VNIR Range Research Articles

The Potential for Hyperspectral Imaging and Machine Learning to Classify Internal Quality Defects in Macadamia Nuts

Tree nuts are rich in nutrients, and global production and consumption have doubled during the last decade. However, nuts have a range of quality defects that must be detected and removed during post-harvest processing. Tree nuts can develop hidden internal discoloration, and current sorting methods are prone to subjectivity and human error. Therefore, non-destructive, real-time methods to evaluate internal nut quality are needed. This study explored the potential for VNIR (400–1000 nm) hyperspectral imaging to classify brown center disorder in macadamias. This study compared the accuracy of classifiers developed using images of kernels imaged in face-up and face-down orientations. Classification accuracy was excellent using face-up (>97.9%) and face-down (>94%) images using ensemble and linear discriminate models before and after wavelength selection. Combining images to form a pooled dataset also provided high accuracy (>90%) using artificial neural network and support vector machine models. Overall, HSI has great potential for commercial application in nut processing to detect internal brown centers using images of the outside kernel surface in the VNIR range. This technology will allow rapid and non-destructive evaluation of intact nut products that can then be marketed as a high-quality, defect-free product, compared with traditional methods that rely heavily on representative sub-sampling.

Lightning radiometry in visible and infrared bands

Calibrated measurements of lightning optical emissions are critical for both quantifying the impacts of lightning in our atmosphere and devising detection instruments with sufficient dynamic range capable of yielding close to 100% detection efficiency. However, to date, there is only a limited number of investigations that have attempted to take such calibrated measurements. In this work, we report the power radiated by lightning in both visible and infrared bands, assuming isotropic emission, and accounting for atmospheric absorption. More precisely, we report peak radiated power and total radiated energy in the combined visible plus near-infrared range (VNIR, 0.34–1.1 μm), around the Hα line (652–667 nm), and for the 2–2.5 μm infrared band. The estimated peak power and total energy radiated by negative cloud-to-ground return strokes in the VNIR range is 130 MW and 20 kJ, respectively. Additionally, we detected peak radiated powers of 12 and 0.19 MW in the Hα and infrared bands, respectively. We cross-reference the optical data set with peak current reported by a lightning detection network. The resulting trend is that optical power emitted around the Hα line scales with peak return stroke current according to a power law with exponent equal to 1.25. This trend, which should be approximately true across the entire visible spectrum, can be attributed to the plasma negative differential resistance of the lightning return stroke channel. We conclude by discussing the challenges in performing calibrated measurements of lightning optical power in different bands and comparing the results with previously-collected data with different experimental setups, observation conditions, and calibration methods.

Coating Condition Detection and Assessment on the Steel Girder of a Bridge through Hyperspectral Imaging

The organic coating of bridge steel girders is subjected to physical scratches, corrosion, and aging in natural weathering. The breakdown of the coating may cause serviceability and safety problems if left unnoticed. Conventional coating inspection is time-consuming and lacks information about the coating’s chemical integrity. A hyperspectral imaging method is proposed to detect the condition of steel coatings based on coating-responsive features in reflectance spectra. A field test was conducted on the real-world bridge, which shows obvious signs of degradation. The hyperspectral signature enables an assessment of the coating’s health and defect severity. The results indicated that the coating scratch can be effectively located in the domain of a hyperspectral image and the scratch depth can be determined by mapping a scratch depth indicator (SDI = R532 nm/R641 nm). Rust sources and products in steel girders can be identified by the unique spectral signatures in the VNIR range, and the rust stains (and thus stain areas) scattered on the coating can be pinpointed at pixel level by the chloride rust (CR) indicators >1.11 (CR = R733 nm/R841 nm). The chemical integrity of a topcoat is demonstrated by the short-wave infrared spectroscopy and the topcoat degradation can be evaluated by the decreased absorption at 8000 cm−1 and 5850 cm−1. Hyperspectral imaging enables faster and more reliable coating condition detection by the spectral features and provides an alternative for multi-object coating detection.

Evaluation of machine learning techniques with AVIRIS-NG dataset in the identification and mapping of minerals

Identifying valuable mineral resources is obligatory for human development and survival due to their scientific and economic implications. Hyperspectral Remote Sensing (HRS) is an efficient technique that empowers us to precisely identify and map altered minerals based on spectral absorption curves in the VNIR and SWIR range of the electromagnetic spectrum. Machine Learning Algorithms (MLAs) are a subclass of Artificial Intelligence (AI) that improves and automate HRS based lithological mapping using spectra based classification approach. The present study evaluates various MLAs in identifying and mapping hydrothermally altered and weathered minerals such as kaolinite, talc, kaosmec, and montmorillonite. The study was performed with the Airborne Visible Infrared Imaging Spectrometer- Next Generation (AVIRIS-NG) hyperspectral dataset of Jahazpur town of Bhilwara district in the state of Rajasthan, India (75° 9′ 57′' E, 25° 32′ 24″ N). The Spectral Angle Mapper (SAM) algorithm was applied to create a reference mineral distribution map for the target mineral classes. Further, the reference map has been verified with the field validation survey. A total of 828 pixels were identified as end member pixels. The obtained endmember pixels were further used for developing predictive machine learning based models for mapping the mineral resources using eight supervised classifiers, namely Support Vector Machine, k-Nearest Neighbor, Decision Tree, Random Forest, Logistic Regression, Artificial Neural Network, Linear Discriminant Analysis, and Naïve Bayes. The performance of the classifiers was measured using the kappa coefficient, overall accuracy, precision, recall, F1-score, Matthews’s correlation coefficient, and ROC curve. It was observed in the study that the Support Vector Machine outperformed all the classifiers in terms of overall accuracy and AVIRIS-NG data poses an excellent potential for mineral mapping using machine learning based models in a diverse mountainous area.

Fusion of spectrum and image features to identify Glycyrrhizae Radix et Rhizoma from different origins based on hyperspectral imaging technology

To identify Glycyrrhizae Radix et Rhizoma from different geographical origins, spectrum and image features were extracted from visible and near-infrared(VNIR, 435-1 042 nm) and short-wave infrared(SWIR, 898-1 751 nm) ranges based on hyperspectral imaging technology. The spectral features of Glycyrrhizae Radix et Rhizoma samples were extracted from hyperspectral data and denoised by a variety of pre-processing methods. The classification models were established by using Partial Least Squares Discriminate Analysis(PLS-DA), Support Vector Classification(SVC) and Random Forest(RF). Meanwhile, Gray-Level Co-occurrence matrix(GLCM) was employed to extract textural variables. The spectrum and image data were implemented from three dimensions, including VNIR and SWIR fusion, spectrum and image fusion, and comprehensive data fusion. The results indicated that the spectrum in SWIR range performed better classification accuracy than VNIR range. Compared with other four pre-processing methods, the second derivative method based on Savitzky-Golay(SG) smoothing exhibited the best performance, and the classification accuracy of PLS-DA and SVC models were 93.40% and 94.11%, separately. In addition, the PLS-DA model was superior to SVC and RF models in terms of classification accuracy and model generalization capability, which were evaluated by confusion matrix and receiver operating characteristic curve(ROC). Comprehensive data fusion on SPA bands achieved a classification accuracy of 94.82% with only 28 bands. As a result, this approach not only greatly improved the classification efficiency but also maintained its accuracy. The hyperspectral imaging system, a non-invasively, intuitively and quickly identify technology, could effectively distinguish Glycyrrhizae Radix et Rhizoma samples from different origins.

Automated ore microscopy based on multispectral measurements of specular reflectance. I – A comparative study of some supervised classification techniques

Automated mineralogy, including quantitative compositional and textural information, is a requirement for an efficient ore processing, and is comprised as an important input for geometallurgical planning. Classical ore microscopy is seen by many potential users as an outdated, time consuming, tool. Thus, SEM-based systems are often the choice for those who can afford them, in spite of their evident limitations for some minerals (e.g. for iron oxide ores). However, automated and quantitative mineral characterisation of metallic ores is also possible with optical (reflected light) microscopes, attaining similar performance at a much lower price than SEM systems, as shown by the AMCO (Automated Microscopic Characterisation of Ores) prototype. This system relies on the measurement of multispectral specular reflectance, R, on polished ore sections, to achieve automated identification of the ore species (reflectance is computed from grey levels of digital images acquired by automated scanning of the sample). The performance of this approach, supported by a multispectral reflectance database covering the VNIR range (370–1000 nm) built with the AMCO System, is analysed in this paper, comparing the reliability of different classification methods to achieve ore identification.The work outlined in this article focuses on checking the actual behaviour of four classification techniques, based respectively on spectral angle mapper, euclidean distance, Mahalanobis distance and linear discriminant analysis. The tests carried out reveal that the last two techniques are powerful tools to determine to which mineral corresponds a pixel based on its reflectance spectrum.The obtained results prove that automated multispectral optical microscopy is a reliable tool for mineral characterisation of common/industrial ores, with few exceptions (distinction of cassiterite and chromite, or sphalerite and wolframite). For optimisation of its performance, the multispectral information may be complemented with some additional criteria, such as paragenesis or type of deposit (e.g. to differentiate cassiterite from chromite). The reflectance database can also be enriched with measures of local ores, to account for existing reflectance variations due to local conditions or to learn new ores. The performance is particularly good for iron oxide ores (hematite, magnetite, goethite), a very important commodity. In the tests carried out, they are successfully identified in more than 99% of the cases.

Combining the use of CNN classification and strength‐driven compression for the robust identification of bacterial species on hyperspectral culture plate images

Huge streams of diagnostic images are expected to be produced daily in the emerging field of digital microbiology imaging because of the ongoing worldwide spread of Full Laboratory Automation systems. This is redefining the way microbiologists execute diagnostic tasks. In this context, the authors want to assess the suitability and effectiveness of a deep learning approach to solve the diagnostically relevant but visually challenging task of directly identifying pathogens on bacterial growing plates. In particular, starting from hyperspectral acquisitions in the VNIR range and spatial‐spectral processing of cultured plates, they approach the identification problem as the classification of computed spectral signatures of the bacterial colonies. In a highly relevant clinical context (urinary tract infections) and on a database of acquired hyperspectral images, they designed and trained a convolutional neural network for pathogen identification, assessing its performance and comparing it against conventional classification solutions. At the same time, given the expected data flow and possible conservation and transmission needs, they are interested in evaluating the combined use of classification and lossy data compression. To this end, after selecting a suitable wavelet‐based compression technology, they test coding strength‐driven operating points looking for configurations able to provably prevent any classification performance degradation.

Microscopía de reflectancia multiespectral: Aplicación al reconocimiento automatizado de menas metálicas

El artículo presenta el sistema CAMEVA, un sistema de microscopía de reflectancia multiespectral especialmente concebido para facilitar la identificación y caracterización de las fases minerales presentes en preparaciones pulidas de menas metálicas, así como para automatizar la realización de distintos tipos de análisis cuantitativos sobre ellas. El sistema CAMEVA permite obtener resultados similares a los de un sistema MEB (microscopía electrónica de barrido), superando algunas de sus limitaciones, como sus rígidas y costosas exigencias de infraestructura y especialización o la dificultad de distinguir especies polimorfas, pero a un coste notablemente inferior. Los ensayos realizados muestran que el sistema permite la identificación automatizada y fiable de las menas de interés industrial, a partir de la información multiespectral en el rango VNIR (visible e infrarrojo cercano, entre 400 y 1000 nm) recogida en una base de datos específica; esta base de datos, que incluye los 70 minerales de mayor interés, es fácilmente ampliable.

The use of hyperspectral imaging in the VNIR (400–1000 nm) and SWIR range (1000–2500 nm) for detecting counterfeit drugs with identical API composition

The risk of death from taking counterfeit drugs is now greater than the probability of dying from malaria and AIDS combined (at least half a million deaths each year). At the same time, counterfeit medicines are falsified more and more “skillfully”. According to WHO about 10% of counterfeit drugs are copies of original products.The methods of hyperspectral imaging and image analysis and processing were used to detect counterfeit drugs.Original Viagra® (Pfizer) and counterfeit tablets were compared. Hyperspectral imaging was used to acquire hyperspectral data cubes from both original and counterfeit tablets in the spectral range of 400–2500nm. Spectral parameters for both the original Viagra® and counterfeit drugs were compared. Grey-Level Co-Occurrence Matrix (GLCM) analysis and Principal Component Analysis (PCA) were performed.Hyperspectral analysis of the surface of the original Viagra® and counterfeit tablets demonstrates significant differences in reflectance (maximum difference for 1619.75nm). The GLCM contrast for the falsified drug is on average higher than for the original one 16±4%.GLCM contrast analysis enables to quantify homogeneity of distribution of tablet ingredients and enables to distinguish tablets with identical chemical composition. SWIR (1000–2500nm) hyperspectral imaging has a definite advantage over imaging in VNIR (400–1000nm) – higher wavelength is less sensitive to non-uniform illumination.

Near-ultraviolet bluing after space weathering of silicates and meteorites

Asteroid surface space weathering has been investigated both observationally and experimentally, mostly focusing on the effects on the visible–near infrared (VNIR, 0.4–2.5μm) spectral range. Here we present laboratory near-ultraviolet (NUV, 200–400nm) reflectance spectra of ion irradiated (30–400keV) silicates and meteorites as a simulation of solar wind ion irradiation. These results show that the induced alteration can reproduce the spread observed in the VNIR vs. NUV slope diagram for S-type asteroids. In particular, the well-known spectral reddening effect induced in the VNIR range is accompanied by a less known but stronger bluing effect at NUV wavelengths. Such trend was previously identified by Hendrix and Vilas (Hendrix, A.R., Vilas, F. [2006]. Astron. J., 132, 1396–1404) but only based on the comparison between observations and laboratory spectra of lunar materials. We attribute the NUV bluing, analogously to the VNIR reddening, to the formation of iron nanoparticles accompanied by structural modifications (amorphization) of surface silicates. We expect the evidence of weathering processes in the NUV part of spectra before these effects become observable at longer wavelengths, thus searching for the space weathering effects in the NUV range would allow establishing the extent of space weathering for very young asteroidal families.It will be important to include in future studies the NUV range both in the observations of specific classes of objects (e.g., the Vestoids) and in the laboratory spectra of meteorites and terrestrial analogues before and after space weather processing.

SpecCal: Novel software for in-field spectral characterization of high-resolution spectrometers

SpecCal software for the spectral calibration of high-resolution spectrometers is presented in this manuscript. The software, written in IDL 7.1, allows estimation of the channel central wavelength and the full width at half maximum of a selected spectrometer at several wavelengths across the VNIR range (350–1050 nm). This is achieved through comparison of the position and width of specific solar and terrestrial absorption features, as observed in the measured data, with those observed in simulated MODTRAN4 irradiance data. SpecCal is operated from a user-friendly graphical user interface that allows semiautomatic application of the spectral calibration algorithm at several wavelengths. The proposed software may be exploited as a useful in situ vicarious spectral calibration tool for field spectrometers operating in the VNIR range, which makes it possible to quickly analyze the spectral characteristics of the instruments and their possible variations with time.

Exponential Gaussian approach for spectral modeling: The EGO algorithm I. Band saturation

Curve fitting techniques are a widespread approach to spectral modeling in the VNIR range [Burns, R.G., 1970. Am. Mineral. 55, 1608–1632; Singer, R.B., 1981. J. Geophys. Res. 86, 7967–7982; Roush, T.L., Singer, R.B., 1986. J. Geophys. Res. 91, 10301–10308; Sunshine, J.M., Pieters, C.M., Pratt, S.F., 1990. J. Geophys. Res. 95, 6955–6966]. They have been successfully used to model reflectance spectra of powdered minerals and mixtures, natural rock samples and meteorites, and unknown remote spectra of the Moon, Mars and asteroids. Here, we test a new decomposition algorithm to model VNIR reflectance spectra and call it Exponential Gaussian Optimization (EGO). The EGO algorithm is derived from and complementary to the MGM of Sunshine et al. [Sunshine, J.M., Pieters, C.M., Pratt, S.F., 1990. J. Geophys. Res. 95, 6955–6966]. The general EGO equation has been especially designed to account for absorption bands affected by saturation and asymmetry. Here we present a special case of EGO and address it to model saturated electronic transition bands. Our main goals are: (1) to recognize and model band saturation in reflectance spectra; (2) to develop a basic approach for decomposition of rock spectra, where effects due to saturation are most prevalent; (3) to reduce the uncertainty related to quantitative estimation when band saturation is occurring. In order to accomplish these objectives, we simulate flat bands starting from pure Gaussians and test the EGO algorithm on those simulated spectra first. Then we test the EGO algorithm on a number of measurements acquired on powdered pyroxenes having different compositions and average grain size and binary mixtures of orthopyroxenes with barium sulfate. The main results arising from this study are: (1) EGO model is able to numerically account for the occurrence of saturation effects on reflectance spectra of powdered minerals and mixtures; (2) the systematic dilution of a strong absorber using a bright neutral material is not responsible for band deformation. Further work is still required in order to analyze the behavior of the EGO algorithm with respect to the saturation phenomena using more complex band shapes than pyroxene bands.

Sediment facies classification of a sandy shoreline by means of airborne imaging spectroscopy

Airborne imaging spectroscopy data (AISA Eagle and HyMap) were applied to classify the sediments of a sandy beach in seven sand type classes. On the AISA‐Eagle data, several classification strategies were tried out and compared with each other. The best classification results were obtained applying a linear discriminant classifier (LDC) in combination with feature selection based on sequential floating forward search (SFFS). The statistical LDC was used in a multiple binary approach. In the first step, the original bands were used in the classification, but transformation of the bands to wavelet coefficients enhanced the accuracy obtained. The combination of LDC with SFFS resulted in an overall accuracy of 82% (using three wavelet coefficients). Replacing the LDC with the non‐statistical SAM algorithm reduced the overall accuracy to 74% (using all bands or wavelet coefficients). When applying LDC, the optimal number of bands/wavelet coefficients to be used was defined: using more than two bands or three wavelet coefficients did not result in a higher classification accuracy. Finally, the HyMap data, featuring 126 bands in the VNIR‐SWIR range, were used to demonstrate that the VNIR range outperforms the SWIR range for this application.