Josée Lévesque

Spectral unmixing of hyperspectral imagery for mineral exploration: comparison of results from SFSI and AVIRIS

Hyperspectral image data sets acquired near Cuprite, Nevada, in 1995 with the Short-Wave Infrared (SWIR) Full Spectrum Imager (SFSI) and in 1996 with the Airborne Visible-Infrared Imaging Spectrometer (AVIRIS) are analysed with a spectral unmixing procedure and the results compared. The nominal pixel centre spacings are 1.0 by 1.5 m for SFSI and 16.2 by 18.1 m for AVIRIS across track and along track, respectively; the region imaged by SFSI is a small portion of the full AVIRIS scene. Both data cubes have nominal spectral band centre spacings of approximately 10 nm. The image data, converted to radiance units, are atmospherically corrected and converted to surface reflectances. Spectral end members are extracted automatically from the two data sets; those representing mineral species common to both are compared to each other and to reference spectra obtained with a field instrument, the Portable Infrared Mineral Analyser (PIMA). The full sets of end members are used in a constrained linear unmixing of the respective hyperspectral image cubes. The resulting unmixing fraction images derived from the AVIRIS and SFSI data sets for the minerals alunite, buddingtonite, kaolinite, and opal correlate well, with correlation coefficients ranging from 0.75 to 0.91, after compensation for shadowing and misregistration effects.

A demonstration of hyperspectral image exploitation for military applications

Defence R&D Canada has been studying the military applications of hyperspectral imagery for a number of years. In other unrelated efforts, the Canadian remote sensing community has also been active in developing hyperspectral algorithms for civilian use. This civilian technology has many potential military applications. In an effort to demonstrate the potential of these defence and civilian technologies to the Canadian Forces, Defence R&D Canada has initiated the Hyperspectral Image Exploitation (HYMEX) Technology Demonstration Project with the collaboration of Canadian industries, academic organizations and other government departments, the project is evaluating and integrating exploitation algorithms into a suite of tools oriented towards the needs of the Canadian Forces. This paper describes the project activities to date and presents some preliminary results.

Mapping mine tailing surface mineralogy using hyperspectral remote sensing

Acid mine drainage resulting from mine tailings poses an environmental threat. An important initial step towards the reclamation of mine tailing sites is to detect the presence of acid-generating, sulphide-rich minerals and determine their spatial distribution. In this study, the potential of hyperspectral remote sensing for characterizing mine tailings is investigated. The study site is located in northern Ontario, Canada, and the data were collected with PROBE-1, an imaging spectrometer that covers the visible, near-infrared, and shortwave-infrared spectral ranges. The results indicate that using the weakly constrained linear spectral unmixing technique PROBE-1 data can provide information on mineral compositions of the tailing surface. The spatial locations and associations of acid-generating source minerals such as pyrite and pyrrhotite along with their oxidation products (e.g., copiapite, jarosite, ferrihydrite, goethite, and hematite) can provide information about the distribution of oxidation processes at the site. This remote mapping technique can be very valuable when attempting to identify abandoned mine-waste sites and the potential risk they might present where there are no a priori knowledge and field samples available.

Monitoring mine tailings revegetation using multitemporal hyperspectral image data

This paper investigates the use of hyperspectral remote sensing imagery in the 400–2500 nm wavelength range for the extraction of information suitable for monitoring mine tailings revegetation. The objectives were twofold: (i) demonstrate the usefulness of fractional texture for monitoring mine tailings revegetation using visible and near-infrared (VNIR) hyperspectral data, and (ii) investigate the benefit of adding the short-wave infrared (SWIR) bands. Compact Airborne Spectrographic Imager (casi) data were acquired over the Copper Cliff mine tailings impoundment area in the VNIR bands during the summers of 1996 and 1998. In addition, Probe-1 data were collected in the VNIR-SWIR region during the summer of 1999. Surface reflectance was retrieved from the three datasets, and spectra of the 1996 and 1998 casi datasets were resampled to match the 1999 Probe-1 spectral sampling characteristics, which resulted in 30 bands covering the 450–890 nm range. The three datasets were concatenated into one file, and 30 endmember spectra were automatically selected. Constrained linear spectral unmixing was performed using the 30 endmembers, which were then grouped into the six endmember categories, namely water, lime, fresh and oxidized tailings, and low and high photosynthetic vegetation. Image fractions were then normalized and image texture was extracted from the total vegetation fraction. Total vegetation fraction (high/low photosynthetic), total tailings fraction (fresh/oxidized), and texture of the vegetation fraction were used in a K-mean unsupervised classification, which produced the best results using seven classes (78.13% overall accuracy, Kappa coefficient of 0.74). Classification results were validated using a set of 34 ground estimates of vegetation cover and tailings. The full set of 128 bands of the 1999 Probe-1 dataset was used to investigate the contribution of the SWIR bands to monitoring the reclamation of mine tailings. A new vegetation endmember was identified as plant litter, which in many cases replaces areas labelled as low photosynthetic vegetation when using the VNIR bands only. Oxidized tailings could be separated into jarosite and goethite endmembers and lime into agricultural lime (CaMgCO3) and calcium oxide (CaO), also known as quicklime.

Potential Discrimination of Toxic Industrial Chemical Effects on Poplar, Canola and Wheat, Detectable in Optical Wavelengths 400–2450 nm

This research examined the spectral response of poplar (Populus deltoides, Populus trichocarpa), wheat (Triticum aestivum), and canola (Brassica napus) leaves subjected to fumigation with gaseous phase toxic industrial chemical gases (TICs). The gases include ammonia (NH3), sulphur dioxide (SO2), hydrogen sulphide (H2S), chlorine (Cl2), and hydrogen cyanide (HCN). This study aimed to determine if: (1) vegetation subjected to TICs could be distinguished from background vegetation during varying growth stages and environmental stresses; and, (2) different TICs could be distinguished based on the spectral response of vegetation. The results showed that both environmental and TICs induced similar spectral features inherent to plants, which are related primarily to chlorophyll and water loss. These features include pigments in the visible and cellulose, lignin, lipids starches, and sugars in the SWIR. Although no specific spectral features could be tied to individual TICs an analysis of the data using vegetation indices showed that the TICs and environmental stresses result in diagnostic trends from healthy mature to highly stressed leaves. In addition combinations of specific indices could be used to distinguish the effects of NH3, SO2, Cl2 and their effect from that of other treatments of the study. The continued goal for this research program is to develop a remote detection capability for hazardous events such as a toxic gas leak. Our findings at the leaf level suggest that damage can be detected within 48 hrs and should last for an extended period. Thus, the next experimental step is to test if the results shown here at the leaf level can also be detected with airborne and satellites systems.

A method for monitoring mine tailings revegetation using hyperspectral remote sensing

This paper investigates the use of airborne hyperspectral remote sensing imagery in the 400-nm to 900-nm spectral range for the extraction of information suitable for monitoring mine tailings revegetation. Compact Airborne Spectrographic Imager (CASI) data were acquired over the Copper Cliff mine tailings impoundment area in the VNIR bands during the summers of 1996 and 1998, and Probe 1 data were collected in the VNIR/SWIR bands during the summer of 1999. Endmember fractions of water, lime, fresh and oxidised tailings, low and high photosynthetic vegetation were obtained using constrained linear spectral unmixing. Vegetation fraction, tailings fraction and texture of the vegetation fraction were used in a K-Mean unsupervised classification, which produced the best results using seven classes (78.13% overall accuracy) and captured the vegetation cover from dense homogenous to low density patched cover.

Reflectance spectra of the boreal forest over mineralized sites

Detecting areas of near surface mineralization from spectral reflectance of geochemically stressed vegetation is very challenging in Canadas boreal forest region. This paper provide an analysis of the field spectra of two boreal forest species growing on mineralized areas. Their results show that there is little difference in reflectance between the mineralized and background sites. Less than 1.5 nm difference at the red edge inflection point was observed between the mineralized and background sites. This indicates that the differences in geochemical values are not large enough to produce significant spectral variations between sires. Therefore, detection from airborne spectrometers will be difficult.

Assessment of noise reduction of hyperspectral imagery using a target detection application

This article presents an evaluation of a previously proposed noise reduction technique for hyperspectral imagery with regard to its use in remote sensing applications. Target detection from hyperspectral imagery was selected as an example for the evaluation. A hyperspectral datacube acquired using the airborne Shortwave Infrared Full Spectrum Imager (SFSI)-II with man-made targets deployed in the scene of the datacube was tested. In addition to an evaluation using the receiver operating characteristic (ROC) curve approach, we used a spectral unmixing technique to generate the fraction images of the target materials, measured the area of the targets derived from the datacube before and after applying the noise reduction technology, and then compared the derived target areas to the real targets to assess the detectability of the targets. The area ratio between a derived target and the real target was used as the criterion in the evaluation. The evaluation results show that the noise reduction technique can help to better serve remote sensing applications. The small targets that cannot be detected from the original datacube were detected after the noise reduction using the technology.

Target detection from noise-reduced hyperspectral imagery using a spectral unmixing approach

We assess the effectiveness of a previously proposed noise reduction technology for hyperspectral imagery to examine whether it can better serve remote sensing applications after noise reduction using the technology. Target detection from hyperspectral imagery using a spectral unmixing approach is selected as an example in the assess- ment. A hyperspectral datacube acquired using an airborne short-wave- infrared Full Spectrum Image II with man-made targets in the scene of the datacube is tested. Three criteria are proposed and used to evaluate the detectability of the targets derived from the datacube before and after noise reduction. The evaluation results show that the detectability of the targets is significantly improved after noise reduction using the technol- ogy. The targets not detected from the original datacube are detected with high confidence after noise reduction using the technology. A noise reduction technique that is based on a smoothing approach is also evaluated for the sake of comparison to the proposed noise reduction technology. It also improves the detectability of the targets, but is less effective than the proposed noise reduction technology.

Spectral angle mapper based assessment of detectability of man-made targets from hyperspectral imagery after SNR enhancement

This paper assesses the effectiveness of a signal-to-noise ratio (SNR) enhancement technology for hyperspectral imagery to examine whether it can better serve remote sensing applications. A hyperspectral data set acquired using an airborne Short-wave-infrared Full Spectrum Image II with man-made targets in the scene of the data set was tested. Spectral angle mapper and end-members of different target materials were used to measure the superficies of the targets and to assess the detectability of the targets before and after applying the SNR enhancement technology to the data set. The experimental results show that small targets, which cannot be detected in the original data set due to inadequate SNR and low spatial resolution, can be detected after the SNR of the data set is enhanced.