Pushbroom Hyperspectral Imager Research Articles

Aerial Imaging of Fluorescent Dye in the Near Shore

Abstract Aerial images are used to quantify the concentration of fluorescent Rhodamine water tracing (WT) dye in turbid and optically deep water. Tracer releases near the shoreline of an ocean beach and near a tidal inlet were observed with a two-band multispectral camera and a pushbroom hyperspectral imager, respectively. The aerial observations are compared with near-surface in situ measurements. The ratio of upwelling radiance near the Rhodamine WT excitation and emission peaks varies linearly with the in situ dye concentrations for concentrations <20 ppb (r2 = 0.70 and r2 = 0.85–0.88 at the beach and inlet, respectively). The linear relationship allows for relative tracer concentration estimates without in situ calibration. The O(1 m) image pixels resolve complex flow structures on the inner shelf that transport and mix tracer.

Geometric Correction of PHI Hyperspectral Image without Ground Control Points

Geometric correction without ground control points (GCPs) is a very important topic. Conventional airborne photogrammetry is difficult to implement in areas where the installation of GCPs is not available. The technical of integrated GPS/INS systems providing the positioning and attitude of airborne systems is a potential solution in such areas. This paper first states the principle of geometric correction based on a combination of GPS and INS then the error of the geometric correction of Pushbroom Hyperspectral Imager (PHI) without GCP was analysed, then a flight test was carried out in an area of Damxung, Tibet. The experiment result showed that the error at straight track was small, generally less than 1 pixel, while the maximum error at cross track direction, was close to 2 pixels. The results show that geometric correction of PHI without GCP enables a variety of mapping products to be generated from airborne navigation and imagery data.

Modeling and compensating non-uniformity in push-broom NIR hyperspectral imaging system

In this paper, a novel pixel-based model for the response of push-broom near infrared (NIR) hyperspectral cameras based on focal-plane array (FPA) photo-detectors is proposed. The model focuses on practical issues affecting the system response, such as the spatial and spectral non-uniformity (NU) observed in NIR hyperspectral imaging systems, and its novelty relies on considering the operating temperature of the FPA as a signal perturbing the response of the entire system. These effects have been estimated by experimental procedures. The applicability and accuracy of the proposed model has been assessed by mitigating for the spatial and spectral non-uniform responses of a real NIR hyperspectral imaging sequence.

基于多核支持向量机的高光谱影像非线性混合像元分解

As the mixed pixel decomposition based on linear spectrum models has lower decomposition accuracy and the nonlinear spectrum model is difficult to be established,a nonlinear mixed pixel decomposition method for the hyperspectral imagery was proposed based on the posterior probability of Multiple Kernel Support Vector Machine(MKSVM).On the basis of the SVM,the multiple kernel function formed by linear weighted combination was taken to replace the single kernel and the simple multiple kernel learning was used to solve the weights iteratively to achieve the classification.Then,the output values of the classifier were converted to pairwise coupling probabilities by thesigmoid function and then to the multi-class posterior probability.Finally,the hyperspectral imagery decomposition was achieved through the posterior probability.The results from experiments of two push-broom Hyperspectral Imagers(PHIs)show that the classification accuracies of hyperspectral imagery nonlinear mixed pixel decomposition based on MKSVM reach 95.62% and 91.51%,respectively,the Root Mean Square Errors(RMSEs)are reduced to 11.15%and 7.55%,and both are less than 15%.In conclusion,the influence of mixed pixel on hyperspectral imagery classification is eliminated,and the classification accuracy is increased.

Application of hyperspectral imaging technology to discriminate different geographical origins of Jatropha curcas L. seeds

A rapid and non-invasive method was investigated to identify the geographical origin of Jatropha curcas L. seeds in China by using near-infrared hyperspectral imaging technique on the wavelength between 874 and 1734nm. Two hundred and forty J. curcas L. seed samples from four different geographical origins (Jiangsu, Sichuan, Hainan and Taiwan) in China were studied and all of them were scanned by a pushbroom hyperspectral imaging system. Then the obtained data sets were analyzed by spectral and image processing technique respectively. Successive projections algorithm (SPA) was used for selecting effective wavelengths. Dimension reduction was carried out on the region of interest (ROI) image by principal component analysis (PCA). The first principal component (PC) explained over 92% of variances of all spectral bands. Gray-level co-occurrence matrix (GLCM) analysis was implemented on the principal component (PC) image to extract 5 textural feature variables in total. Moreover, 7 morphological features of samples were computed additionally. Then least squares-support vector machine (LS-SVM) classification models were built based on the extracted spectral, textural, morphological, combined spectral and textural, combined spectral and morphological, combined textural and morphological, combined spectral, textural and morphological features, respectively. The satisfactory results show the correct discrimination rate of 93.75% for the prediction samples based on spectral and morphological features. The study demonstrated that hyperspectral image technique can be a reliable tool for discriminating different geographical origins of J. curcas L. seeds. The above results indicated that this objective and non-destructive method can be utilized for quality control purposes and seed breeding in future.

Non-invasive measurement of frog skin reflectivity in high spatial resolution using a dual hyperspectral approach.

BackgroundMost spectral data for the amphibian integument are limited to the visible spectrum of light and have been collected using point measurements with low spatial resolution. In the present study a dual camera setup consisting of two push broom hyperspectral imaging systems was employed, which produces reflectance images between 400 and 2500 nm with high spectral and spatial resolution and a high dynamic range.Methodology/Principal FindingsWe briefly introduce the system and document the high efficiency of this technique analyzing exemplarily the spectral reflectivity of the integument of three arboreal anuran species (Litoria caerulea, Agalychnis callidryas and Hyla arborea), all of which appear green to the human eye. The imaging setup generates a high number of spectral bands within seconds and allows non-invasive characterization of spectral characteristics with relatively high working distance. Despite the comparatively uniform coloration, spectral reflectivity between 700 and 1100 nm differed markedly among the species. In contrast to H. arborea, L. caerulea and A. callidryas showed reflection in this range. For all three species, reflectivity above 1100 nm is primarily defined by water absorption. Furthermore, the high resolution allowed examining even small structures such as fingers and toes, which in A. callidryas showed an increased reflectivity in the near infrared part of the spectrum.Conclusion/SignificanceHyperspectral imaging was found to be a very useful alternative technique combining the spectral resolution of spectrometric measurements with a higher spatial resolution. In addition, we used Digital Infrared/Red-Edge Photography as new simple method to roughly determine the near infrared reflectivity of frog specimens in field, where hyperspectral imaging is typically difficult.

Hyperspectral Imaging as a Visual Quality Assessment Method for South African Herbal Teas

Rooibos (Aspalathus linearis) and Honeybush (Cyclopia species) are teas indigenous to South Africa that are popularly consumed locally as well as internationally for its purported health benefits. Studies have confirmed anti-oxidant, antimutagenic and anticancer activity mainly due to the presence of polyphenols. Hyperspectral imaging (HSI) integrates conventional spectroscopy and imaging to obtain spectral and spatial information from a sample. The potential of short wave infrared (SWIR) hyperspectral imaging in combination with chemometric data analysis as a rapid quality control method for commercially important herbal teas was investigated. Images were acquired with the sisuChema short wave infrared (SWIR) hyperspectral pushbroom imaging system at a spectral range of 920 – 2514nm. Principal Component Analysis (PCA) was performed to remove background pixels and to visualize the data. Classes (i.e. Rooibos or Honeybush) were interactively assigned using the score images and plots. Partial Least Squares Discriminant Analysis (PLS-DA) classification models developed were subsequently used to accurately predict the species identity of commercially available herbal tea samples introduced as an external dataset, producing a visually interpretable result. The results showed that the labeling information is correct but that Honeybush tea is in most cases present in very low amounts. It is evident that hyperspectral imaging is an objective and non-destructive quality control method that can be successfully used to determine the botanical species included in herbal teas in addition to giving an indication of the abundance of each species in polyherbal tea mixtures.

Chemical-free assessment and mapping of major constituents in beef using hyperspectral imaging

Developing a rapid and non-destructive method for food safety and quality monitoring has become a crucial request from the meat industry. Hyperspectral imaging technique provides extraordinary advantages over the traditional imaging and spectroscopy techniques in food quality evaluation due to the spatial and spectral information that it can offer. In this study, a laboratory-based pushbroom hyperspectral imaging system in reflectance mode was developed in the near infrared (NIR) range (900–1700nm) for non-invasive determination of the major chemical compositions of beef. Beef samples collected from different breeds were scanned by the system followed by traditional assessment of their chemical composition by using the ordinary wet-chemical methods. The extracted spectral data and the measured quality parameters were modeled by partial least squares regression (PLSR) for predicting water, fat and protein contents yielding a reasonable accuracy with determination coefficients (RP2) of 0.89, 0.84 and 0.86 concomitant with standard error of prediction (SEP) of 0.46%, 0.65% and 0.29%, respectively. Some image processing algorithms were developed and the most relevant wavelengths were selected to visualize the predicted chemical constituents in each pixel of the hyperspectral image yielding the spatially distributed visualizations of the sample contents. The results were promising and implied that hyperspectral imaging technique associated with appropriate chemometric multivariate analyses has a great potential for simultaneous assessment of various chemical constituents without using hazardous chemical reagents.

NIR hyperspectral imaging as non-destructive evaluation tool for the recognition of fresh and frozen–thawed porcine longissimus dorsi muscles

Abstract Identification of fresh and frozen–thawed meat is an important authenticity issue, although a challenging task. The potential application of a pushbroom hyperspectral imaging system in the near-infrared (NIR) range (900–1700 nm) was explored as a rapid and non-destructive technique for the investigation of meat freshness. The effect of standard freeze–thaw routines was studied in pork samples from the longissimus dorsi muscle. Partial least squares discriminant analysis (PLS-DA) models were used to distinguish between fresh and frozen–thawed samples. Optimal wavelengths were selected and used for sample discrimination with reduced spectral data and image processing. Classification models with reduced spectral data achieved an overall correct classification of 100% for an independent set of samples. An image processing algorithm was also developed for visualizing the classification results. The best classification model obtained was successfully applied to the images to produce classification maps with high overall accuracy. Results confirmed the possibility of developing a fast and reliable system for discrimination between fresh and frozen–thawed pork based on reflectance in the NIR wavelength range. Industrial relevance The application of NIR hyperspectral analysis could enable the development of a rapid, reliable and non-destructive method for authentication of fresh meat samples for the benefit of the retail sector and the consumer.

Hyperspectral imaging in the quality control of herbal medicines – The case of neurotoxic Japanese star anise

Illicium verum (Chinese star anise) dried fruit is popularly used as a remedy to treat infant colic. However, instances of life-threatening adverse events in infants have been recorded after use, in some cases due to substitution and/or adulteration of I. verum with Illicium anisatum (Japanese star anise), which is toxic. It is evident that rapid and efficient quality control methods are of utmost importance to prevent re-occurrence of such dire consequences. The potential of short wave infrared (SWIR) hyperspectral imaging and image analysis as a rapid quality control method to distinguish between I. anisatum and I. verum whole dried fruit was investigated. Images were acquired using a sisuChema SWIR hyperspectral pushbroom imaging system with a spectral range of 920–2514nm. Principal component analysis (PCA) was applied to the images to reduce the high dimensionality of the data, remove unwanted background and to visualise the data. A classification model with 4 principal components and an R2X_cum of 0.84 and R2Y_cum of 0.81 was developed for the 2 species using partial least squares discriminant analysis (PLS-DA). The model was subsequently used to accurately predict the identity of I. anisatum (98.42%) and I. verum (97.85%) introduced into the model as an external dataset. The results show that SWIR hyperspectral imaging is an objective and non-destructive quality control method that can be successfully used to identify whole dried fruit of I. anisatum and I. verum. In addition, this method has the potential to detect I. anisatum whole dried fruits within large batches of I. verum through upscaling to a conveyor belt system.

Determination of total viable count (TVC) in chicken breast fillets by near-infrared hyperspectral imaging and spectroscopic transforms

Near infrared (NIR) hyperspectral imaging (HSI) and different spectroscopic transforms were investigated for their potential in detecting total viable counts in raw chicken fillets. A laboratory-based pushbroom hyperspectral imaging system was utilized to acquire images of raw chicken breast fillets and the resulting reflectance images were corrected and transformed into hypercubes in absorbance and Kubelka–Munck (K–M) units. Full wavelength partial least regression models were established to correlate the three spectral profiles with measured bacterial counts, and the best calibration model was based on absorbance spectra, where the correlation coefficients (R) were 0.97 and 0.93, and the root mean squared errors (RMSEs) were 0.37 and 0.57log10 colony forming units (CFU) per gram for calibration and cross validation, respectively. To simplify the models, several wavelengths were selected by stepwise regression. More robustness was found in the resulting simplified models and the model based on K–M spectra was found to be excellent with an indicative high ratio of performance to deviation (RPD) value of 3.02. The correlation coefficients and RMSEs for this model were 0.96 and 0.40log10 CFU per gram as well as 0.94 and 0.50log10 CFU per gram for calibration and cross validation, respectively. Visualization maps produced by applying the developed models to the images could be an alternative to test the adaptability of a calibration model. Moreover, multi-spectral imaging systems were suggested to be developed for online applications.

Non-destructive assessment of microbial contamination in porcine meat using NIR hyperspectral imaging

Abstract Temperature fluctuation during cold storage of meat products usually leads to undesirable microbial growths, which affect the overall product quality. In this study, a pushbroom hyperspectral imaging system in the near-infrared (NIR) range (900–1700 nm) as a rapid and non-destructive technique was exploited for determining the total viable count (TVC) and psychrotrophic plate count (PPC) in chilled pork during storage. Fresh pork samples from the longissimus dorsi muscle were obtained directly from a commercial slaughtering plant, and stored in the refrigerated temperatures at 0 °C and 4 °C for 21 days. Every 48 h, a NIR hyperspectral image in the reflectance mode was acquired directly for each sample. The TVC and PPC were determined simultaneously by classical microbiological plating methods and multivariate statistical models for predicting contamination and spoilage conditions in the samples were then developed. Partial least squares regression (PLS) was applied to fit the spectral information extracted from the samples to the logarithmic values of TVC and PPC. The best regressions were obtained with R2 of 0.86 and 0.89 for log (TVC) and log (PPC), respectively. The most important wavelengths were then selected for regression and for spatial visualization of contamination. Results are encouraging and show the promising potential of hyperspectral technology for detecting bacterial spoilage in pork and tracking the increase of microbial growth of chilled pork during storage at different temperatures. Industrial relevance: A novel method based on hyperspectral imaging technique has been successfully developed for determining the total viable count (TVC) and psychrotrophic plate count (PPC) in chilled pork during storage non-destructively for the meat industry.

Prediction of firmness and soluble solids content of blueberries using hyperspectral reflectance imaging

Currently, blueberries are inspected and sorted by color, size and/or firmness (or softness) in packing houses, using different inspection techniques like machine vision and mechanical vibration or impact. A new inspection technique is needed for effectively assessing both external features and internal quality attributes of individual blueberries. This paper reports on the use of hyperspectral imaging technique for predicting the firmness and soluble solids content (SSC) of blueberries. A pushbroom hyperspectral imaging system was used to acquire hyperspectral reflectance images from 302 blueberries in two fruit orientations (i.e., stem and calyx ends) for the spectral region of 500–1000nm. Mean spectra were extracted from the regions of interest for the hyperspectral images of each blueberry. Prediction models were developed based on partial least squares method using cross validation and were externally tested with 25% of the samples. Better firmness predictions (R=0.87) were obtained, compared to SSC predictions (R=0.79). Fruit orientation had no or insignificant effect on the firmness and SSC predictions. Further analysis showed that blueberries could be sorted into two classes of firmness. This research has demonstrated the feasibility of implementing hyperspectral imaging technique for sorting blueberries for firmness and possibly SSC to enhance the product quality and marketability.

Study on the Hierarchical Classification of Urban Manmade Objects Based on Spectral Features.

An experiment fleight was made by pushbroom hyperspectral imager PHI in Shahe, Beijng during June, 1997. The main aim was to study the key technology of urban land use and construct classification by using hyperspectral technology. By analyzing the acquired 15 bands visible near infrared data, we try to classify the urban manmade objects through the analysis of the spectral signatures. Because of the complexity of the spectral features of urban manmade objects, it is difficult to recognize the diffrent classes by traditional classification algorithm. Based on the analysis of the spectral features, we proffer a strategy of classification that calssify the classes by composite algoritnm hierarchically. The result states itself that this method performs better than the familiar method.

Hyperspectral imaging for the age estimation of blood stains at the crime scene

The age estimation of blood stains can provide important information on the temporal aspects of a crime. As previously shown, visible spectroscopy of blood stains can successfully be used for their age estimation. In the present study we evaluated the feasibility to use hyperspectral imaging for this purpose. Visible reflectance spectra of blood stains were recorded using a pushbroom hyperspectral imaging system. From these spectra, the relative amounts of oxyhemoglobin, methemoglobin and hemichrome within the blood stains were derived. By comparison of the hemoglobin derivative fractions with a reference dataset, the age of blood stains up to 200 days old was estimated. The absolute error of the age estimation task increased with age, with a median relative error of 13.4% of the actual age. To test the practical applicability of this method, a simulated crime scene was analyzed, in which blood stains of several ages were deposited. Hyperspectral imaging combined with the proposed analysis provided insight in the absolute age of the blood stains. Additionally, the blood stains were clustered based on their hemoglobin derivative fractions, without the use of a reference dataset. Results demonstrated that the order of formation of blood stains can be determined, even under unknown environmental circumstances, when no proficient reference dataset is available. These findings are an important step toward the practical implementation of blood stain age estimation in forensic casework.

Classification of urban vegetation patterns from hyperspectral imagery: hybrid algorithm based on genetic algorithm tuned fuzzy support vector machine

Hyperspectral imagery has been widely used in terrain classification for its high resolution. Urban vegetation, known as an essential part of the urban ecosystem, can be difficult to discern due to high similarity of spectral signatures among some land-cover classes. In this paper, we investigate a hybrid approach of the genetic-algorithm tuned fuzzy support vector machine (GA-FSVM) technique and apply it to urban vegetation classification from aerial hyperspectral urban imagery. The approach adopts the genetic algorithm to optimize parameters of support vector machine, and employs the K-nearest neighbor algorithm to calculate the membership function for each fuzzy parameter, aiming to reduce the effects of the isolated and noisy samples. Test data come from push-broom hyperspectral imager (PHI) hyperspectral remote sensing image which partially covers a corner of the Shanghai World Exposition Park, while PHI is a hyper-spectral sensor developed by Shanghai Institute of Technical Physics. Experimental results show the GA-FSVM model generates overall accuracy of 71.2%, outperforming the maximum likelihood classifier with 49.4% accuracy and the artificial neural network method with 60.8% accuracy. It indicates GA-FSVM is a promising model for vegetation classification from hyperspectral urban data, and has good advantage in the application of classification involving abundant mixed pixels and small samples problem.

Application of Visible and Near Infrared Hyperspectral Imaging to Differentiate Between Fresh and Frozen–Thawed Fish Fillets

The potential of visible and near infrared (VIS/NIR) hyperspectral imaging was investigated as a rapid and nondestructive technique to determine whether fish has been frozen–thawed. A total of 108 halibut (Psetta maxima) fillets were studied, including 48 fresh and 60 frozen–thawed (F-T) samples. Regarding the F-T samples, two speeds of freezing (fast and slow) were tested. The hyperspectral images of fillets were captured using a pushbroom hyperspectral imaging system in the spectral region of 380 to 1,030 nm. All images were calibrated for reflectance, followed by the minimum noise fraction rotation to reduce the noise. A region-of-interest (ROI) at the image center was selected, and the average spectral data were generated from the ROI image. Dimension reduction was carried out on the ROI image by principal component analysis. The first three principal components (PCs) explained over 98 % of variances of all spectral bands. Gray-level co-occurrence matrix analysis was implemented on the three PC images to extract 36 textural feature variables in total. Least squares-support vector machine classification models were developed to differentiate between fresh and F-T fish based on (1) spectral variables; (2) textural variables; (3) combined spectral and textural variables, respectively. Satisfactory average correct classification rate of 97.22 % for the prediction samples based on (3) was achieved, which was superior to the results based on (1) or (2). The results turned worse when different freezing rates were taken into consideration to classify three groups of fish. The overall results indicate that VIS/NIR hyperspectral imaging technique is promising for the reliable differentiation between fresh and F-T fish.

Investigation of fungal development in maize kernels using NIR hyperspectral imaging and multivariate data analysis

Near infrared (NIR) hyperspectral imaging and hyperspectral image analysis were evaluated for their potential to track changes in fungal contamination on and fungal activity immediately under the surface of whole maize kernels (Zea mays L.) infected with Fusarium verticillioides. Hyperspectral images of clean and infected kernels were acquired using a SisuChema hyperspectral pushbroom imaging system with a spectral range of 1000–2498 nm at predetermined time intervals after infection. Background, bad pixels and shading of acquired absorbance images were removed using exploratory principal component analysis (PCA). When plotting PC4 against PC5, with percentage sum of squares (%SS) 0.49% and 0.34%, three distinct clusters were apparent in the score plot and this was associated with degree of infection. Loading line plots, with prominent peaks at 1900 nm and 2136 nm, confirmed that the source of variation was due to changes in starch and protein. Partial least squares (PLS) regression models, with time as the Y variable, were calculated and also indicated that changes over time were apparent. Variable importance plots (VIP) confirmed the peaks observed in the PCA loading line plots. More systematic future experiments are needed to confirm this, but it can already be concluded that early detection of fungal contamination and activity is possible.

Push-broom hyperspectral imaging for elemental mapping with glow discharge optical emission spectrometry

Glow discharge optical emission spectroscopy (GDOES) has been recognized for allowing direct solid sample elemental analysis with high depth resolution. However, the lateral resolution it affords has been historically restricted to some millimetres or the diameter of the sputtered area. Recently, it was shown that one can obtain laterally resolved information from within the sputtered area by operating the discharge in pulsed power mode. The newly available data dimensions require a new approach to the collection of the GDOES signal with lateral (to recover X and Y positions), spectral (to qualify and quantify elemental information), and temporal resolution (to improve lateral resolution and allow depth profiling). Previous studies have utilized spectral imagers of whisker-broom and staring geometries. In this study we characterize the advantages and disadvantages of using a push-broom geometry hyperspectral imager for GDOES elemental mapping. The results show that the higher light throughput of the push-broom geometry allows faster image acquisition times, compared to other spectral imaging systems with the same components, and thus maintain depth resolutions below 10 nm.

New methodology of hyperspectral information extraction and accuracy assessment based on a neural network

Spectral analysis is an important research topic in the field of hyperspectral remote sensing. In this paper, the pixel vectors of hyperspectral imagery are taken as research object. A new method of spectral analysis and hyperspectral image information extraction were studied based on a Neural Network (NN). An application of this method in disaster measurement of wheat stripe rust disease is described. The experimental data was hyperspectral imagery of wheat-infected with stripe rust and came from the Push-broom hyperspectral imager (PHI). The NN model of Self-organizing Feature Maps (SOFM) was introduced to cluster and analysis of the disease severity. By means of computing the spectral index data (SID) and spectral angle data (SAD) as well as spectral average reflectance data (ARD) of hyperspectral imagery in experimental area, the three kinds of two-dimensional data matrixes were used as the inputs of the SOFM model. After iterative learning and self-organized clustering, the outputs of the model were mapped in 3-dimensional severity space of wheat stripe rust. The computed data and image data in the experimental area were simulated through net-training. The trees and soil in the experimental area are easily interpreted by comparison of sample’s spectral curves. The disease and healthy sampling points were identified by “Red-edge” positions. Then, the degrees of disease severity were established by spectral threshold analysis of samples in near-infrared bands. Finally, a cross correlogram spectral matching technique was used to assess the accuracy. The study results show that the disease of wheat stripe rust was severe. The whole trial spot was divided into six feature classes, including four severity degrees of wheat stripe rust disease, they are healthy, light, mid and heavy. These results are reliable and accurate.