An innovative multivariate strategy for HSI-NIR images to automatically detect defects in green coffee

Research on variety identification of common bean seeds based on hyperspectral and deep learning

Accurate, fast and non-destructive identification of varieties of common bean seeds is important for the cultivation and efficient utilization of common beans. This study is based on hyperspectral and deep learning to identify the varieties of common bean seeds non-destructively. In this study, the average spectrum of 3078 hyperspectral images from 500 varieties was obtained after image segmentation and sensitive region extraction, and the Synthetic Minority Over-sampling Technique (SMOTE) was used to achieve the equilibrium of the samples of various varieties. A one-dimensional convolutional neural network model (IResCNN) incorporating Inception module and residual structure was proposed to identify seed varieties, and Support Vector Machine (SVM), K-Nearest Neighbor (KNN), VGG19, AlexNet, ResNet50 were established to compare the identification effect. After analyzing the effects of multiple spectral preprocessing methods on the model, the study selected Savitzky-Golay smoothing correction (SG) for spectral preprocessing and extracted 66 characteristic wavelengths using Successive Projections Algorithm (SPA) as inputs to the discriminative model. Ultimately, the IResCNN model achieved the highest accuracy of 93.06 % on the test set, indicating that hyperspectral technology can accurately identify bean varieties, and the study provides a correct method of thinking for the non-destructive classification of multi-species small-sample bean varieties.

Chemical analysis and NIR spectroscopy in the determination of the origin, variety and roast time of Mexican coffee

Coffee is a product whose quality and price are associated with its geographical, genetic and processing origin; therefore, the development of analytical techniques to authenticate the above mentioned is important to avoid adulteration. The objective of this study was to compare conventional analytical methods with NIR technology for the authentication of roasted and ground coffee samples from different producing regions in Mexico (origins) and different varieties. A second objective was to determine, under the same processing conditions, if roasting times can be differentiated by using this technology. A total of 120 samples of roasted and ground commercial coffee were obtained from the states of Chiapas, Oaxaca, Tabasco and Veracruz in Mexico, 30 locally available samples per state. Samples from Veracruz included three different varieties, grown on the same farm and processed under the same conditions. One of these varieties was selected to evaluate the chemical composition of samples roasted at 185 °C using four different roasting times (15, 17, 19 and 21 min). Samples from different producing regions showed significant differences (P < 0.05) in fat content (from 7.45 ± 0.42% in Tabasco to 18.40 ± 2.95% in Chiapas), which was associated with the altitude of coffee plantations (Pearson's r = 0.96). The results indicate that NIR technology generates sufficient useful information to authenticate roasted and ground coffee from different geographical origins in Mexico and different varieties from the same coffee plantation, with similar results to those obtained by conventional analytical methods.

A comprehensive insight on modern green analyses for quality control determination and processing monitoring in coffee and cocoa seeds

Green analysis is defined as the analysis of chemicals in a manner where sample extraction and analysis are performed with least amounts of steps, low hazardous materials, while maintaining efficiency in terms of analytes detection. Coffee and cocoa represent two of the most popular and valued beverages worldwide in addition to their several products i.e., cocoa butter, chocolates. This study presents a comprehensive overview of green methods used to evaluate cocoa and coffee seeds quality compared to other conventional techniques highlighting advantages and or limitations of each. Green techniques discussed in this review include solid phase microextraction, spectroscopic techniques i.e., infra-red (IR) spectroscopy and nuclear magnetic resonance (NMR) besides, e-tongue and e-nose for detection of flavor. The employment of multivariate data analysis in data interpretation is also highlighted in the context of identifying key components pertinent to specific variety, processing method, and or geographical origin.

Chemometric strategies for near infrared hyperspectral imaging analysis: classification of cotton seed genotypes

Hyperspectral images have been increasingly employed in the agricultural sector for seed classification for different purposes. In the present paper we propose a new methodology based on HSI in the near infrared range (HSI-NIR) to distinguish conventional from transgenic cotton seeds. Three different chemometric approaches, one pixel-based and two object-based, using partial least squares discriminant analysis (PLS-DA) were built and their performances were compared considering the pros and cons of each approach. Specificity and sensitivity values ranged from 0.78-0.92 and 0.62-0.93, respectively, for the different approaches.

Deep Learning Model for the Inspection of Coffee Bean Defects

The detection of coffee bean defects is the most crucial step prior to bean roasting. Existing defect detection methods used in the specialty coffee bean industry entail manual screening and sorting, require substantial human resources, and are not standardized. To solve these problems, this study developed a deep learning algorithm to detect defects in coffee beans. The results reveal that when the pooling layer was used to enhance features and reduce neural dimensionality, some of the coffee been features were lost or misclassified. Therefore, a novel dimensionality reduction method was adopted to increase the ability of feature extraction. The developed model also overcame the drawbacks of padding causing blurred image boundaries and the dead neurons causing impeding feature propagation. Images of eight types of coffee beans were used to train and test the proposed detection model. The proposed method was verified to reduce the bias when classifying defects in coffee beans. The detection accuracy rate of the proposed model was 95.2%. When the model was only used to detect the presence of defects, the accuracy rate increased to 100%. Thus, the proposed model is highly accurate in coffee bean defect detection in the classification of eight types of coffee beans.

Ecometabolic mixture design-fingerprints from exploratory multi-block data analysis in Coffea arabica beans from climate changes: Elevated carbon dioxide and reduced soil water availability

Ecometabolic mixture design-fingerprinting in coffee cultivated under climate change was chemically explored using ComDim. Multi-blocks were formed using UV, NIRS, ¹H NMR, SWV, and FT-IR data. ComDim investigated all these different fingerprints according to the extractor solvent and in virtue of atmospheric CO₂ increase. Ethanol and ethanol-dichloromethane showed the best separations due to CO₂ environment. ¹H NMR loading indicate increases of fatty acids, caffeine, trigonelline, and glucose in beans under current CO₂ levels, whereas quinic acid/chlorogenic acids, malic acid, and kahweol/cafestol increased in beans under elevated CO₂ conditions. SWV indicated quercetin and chlorogenic acid as important compounds in coffee beans cultivated under current and elevated CO₂, respectively. Based on the ethanol and ethanol-dichloromethane fingerprints, k-NN correctly classified the beans cultivated under different carbon dioxide environments and water availabilities, confirming the existence of metabolic changes due to climate changes. SWV proved to be promising compared with widely used spectrometric methods.

Rapid and direct detection of small microplastics in aquatic samples by a new near infrared hyperspectral imaging (NIR-HSI) method

Microplastic (MP) contamination is a critical environmental challenge with a strong impact on the ecosystems, economy and potentially for human health. The smaller the MP size, the greater is the environmental risks as well as the analytical difficulties in detecting and characterising the particles. .We propose a rapid near infrared hyperspectral imaging (NIR-HSI) method that enables the chemical identification and characterisation of small MP (down to 80 μm) in aquatic samples, directly on filters, with no pre-sorting step needed. By considerably reducing the procedural steps, the time of analysis and costs our method addresses the urgent need of cost-effective and robust tools for extensive monitoring of MP in natural systems.

An in-depth study of cheese ripening by means of NIR hyperspectral imaging: Spatial mapping of dehydration, proteolysis and lipolysis

Cheese represents one of the most complex food matrices, for the high number of factors contributing to the chemical composition, and so its evaluation represents an important analytical challenge. The present study describes an innovative and non-destructive analytical approach, based on hyperspectral imaging in the near-infrared region (HSI-NIR) and multivariate pattern recognition, to study and monitor the extent - spatial and temporal - of biochemical phenomena responsible for cheese ripening. NIR spectral bands characterising dehydration, proteolysis and lipolysis were individuated and studied by exploiting a representative sample set of characteristic cheeses. The information obtained was employed to develop score maps based on principal component analysis (PCA), which permitted to monitor and visualise the ripening of Formaggetta, a commercial semi-hard cheese typical of Liguria, an Italian region, providing a deep understanding of the evolution of dehydration, proteolysis and lipolysis during the maturation period that precedes the placing on the market.

Monitoring the Processing of Dry Fermented Sausages with a Portable NIRS Device

This work studies the ability of a MicroNIR (VIAVI, Santa Rosa, CA) device to monitor the dry fermented sausage process with the use of multivariate data analysis. Thirty sausages were made and subjected to dry fermentation, which was divided into four main stages. Physicochemical (weight lost, pH, moisture content, water activity, color, hardness, and thiobarbiruric reactive substances analysis) and sensory (quantitative descriptive analysis) characterizations of samples on different steps of the ripening process were performed. Near-infrared (NIR) spectra (950-1650 nm) were taken throughout the process at three points of the samples. Physicochemical data were explored by distance to K-Nearest Neighbor (K-NN) cluster analysis, while NIR spectra were studied by partial least square-discriminant analysis; before these models, Principal Component Analysis (PCA) was performed in both databases. The results of multivariate data analysis showed the ability to monitor and classify the different stages of ripening process (mainly the fermentation and drying steps). This study showed that a portable NIR device (MicroNIR) is a nondestructive, simple, noninvasive, fast, and cost-effective tool with the ability to monitor the dry fermented sausage processing and to classify samples as a function of the stage, constituting a feasible decision method for sausages to progress to the following processing stage.

Detection and identification of Cannabis sativa L. using near infrared hyperspectral imaging and machine learning methods. A feasibility study

Remote identification of illegal plantations of Cannabis sativa Linnaeus is an important task for the Brazilian Federal Police. The current analytical methodology is expensive and strongly dependent on the expertise of the forensic investigator. A faster and cheaper methodology based on automatic methods can be useful for the detection and identification of Cannabis sativa L. in a reliable and objective manner. In this work, the high potential of Near Infrared Hyperspectral Imaging (HSI-NIR) combined with machine learning is demonstrated for supervised detection and classification of Cannabis sativa L. This plant, together with other plants commonly found in the surroundings of illegal plantations and soil, were directly collected from an illegal plantation. Due to the high correlation of the NIR spectra, sparse Principal Component Analysis (sPCA) was implemented to select the most important wavelengths for identifying Cannabis sativa L. One class Soft Independent Class Analogy model (SIMCA) was built, considering just the spectral variables selected by sPCA. Sensitivity and specificity values of 89.45% and 97.60% were, respectively, obtained for an external validation set subjected to the s-SIMCA. The results proved the reliability of a methodology based on NIR hyperspectral cameras to detect and identify Cannabis sativa L., with only four spectral bands, showing the potential of this methodology to be implemented in low-cost airborne devices.

Detection of Insect Damage in Green Coffee Beans Using VIS-NIR Hyperspectral Imaging

The defective beans of coffee are categorized into black beans, fermented beans, moldy beans, insect damaged beans, parchment beans, and broken beans, and insect damaged beans are the most frequently seen type. In the past, coffee beans were manually screened and eye strain would induce misrecognition. This paper used a push-broom visible-near infrared (VIS-NIR) hyperspectral sensor to obtain the images of coffee beans, and further developed a hyperspectral insect damage detection algorithm (HIDDA), which can automatically detect insect damaged beans using only a few bands and one spectral signature. First, by taking advantage of the constrained energy minimization (CEM) developed band selection methods, constrained energy minimization-constrained band dependence minimization (CEM-BDM), minimum variance band prioritization (MinV-BP), maximal variance-based bp (MaxV-BP), sequential forward CTBS (SF-CTBS), sequential backward CTBS (SB-CTBS), and principal component analysis (PCA) were used to select the bands, and then two classifier methods were further proposed. One combined CEM with support vector machine (SVM) for classification, while the other used convolutional neural networks (CNN) and deep learning for classification where six band selection methods were then analyzed. The experiments collected 1139 beans and 20 images, and the results demonstrated that only three bands are really need to achieve 95% of accuracy and 90% of kappa coefficient. These findings show that 850–950 nm is an important wavelength range for accurately identifying insect damaged beans, and HIDDA can indeed detect insect damaged beans with only one spectral signature, which will provide an advantage in the process of practical application and commercialization in the future.

Chemometric challenges in development of paper-based analytical devices: Optimization and image processing

Although microfluidic paper-based analytical devices (μPADs) get a lot of attention in the scientific literature, they rarely reach the level of commercialization. One possible reason for this is a lack of application of machine learning techniques supporting the design, optimization and fabrication of such devices. This work demonstrates the potential of two chemometric techniques including design of experiments (DoE) and digital image processing to support the production of μPADs. On the example of a simple colorimetric assay for isoniazid relying on the protonation equilibrium of methyl orange, the experimental conditions were optimized using a D-optimal design (DO) and the impact of multiple factors on the μPAD response was investigated. In addition, this work demonstrates the impact of automatic image processing on accelerating color value analysis and on minimizing errors caused by manual detection area selection. The employed algorithm is based on morphological recognition and allows the analysis of RGB (red, green, and blue) values in a repeatable way. In our belief, DoE and digital image processing methodologies are keys to overcome some of the remaining weaknesses in μPAD development to facilitate their future market entry.

Macroscopic mid-FTIR mapping and clustering-based automated data-reduction: An advanced diagnostic tool for in situ investigations of artworks

The present study describes a multivariate strategy that can be used for automatic on-site processing of reflection mode macro FTIR mapping (MA-rFTIR) data obtained during investigation of artworks. The chemometric strategy is based on the integration of principal component analysis (PCA) with a clustering approach in the space subtended by the three lowest-order principal components and allows to automatically identify the regions of interest (ROIs) of the area scanned and to extract the average FTIR spectra related to each ROI. Thanks to the automatic data management, in-field HSI (hyperspectral imaging)-based analyses may be performed even by staff lacking specific advanced chemometric expertise, as it is sometimes the case for conservation scientists or conservators with a scientific background. MA-rFTIR was only recently introduced in the conservation field and, in this work the technique was employed to characterize the surface of metallic artefacts. The analytical protocol was employed as part of a rapid procedure to evaluate the conservation state and the performance of cleaning methods on bronze objects. Both activities are commonly part of restoration campaigns of bronzes and require an on-site analytical procedure for efficient and effective diagnosis. The performance of the method was first evaluated on aged standard samples (bronzes with a layer of green basic copper hydroxysulphate, treated with different organic coatings) and then scrutinized in situ on areas of the 16th century Neptune fountain statue (Piazza del Nettuno, Bologna, Italy) by Gianbologna.