Keshav Dev Singh

A field technique for rapid lithological discrimination and ore mineral identification: Results from Mamandur Polymetal Deposit, India

This work illustrates the efficiency of field spectroscopy for rapid identification of minerals in ore body, alteration zone and host rocks. The adopted procedure involves collection of field spectra, their processing for noise, spectral matching and spectral un-mixing with selected library end-members. Average weighted spectral similarity and effective peak matching techniques were used to draw end-members from library. Constrained linear mixture modelling technique was used to convolve end-member spectra. Linear mixture model was optimized based on root mean square error between field- and modelled-spectra. Estimated minerals and their abundances were subsequently compared with conventional procedures such as petrography, X-ray diffraction and X-ray fluorescence for accuracy assessment. The mineralized zone is found to contain azurite, galena, chalcopyrite, bornite, molybdenite, marcacite, gahnite, hematite, goethite, anglesite and malachite. The alteration zone contains chlorite, kaolinite, actinolite and mica. These mineral assemblages correlate well with the petrographic measurements (R2 = 0.89). Subsequently, the bulk chemistry of field samples was compared with spectroscopically derived cumulative weighted mineral chemistry and found to correlate well (R2 = 0.91–0.98) at excellent statistical significance levels (90–99%). From this study, it is evident that field spectroscopy can be effectively used for rapid mineral identification and abundance estimation.

Relevance of transformation techniques in rapid endmember identification and spectral unmixing: A hypespectral remote sensing perspective

One of the tedious and time consuming tasks related to hyperspectral data analysis is the identification of library candidates for spectral unmixing. In this study, we evaluated the relevance of different transformation procedures such as First Derivative (FD), Fast Fourier Transform (FFT), Discrete Wavelet Transform (DWT), Hilbert-Huang Transform (HHT) and S-transform (ST) in automated retrieval of library endmembers for linear spectral unmixing. The spectral similarity between the target and library candidates were estimated using Pearsons Correlation Coefficient (PCC) and student t-test based approach. Subsequently, these endmembers are used to estimate the fractional abundances by Fully Constrained Least Square Estimation (FCLSE) based Quadratic Programming (QP) optimization approach. The match between the target and modeled spectrum was calculated based on Root Mean Squared Error (RMSE) and spectral similarity scores estimated using Spectral Angle Mapper (SAM). In addition to RMSE and SAM scores, the simulation processing time and appropriateness of identified endmembers are considered to estimate the effectiveness of each transformation procedure. It is observed that DWT, HHT and ST based approaches are more efficient in identifying correct library endmembers than the FD and FFT based approaches.

Relevance of mineral texture on bidirectional reflectance and emission spectroscopy: Implications for geological remote sensing

This study evaluates the effect of grain-size, fabric, and composition of minerals on bidirectional reflectance and emission spectra. For this purpose, mineral grains of quartz (high albedo), feldspar (moderate albedo), and amphibole (low albedo) with four different grain-size ranges were analyzed using a goniometer system equipped with a spectroradiometer. Reflectance and emission spectra of bi-component mineral assemblages prearranged in seven different fabrics were analysed. It was observed that the parameters under investigation significantly influenced the intensity of reflection, scattering directions, spectral shape, and absorption depth, which in turn affected quantitative abundances estimated by techniques based on spectral similarity like Linear Mixture Model (LMM). It was observed that the error in abundances estimated by LMM varied from 3% to 15%. The authors conclude that the parameters under investigation, if neglected can introduce significant errors in geological interpretations.

A comparative study of signal transformation techniques in automated spectral unmixing of infrared spectra for remote sensing applications

ABSTRACT From geological and planetary exploration perspectives, automated sub-pixel classification of hyperspectral data is the most difficult task as it involves blind unmixing with library spectra of minerals. In this study, we demonstrate a procedure involving spectral transformation and linear unmixing to achieve the above task. For this purpose, infrared spectra of rocks from the spectral library, field, and remotely sensed hyperspectral image cube were used. Potential spectra of minerals for unmixing rock spectra were drawn from the library based on similarity of absorption features measured using Pearson correlation coefficient. Eight transformation techniques namely, first derivative, fast Fourier transform, discrete wavelet transform, Hilbert–Huang transform, crude low pass filter, S-transform, binary encoding, spectral effective peak matching, and two sparsity-based techniques (orthogonal matching pursuit, sparse unmixing via variable splitting, and augmented Lagrangian) were evaluated. Subsequently, minerals identified by above techniques were unmixed by linear mixture model (LMM) to decipher mineralogical composition and abundance. Results of LMM achieved using fully constrained least-square-estimation-based quadratic programming optimization approach were evaluated by conventional procedures such as X-ray diffraction and microscopy. In the case of image cube, endmembers derived using minimum noise fraction and pixel purity index were subjected to above procedure. It is evident that the discrete-wavelet-transformation-based approach produced excellent and meaningful results due to its flexibility in scaling the data and capability to handle noisy spectra. It is interesting to note that the adopted procedure could perform sub-pixel classification of image cube automatically and identify predominance of dolomite in limestone and sodium in alunite based on subtle differences in absorption positions.

A novel method for automatic minerals identification and their abundance estimation for material type discrimination using hyperspectral data

In hyperspectral remote sensing, the conventional endmember extraction and unmixing procedures are often complex and associated with uncertainties. In this work, we have designed an algorithm that uses Crude Low Pass Filter (CLoPF) and Pearsons Correlation Coefficient (PCC) to identify the endmember spectra from spectral library. Subsequently, a Non-Negativity Fully Constrained Least Square (NNFCLS) optimization approach was used to determine the fractional abundances of identified end-members. The efficacy of adopted procedure was estimated by Normalized Root Mean Squared Deviation (NRMSD), Spectral Angular Mapper (SAM), computation timing and appropriateness of identified candidates. It is observed that this procedure can be effectively used to resolve the mix-pixel spectra into library constituents and its fractional abundances.

Uncertainties in TIR hyperspectral image cube unmixing

The spectral unmixing in the thermal infrared (TIR) region is not linear case. For simplicity, the spectral mixing for pixel deconvolution is assumed to be linear, but in reality, the intimate mixture spectra are non-linear. The multiple scattering effects due to texture and fabric do affect the spectral shape and form. In this paper, the uncertainties in unmixing of TIR hyperspectral image cube are investigated.

Non-linear spectral unmixing of moon mineralogy mapper (M 3 ) data

Most of the approaches to solve the unmixing problem are based on the Linear Mixing Model (LMM) which is questionable. Therefore, nonlinear spectral model is generally used to study the effects of multiple scattering in the complex surfaces. In this paper, we have demonstrated the application of Radiative Transform Equation (RTE) based Hapke multi scattering model. The Hapke model based non-linear spectral unmixing is carried out on a Moon Mineralogy Mapper (M3) data. The values of six non-linear Hapkes model parameters are estimated using a MATLAB® based algorithm after optimizing the Hapke corrected M3-Endmembers and modeled spectra through minimum Root Mean Square Error (RMSE). After addressing the non-linear components in the image derived M3-Endmembers, the automated library candidate selection scheme is followed to estimate the corrected mineralogy over lunar surface.