Hyperspectral Image Acquisition Research Articles

Quality Detection of Fresh‐Cut Pineapples During Cold Storage Based on Diffuse Reflectance and Fluorescence Hyperspectral Imaging Technique

ABSTRACTThe fresh‐cut fruits and vegetables industry has developed rapidly in recent years, with pineapple being a popular fruit. The quality of pineapple plays a crucial role in determining its market value. This research explored the potential of diffuse reflectance hyperspectral imaging (HSI) and fluorescence hyperspectral imaging (FHSI) in monitoring L*a*b*, pH, and soluble solids content (SSC) of fresh‐cut pineapples during cold storage. Fresh‐cut pineapples were stored at a temperature of 4°C for 0–5 days, with daily acquisition of diffuse reflectance and fluorescence hyperspectral images at 380–1000 nm, along with measurement of quality indices. Changes in spectral properties and quality indices during cold storage were analyzed. Seven preprocessing algorithms were used to preprocess two types of spectra. Subsequently, five feature selection methods were employed to extract feature variables from the preprocessed spectra. The Partial Least Squares Regression (PLSR) models were constructed to predict the various quality indices. Data fusion methods were introduced to leverage the complementary information from different spectra. Both feature‐level and decision‐level fusion methods demonstrated improvement in model accuracy. The hybrid fusion method, combining the advantages of these two fusion methods, effectively enhanced the prediction accuracy of all quality indices. The determination coefficients (R2) for predicting L*a*b* consistently exceeded 0.8, while the R2 for predicting pH was close to 0.8, and the R2 for predicting SSC reached 0.91. In summary, the quality indices of fresh‐cut pineapples can be accurately predicted using diffuse reflectance HSI and FHSI, with model accuracy further enhanced through multi‐level data fusion methods.

Advancing hyperspectral imaging techniques for root systems: a new pipeline for macro- and microscale image acquisition and classification

BackgroundUnderstanding the environmental impacts on root growth and root health is essential for effective agricultural and environmental management. Hyperspectral imaging (HSI) technology provides a non-destructive method for detailed analysis and monitoring of plant tissues and organ development, but unfortunately examples for its application to root systems and the root-soil interface are very scarce. There is also a notable lack of standardized guidelines for image acquisition and data analysis pipelines.MethodsThis study investigated HSI techniques for analyzing rhizobox-grown root systems across various imaging configurations, from the macro- to micro-scale, using the imec VNIR SNAPSCAN camera. Focusing on three graminoid species with different root architectures allowed us to evaluate the influence of key image acquisition parameters and data processing techniques on the differentiation of root, soil, and root-soil interface/rhizosheath spectral signatures. We compared two image classification methods, Spectral Angle Mapper (SAM) and K-Means clustering, and two machine learning approaches, Random Forest (RF) and Support Vector Machine (SVM), to assess their efficiency in automating root system image classification.ResultsOur study demonstrated that training a RF model using SAM classifications, coupled with wavelength reduction using the second derivative spectra with Savitzky-Golay (SG) smoothing, provided reliable classification between root, soil, and the root-soil interface, achieving 88–91% accuracy across all configurations and scales. Although the root-soil interface was not clearly resolved, it helped to improve the distinction between root and soil classes. This approach effectively highlighted spectral differences resulting from the different configurations, image acquisition settings, and among the three species. Utilizing this classification method can facilitate the monitoring of root biomass and future work investigating root adaptations to harsh environmental conditions.ConclusionsOur study addressed the key challenges in HSI acquisition and data processing for root system analysis and lays the groundwork for further exploration of VNIR HSI application across various scales of root system studies. This work provides a full data analysis pipeline that can be utilized as an online Python-based tool for the semi-automated analysis of root-soil HSI data.

Hyperspectral imaging of human liver allografts for prediction of initial graft function

PurposeIschemia reperfusion injury represents a significant yet difficult to assess risk factor for short- and long-term graft impairment in human liver transplantation (LT). As a non-invasive, non-ionizing tool, hyperspectral imaging (HSI) is capable of correlating optical properties with organ microperfusion. Hence, we here performed a study of human liver allografts assessed by HSI for microperfusion and prediction of initial graft function.MethodsImages of liver parenchyma of 37 human liver allografts were acquired at bench preparation, during normothermic machine perfusion (NMP), if applicable, and after reperfusion in the recipient. A specialized HSI acquisition software computed oxygen saturation (StO2), tissue hemoglobin indices (THI), near infrared perfusion indices (NIR), and tissue water indices (TWI). HSI parameters were analyzed for differences with regard to preservation technique, reperfusion sequence and presence of early allograft dysfunction (EAD).ResultsOrgan preservation was performed by means of NMP (n = 31) or static cold storage (SCS; n = 6). Patients’ demographics, donor characteristics, presence of EAD (NMP 36.7% vs. SCS 50%, p = 0.6582), and HSI parameters were comparable between both groups of preservation method. In organs developing EAD, NIR at 1, 2, and 4 h NMP and after reperfusion in the recipient was significantly lower (1 h NMP: 18.6 [8.6–27.6] vs. 28.3 [22.5–39.4], p = 0.0468; 2 h NMP: 19.4 [8.7–30.4] vs. 37.1 [27.5–44.6], p = 0.0011; 4 h NMP: 26.0 [6.8–37.1] vs. 40.3 [32.3–49.9], p = 0.0080; reperfusion: 13.0 [11.5–34.3] vs. 30.6 [19.3–44.0], p = 0.0212).ConclusionHSI assessment of human liver allografts is feasible during organ preservation and in the recipient. NIR during NMP and after reperfusion might predict the onset of EAD. Larger trials are warranted for assessment of this novel technique in human LT.

Efficient microplastic identification by hyperspectral imaging: A comparative study of spatial resolutions, spectral ranges and classification models to define an optimal analytical protocol

Microplastics (MPs) pollution is a global and challenging issue, necessitating the development of efficient analytical strategies for their detection to monitor their environmental impact.This study aims to define an optimal analytical protocol for characterizing MPs by hyperspectral imaging (HSI), comparing different setups based on spatial resolution, spectral range and classification models. The investigated MPs include polymers commonly found in the environment, such as polystyrene (PS), polypropylene (PP) and high-density polyethylene (HDPE), subdivided in three size classes (1000–2000 μm, 500–1000 μm, 250–500 μm). Furthermore, MP particles with diameters ranging from 30 to 250 μm were assessed to determine the limit of detection (LOD) in the different configurations. Hyperspectral images were acquired with two spatial resolutions, 150 and 30 μm/pixel, and two spectral ranges, 1000–1700 nm (NIR) and 1000–2500 nm (SWIR). Three classification models, Partial Least Square-Discriminant Analysis (PLS-DA), Error Correction Output Coding-Support Vector Machine (ECOC-SVM) and Neural Network Pattern Recognition (NNPR) were tested on the acquired images. The correctness of these models was evaluated by prediction maps and statistical parameters (Recall, Specificity and Accuracy). The results demonstrated that for MP particles larger than 250 μm, the optimal setup is a spatial resolution of 150 μm/pixel and a spectral range of 1000–1700 nm, utilizing a linear classification model like PLS-DA. This approach offers accurate predictions while being time- and cost-efficient. For MPs smaller than 250 μm, a higher spatial resolution of 30 μm/pixel with a spectral range of 1000–2500 nm and a non-linear classification method like ECOC-SVM is preferable. The LOD is 250 μm for the 150 μm/pixel resolution and ranges from 100 to 200 μm for the 30 μm/pixel resolution. These findings provide a valuable guide for selecting the appropriate HSI acquisition conditions and data processing methods to optimally characterize MPs of different sizes.

EELS hyperspectral images unmixing - using autoencoders

Spatially resolved Electron Energy-Loss Spectroscopy conducted in a Scanning Transmission Electron Microscope enables the acquisition of hyperspectral images. Spectral unmixing is the process of decomposing each spectrum of an hyperspectral image into a combination of representative spectra (endmembers) corresponding to compounds present in the sample along with their local proportions (abundances). Spectral unmixing is a complex task, and various methods have been developed in different communities using hyperspectral images. However, none of these methods fully satisfy the spatially resolved Electron Energy-Loss Spectroscopy requirements. Recent advancements in remote sensing, which focus on Deep Learning techniques, have the potential to meet these requirements, particularly Autoencoders. In this study, the performance of Deep Learning methods using autoencoders for spectral unmixing is evaluated, and their results are compared with traditional methods. Synthetic hyperspectral images have been created to quantitatively assess the outcomes of the unmixing process using specific metrics. The methods are subsequently applied to a series of experimental data. The findings demonstrate the promising potential of autoencoders as a tool for Electron Energy-Loss Spectroscopy hyperspectral images unmixing, marking a starting point for exploring more sophisticated Neural Networks.

UAV hyperspectral image acquisition and processing, an application for nutrient estimation of rice in Vietnam

Hyperspectral imagery obtained from Unmanned Aerial Vehicles (UAVs) is increasingly employed to investigate nutrient concentrations in vegetation. The deployment of a hyperspectral camera on a UAV, flight planning, image acquisition, preprocessing of hyperspectral data, and the subsequent estimation of nutrient concentrations in vegetation are facing challenges. These challenges manifest as geometric, spectral distortions, and the abundance of numerous spectral bands. This study seeks to guide on mitigating the impact of issues encountered during an experiment to estimate nutrient concentrations in rice leaves using UAV hyperspectral images. An industrial hexagonal drone equipped with a push-broom hyperspectral camera featuring 122 bands within the Visible to Near-Infrared (VIS-NIR) wavelength range (400-960 nm) is employed to collect data over a 1-hectare testing rice field. Models for estimating Leaf Phosphorus Concentration (LPC) and Leaf Potassium Concentration (LKC) are developed based on the correlation between hyperspectral images, characterized by a 3 cm spatial resolution, and 162 LPC and 162 LKC reference data points. Utilizing various vegetation indices for LPC and LKC estimation, the outcomes reveal that a combination of band wavelengths at 838 nm and 734 nm is effective for LPC estimation, yielding a Root Mean Square Error (RMSE) of 27.1%. Conversely, LKC estimation exhibits an RMSE of 38.8% with an insignificant correlation between LKC and the current wavelength ranges. Above all, this study is a primary example of the utilization of UAV hyperspectral data in precision agriculture in Vietnam.

Nondestructive freshness evaluation of mackerel fish using Vis/NIR hyperspectral imaging and multivariate analysis

Nondestructive freshness evaluation models for chub mackerel (Scomber japonicus) fillets were developed using visible/near-infrared (Vis/NIR) hyperspectral imaging and multivariate regression analysis. Total 96 mackerel samples were investigated during 6 days of storage under five different conditions for measurement of pH, total volatile basic nitrogen (TVB-N), and K values along with acquisition of hyperspectral images. With partial least squares regression (PLSR) and support vector regression (SVR) along with wavelength selection method using Variables Importance in Projection (VIP) scores, performances of PLSR, VIP-PLSR, SVR, and VIP-SVR models were evaluated and compared. The VIP-PLSR models showed the best performance for predicting the freshness indicators, with R2 values of 0.86, 0.86, and 0.91 for pH, TVB-N, and K values, respectively. Furthermore, it was shown that the identification and removal of noise pixels from the hyperspectral data based on correlation analysis was effective in improving the regression results.

Hyperspectral imaging combined with GA-SVM for maize variety identification.

The demand for identification of maize varieties has increased dramatically due to the phenomenon of mixed seeds and inferior varieties pretending to be high-quality varieties continuing to occur. It is urgent to solve the problem of efficient and accurate identification of maize varieties. A hyperspectral image acquisition system was used to acquire images of maize seeds. Regions of interest (ROI) with an embryo size of 10 × 10 pixel were extracted, and the average spectral information in the range of 949.43-1709.49 nm was intercepted for the subsequent study in order to eliminate random noise at both ends. Savitzky-Golay (SG) smoothing algorithm and multiple scattering correction (MSC) were used to pretreat the full-band spectrum. The feature wavelengths were screened by successive projection algorithms (SPA), competitive adaptive reweighted sampling (CARS) single screening, and two combinations of CARS-SPA and CARS + SPA, respectively. Support vector machines (SVMs) and models optimized based on genetic algorithm (GA), particle swarm optimization (PSO) were established by using full bands (FB) and feature bands as the model input. The results showed that the MSC-(CARS-SPA)-GA-SVM model had the best performance with 93.00% of the test set accuracy, 8 feature variables, and a running time of 24.45 s. MSC pretreatment can effectively eliminate the scattering effect of spectral data, and the feature wavelengths extracted by CARS-SPA can represent all wavelength information. The study proved that hyperspectral imaging combined with GA-SVM can realize the identification of maize varieties, which provided a theoretical basis for maize variety classification and authenticity identification.

A hyperspectral surgical microscope with super-resolution reconstruction for intraoperative image guidance.

Hyperspectral imaging (HSI) is an emerging imaging modality in medical applications, especially for intraoperative image guidance. A surgical microscope improves surgeons' visualization with fine details during surgery. The combination of HSI and surgical microscope can provide a powerful tool for surgical guidance. However, to acquire high-resolution hyperspectral images, the long integration time and large image file size can be a burden for intraoperative applications. Super-resolution reconstruction allows acquisition of low-resolution hyperspectral images and generates high-resolution HSI. In this work, we developed a hyperspectral surgical microscope and employed our unsupervised super-resolution neural network, which generated high-resolution hyperspectral images with fine textures and spectral characteristics of tissues. The proposed method can reduce the acquisition time and save storage space taken up by hyperspectral images without compromising image quality, which will facilitate the adaptation of hyperspectral imaging technology in intraoperative image guidance.

Determination of Dry-Matter Content of Kiwifruit before Harvest Based on Hyperspectral Imaging

Determining pre-harvest fruit maturity is vital to ensure the quality of kiwifruit, and dry-matter content is an important indicator of kiwifruit ripeness. To predict the pre-harvest dry-matter content of kiwifruit continuously in real-time with high accuracy, this study uses hyperspectral data of pre-harvest Jintao kiwifruit obtained by using a hyperspectral image acquisition device. The raw data underwent whiteboard correction, spectral data extraction, spectral pre-processing, and feature-band extraction, following which the dry-matter content of the fruit was predicted by using partial least squares (PLS) regression. The feature bands extracted by the random frog method were 538.93, 671.14, 693.41, 770.61, 796.98, 813.24, 841.21, 843.29, and 856.80 nm, which improve the accuracy of the PLS method for predicting dry-matter content, with R2 = 0.92 and a root mean square error (RMSE) of 0.41% for the training set, and R2 = 0.85 and a RMSE of 0.50% for the test set. These results show that the proposed method reduces the number of required bands while maintaining the prediction accuracy, thereby demonstrating the reliability of using hyperspectral data to predict the pre-harvest dry-matter content of kiwifruit. This method can effectively guide the management of kiwifruit harvesting period, establishing a theoretical foundation for precise unmanned harvesting.

Integrating spectral and image information for prediction of cottonseed vitality.

Cotton plays a significant role in people's lives, and cottonseeds serve as a vital assurance for successful cotton cultivation and production. Premium-quality cottonseeds can significantly enhance the germination rate of cottonseeds, resulting in increased cotton yields. The vitality of cottonseeds is a crucial metric that reflects the quality of the seeds. However, currently, the industry lacks a non-destructive method to directly assess cottonseed vitality without compromising the integrity of the seeds. To address this challenge, this study employed a hyperspectral imaging acquisition system to gather hyperspectral data on cottonseeds. This system enables the simultaneous collection of hyperspectral data from 25 cottonseeds. This study extracted spectral and image information from the hyperspectral data of cottonseeds to predict their vitality. SG, SNV, and MSC methods were utilized to preprocess the spectral data of cottonseeds. Following this preprocessing step, feature wavelength points of the cottonseeds were extracted using SPA and CARS algorithms. Subsequently, GLCM was employed to extract texture features from images corresponding to these feature wavelength points, including attributes such as Contrast, Correlation, Energy, and Entropy. Finally, the vitality of cottonseeds was predicted using PLSR, SVR, and a self-built 1D-CNN model. For spectral data analysis, the 1D-CNN model constructed after MSC+CARS preprocessing demonstrated the highest performance, achieving a test set correlation coefficient of 0.9214 and an RMSE of 0.7017. For image data analysis, the 1D-CNN model constructed after SG+CARS preprocessing outperformed the others, yielding a test set correlation coefficient of 0.8032 and an RMSE of 0.9683. In the case of fused spectral and image data, the 1D-CNN model built after SG+SPA preprocessing displayed the best performance, attaining a test set correlation coefficient of 0.9427 and an RMSE of 0.6872. These findings highlight the effectiveness of the 1D-CNN model and the fusion of spectral and image features for cottonseed vitality prediction. This research contributes significantly to the development of automated detection devices for assessing cottonseed vitality.

Non-destructive detection of protein content in mulberry leaves by using hyperspectral imaging.

Protein content is one of the most important indicators for assessing the quality of mulberry leaves. This work is carried out for the rapid and non-destructive detection of protein content of mulberry leaves using hyperspectral imaging (HSI) (Specim FX10 and FX17, Spectral Imaging Ltd., Oulu, Finland). The spectral range of the HSI acquisition system and data processing methods (pretreatment, feature extraction, and modeling) is compared. Hyperspectral images of three spectral ranges in 400-1,000 nm (Spectral Range I), 900-1,700 nm (Spectral Range II), and 400-1,700 nm (Spectral Range III) were considered. With standard normal variate (SNV), Savitzky-Golay first-order derivation, and multiplicative scatter correction used to preprocess the spectral data, and successive projections algorithm (SPA), competitive adaptive reweighted sampling, and random frog used to extract the characteristic wavelengths, regression models are constructed by using partial least square and least squares-support vector machine (LS-SVM). The protein content distribution of mulberry leaves is visualized based on the best model. The results show that the best results are obtained with the application of the model constructed by combining SNV with SPA and LS-SVM, showing an R 2 of up to 0.93, an RMSE of just 0.71 g/100 g, and an RPD of up to 3.83 based on the HSI acquisition system of 900-1700 nm. The protein content distribution map of mulberry leaves shows that the protein of healthy mulberry leaves distributes evenly among the mesophyll, with less protein content in the vein of the leaves. The above results show that rapid, non-destructive, and high-precision detection of protein content of mulberry leaves can be achieved by applying the SWIR HSI acquisition system combined with the SNV-SPA-LS-SVM algorithm.

Design and Implementation of Airborne Hyperspectral Camera Control and Data Storage System

With the great progress of hyperspectral imaging technology, airborne hyperspectral cameras have been widely used in the field of earth detection and remote sensing. However, with the increasing resolution of airborne hyperspectral cameras, the amount of data acquired is also growing explosive, The speed and accuracy in the transmission process are easily affected, which poses challenges to the real-time control and data acquisition and storage of airborne hyperspectral cameras. In this paper, a general airborne hyperspectral image acquisition and storage system is designed, which converts the hyperspectral data from the Cameralink interface into a high-speed serial PCIe data stream and realizes high-speed storage of image data. The image width is 2048 pixels, the number of spectral segments is 120, the pixel size is 12bit, and the maximum frame rate is 300 fps. The maximum transmission rate is 2048×120×12bit×300fps, and the data rate is 844Mbps. Based on this, we perform thermal analysis and vibration analysis of the system platform. Experimental results show that the system can achieve stable data transmission and storage, and the data is intact and reliable.

Lightfield hyperspectral imaging in neuro-oncology surgery: an IDEAL 0 and 1 study.

Hyperspectral imaging (HSI) has shown promise in the field of intra-operative imaging and tissue differentiation as it carries the capability to provide real-time information invisible to the naked eye whilst remaining label free. Previous iterations of intra-operative HSI systems have shown limitations, either due to carrying a large footprint limiting ease of use within the confines of a neurosurgical theater environment, having a slow image acquisition time, or by compromising spatial/spectral resolution in favor of improvements to the surgical workflow. Lightfield hyperspectral imaging is a novel technique that has the potential to facilitate video rate image acquisition whilst maintaining a high spectral resolution. Our pre-clinical and first-in-human studies (IDEAL 0 and 1, respectively) demonstrate the necessary steps leading to the first in-vivo use of a real-time lightfield hyperspectral system in neuro-oncology surgery. A lightfield hyperspectral camera (Cubert Ultris ×50) was integrated in a bespoke imaging system setup so that it could be safely adopted into the open neurosurgical workflow whilst maintaining sterility. Our system allowed the surgeon to capture in-vivo hyperspectral data (155 bands, 350-1,000 nm) at 1.5 Hz. Following successful implementation in a pre-clinical setup (IDEAL 0), our system was evaluated during brain tumor surgery in a single patient to remove a posterior fossa meningioma (IDEAL 1). Feedback from the theater team was analyzed and incorporated in a follow-up design aimed at implementing an IDEAL 2a study. Focusing on our IDEAL 1 study results, hyperspectral information was acquired from the cerebellum and associated meningioma with minimal disruption to the neurosurgical workflow. To the best of our knowledge, this is the first demonstration of HSI acquisition with 100+ spectral bands at a frame rate over 1Hz in surgery. This work demonstrated that a lightfield hyperspectral imaging system not only meets the design criteria and specifications outlined in an IDEAL-0 (pre-clinical) study, but also that it can translate into clinical practice as illustrated by a successful first in human study (IDEAL 1). This opens doors for further development and optimisation, given the increasing evidence that hyperspectral imaging can provide live, wide-field, and label-free intra-operative imaging and tissue differentiation.

Compressive hyperspectral microscopy for cancer detection.

Hyperspectral microscopy grants the ability to characterize unique properties of tissues based on their spectral fingerprint. The ability to label and measure multiple molecular probes simultaneously provides pathologists and oncologists with a powerful tool to enhance accurate diagnostic and prognostic decisions. As the pathological workload grows, having an objective tool that provides companion diagnostics is of immense importance. Therefore, fast whole-slide spectral imaging systems are of immense importance for automated cancer prognostics that meet current and future needs. We aim to develop a fast and accurate hyperspectral microscopy system that can be easily integrated with existing microscopes and provide flexibility for optimizing measurement time versus spectral resolution. The method employs compressive sensing (CS) and a spectrally encoded illumination device integrated into the illumination path of a standard microscope. The spectral encoding is obtained using a compact liquid crystal cell that is operated in a fast mode. It provides time-efficient measurements of the spectral information, is modular and versatile, and can also be used for other applications that require rapid acquisition of hyperspectral images. We demonstrated the acquisition of breast cancer biopsies hyperspectral data of the whole camera area within . This means that a typical biopsy can be measured in . The hyperspectral images with 250 spectral bands are reconstructed from 47 spectrally encoded images in the spectral range of 450 to 700nm. CS hyperspectral microscopy was successfully demonstrated on a common lab microscope for measuring biopsies stained with the most common stains, such as hematoxylin and eosin. The high spectral resolution demonstrated here in a rather short time indicates the ability to use it further for coping with the highly demanding needs of pathological diagnostics, both for cancer diagnostics and prognostics.

Hyperspectral imaging combined with CNN for maize variety identification.

As the third largest food crop in the world, maize has wide varieties with similar appearances, which makes identification difficult. To solve the problem of identification of hybrid maize varieties, a method based on hyperspectral image technology combined with a convolutional neural network (CNN) is proposed to identify maize varieties. In this study, 735 maize seeds from seven half-parent hybrid maize varieties were regarded as the research object. The maize seed images in the range of 900 ~ 1700nm were obtained by hyperspectral image acquisition system. The region of interest (ROI) of the embryo surface was selected, and the spectral reflectance of maize seeds was extracted. After Savitzky-Golay (SG) Smoothing pretreatment, Maximum Normalization (MN) pretreatment was performed. The 56 feature wavelengths were selected by Competitive Adaptive Reweighting Algorithm (CARS) and Successive Projection Algorithm (SPA). And the 56 wavelengths were mapped to high-dimensional space by high-dimensional feature mapping and then reconstructed into three-dimensional image features. A five-layer convolution neural network was used to identify three-dimensional image features, and nine (SG+MN)-(CARS+SPA)-CNN maize variety identification models were established by changing the input feature dimension and the depth factor size of the model layer. The results show that the maize variety classification model works best, when the input feature dimension is 768 and the layer depth factor d is 1.0. At this point, the model accuracy of the test set is 96.65% and the detection frame rate is1000 Fps/s in GPU environment, which can realize the rapid and effective non-destructive detection of maize varieties. This study provides a new idea for the rapid and accurate identification of maize seeds and seeds of other crops.

Identification of Turtle-Shell Growth Year Using Hyperspectral Imaging Combined with an Enhanced Spatial-Spectral Attention 3DCNN and a Transformer.

Turtle shell (Chinemys reecesii) is a prized traditional Chinese dietary therapy, and the growth year of turtle shell has a significant impact on its quality attributes. In this study, a hyperspectral imaging (HSI) technique combined with a proposed deep learning (DL) network algorithm was investigated for the objective determination of the growth year of turtle shells. The acquisition of hyperspectral images was carried out in the near-infrared range (948.72-2512.97 nm) from samples spanning five different growth years. To fully exploit the spatial and spectral information while reducing redundancy in hyperspectral data simultaneously, three modules were developed. First, the spectral-spatial attention (SSA) module was developed to better protect the spectral correlation among spectral bands and capture fine-grained spatial information of hyperspectral images. Second, the 3D convolutional neural network (CNN), more suitable for the extracted 3D feature map, was employed to facilitate the joint spatial-spectral feature representation. Thirdly, to overcome the constraints of convolution kernels as well as better capture long-range correlation between spectral bands, the transformer encoder (TE) module was further designed. These modules were harmoniously orchestrated, driven by the need to effectively leverage both spatial and spectral information within hyperspectral data. They collectively enhance the model's capacity to extract joint spatial and spectral features to discern growth years accurately. Experimental studies demonstrated that the proposed model (named SSA-3DTE) achieved superior classification accuracy, with 98.94% on average for five-category classification, outperforming traditional machine learning methods using only spectral information and representative deep learning methods. Also, ablation experiments confirmed the effectiveness of each module to improve performance. The encouraging results of this study revealed the potentiality of HSI combined with the DL algorithm as an efficient and non-destructive method for the quality control of turtle shells.

Towards automated detection of hyperhydricity in plant in vitro culture

Hyperhydricity (HH) is one of the most important physiological disorders that negatively affects various plant tissue culture techniques. The objective of this study was to characterize optical features to allow an automated detection of HH. For this purpose, HH was induced in two plant species, apple and Arabidopsis thaliana, and the severity was quantified based on visual scoring and determination of apoplastic liquid volume. The comparison between the HH score and the apoplastic liquid volume revealed a significant correlation, but different response dynamics. Corresponding leaf reflectance spectra were collected and different approaches of spectral analyses were evaluated for their ability to identify HH-specific wavelengths. Statistical analysis of raw spectra showed significantly lower reflection of hyperhydric leaves in the VIS, NIR and SWIR region. Application of the continuum removal hull method to raw spectra identified HH-specific absorption features over time and major absorption peaks at 980 nm, 1150 nm, 1400 nm, 1520 nm, 1780 nm and 1930 nm for the various conducted experiments. Machine learning (ML) model spot checking specified the support vector machine to be most suited for classification of hyperhydric explants, with a test accuracy of 85% outperforming traditional classification via vegetation index with 63% test accuracy and the other ML models tested. Investigations on the predictor importance revealed 1950 nm, 1445 nm in SWIR region and 415 nm in the VIS region to be most important for classification. The validity of the developed spectral classifier was tested on an available hyperspectral image acquisition in the SWIR-region.

Full-surface defect detection of navel orange based on hyperspectral online sorting technology.

The whole-surface hyperspectral image acquisition of navel orange is particularly important for surface defect detection and quality classification. Because the light intensity at the edge of the navel orange is lower than that in the middle, the defects on the surface of the navel orange cannot be effectively identified. In this paper, a hyperspectral online sorting device for the whole-surface defects of navel orange is proposed. First of all, the image data of navel orange is collected by online detection sorting equipment and the spectral image of the characteristic wave peak of 1655.72nm was extracted. Then, the light intensity at the edge of the navel orange is enhanced by nonuniformity correction based on quadratic curve fitting, and the light intensity correction of the navel orange is realized. Finally, the corrected image is segmented by the threshold to obtain surface defects, and the number of surface defect pixels is improved effectively compared with that before light intensity correction. Ultimately, the online sorting test is carried out, and the detection accuracy is 100%. This indicates that this method effectively improves the sensitivity of defect detection. At the same time, the dimensionality reduction of hyperspectral data is also carried out, which is conducive to improving the efficiency of online detection.

An attention-driven nonlinear optimization method for CS-based hyperspectral image reconstruction

Hyperspectral images contain both spatial and spectral information, which makes it has great advantages in the field of target identification, material determination, etc. However, hyperspectral image acquisition systems usually output massive amounts of data. Compressed sensing (CS) and single dispatchers coded aperture snapshot spectral imaging (SD-CASSI) were proposed to compress the amount of data. These systems usually use numerical optimization methods to do image reconstruction by finding the optimal reconstructed image. This method generates a large number of iterative calculations, which is time-consuming. In this paper, we modeled the image reconstruction for SC-CASSI as a non-linear optimization problem and proved that the number of iterations of the solution process is exponential. We proposed an attention-driven reconstruction algorithm for SD-CASSI systems. The algorithm mainly consists of an initial convolution block and residual attention modules. A channel attention module is also introduced to mine spectrally correlated features in raw signals. Experimental results show that the proposed algorithm is superior to three states-of-art reconstruction algorithms for CS-based systems. The reconstruction speed of the system is also increased by 94% compared with the numerical method.