest in optical imaging through scattering media, including through body tissue, battlefield smoke, fog, rain, and mist, and muddy water. When an image is formed of an object embedded in a scattering medium, the image is comprised of the superposition of a sharp image plus a diffuse background. The light that propagates the shortest distance between the object and image plane forms the sharp image. The diffuse image is the result of the scattered light taking a random path between the object and image. In order to observe the sharp image formed by the weakly scattered light, the diffuse background must be removed from the image. Polarization based discrimination between weakly scattered and scattered light is based on the fact that weakly scattered light retains its incident polarization state whereas diffuse scattered light carries a random polarization state. In this paper, we will describe a polarimetric sensor and laboratory setup for imaging through scattering media. We will report the results from the laboratory measurements demonstrating imaging through several types of scattering media. Finally, we will discuss application of this technique to medical imaging, imaging through battlefield obscurants, and wireless communications.

Polarization imaging through scattering media

A technique to measure linear diattenuation and retardance spectra of infrared materials in transmission is described. A sample is rotated between two stationary linear polarizers in the sample compartment of a Fourier transform infrared spectrometer. The intensity modulation that results from the rotation of the sample is Fourier analyzed, and the linear diattenuation and linear retardance of the sample are calculated from the Fourier series coefficients for each wavelength. The advantages of this technique include immunity of the measurement to instrumental polarization, to circular diattenuation, and to circular retardance. The rotating sample polarimeter does not require retarders. Compensation for systematic errors from polarizers with diattenuation less than one is included in the data reduction. This technique is useful for the calibration of retarders and polarizers and hence for the bootstrap calibration of more elaborate polarimeters such as Mueller matrix polarimeters. We present as an example of the technique the caliration spectra 3–14 μm of an infrared quasi-achromatic retarder whose fast axis orientation oscillates with wavelength.

David B. Chenault

Papers

Review of passive imaging polarimetry for remote sensing applications

A division of aperture MWIR imaging polarimeter

Mueller matrix algorithms

Polarization imaging through scattering media

Measurements of linear diattenuation and linear retardance spectra with a rotating sample spectropolarimeter.