Sensitive Wavebands Research Articles

Identificating vegetation stress under natural gas micro-leakage based on leaf scale temporal hyperspectrum

ABSTRACT Timely detection of underground natural gas leakage is significant for the safe production of gas storage and the protection of ecological environment. Direct detection makes it difficult to identify natural gas micro-leakage accurately, while hyperspectral remote sensing can identify micro-leakage based on spectral changes of vegetation near the leakage area. However, the spectral morphological characteristics of vegetation leaves under natural gas micro-leakage will undergo complex changes over time, making it difficult to identify vegetation stress. Temporal spectral characteristics improve the capability of capturing spectral differences of stressed vegetation at different times, allowing for indirect evaluation of leaked information. This study simulates the stress of underground natural gas storage micro leaks on bean growth through field experiments. Based on the changes in spectral reflectance and chlorophyll content of leaves, temporal features were constructed based on multi-temporal reflectance data, and feature selection methods were used to obtain bands sensitive to natural gas stress. This paper proposed a leaf scale natural gas micro-leakage identification index based on temporal spectral features (TNII) to identify bean stress under natural gas leakage. The results are summarized as follows: (1) With the leakage of underground natural gas, the spectral reflectance of leaves increased in the visible and near-infrared range, and the chlorophyll content decreased. (2) The TNII ( R 676 − R 526 ) Tn − ( R 676 − R 526 ) T 1 ( R 676 − R 526 ) Tn + ( R 676 − R 526 ) T 1 constructed based on sensitive wavebands (676 and 526 nm) can be employed to identify bean stress under natural gas after 17 days of leakage. The JM (Jefferies Matusita) distance test indicates that TNII could timely and stably identify bean stress under natural gas micro-leakage (JM > 1.8). This study provides the theoretical basis and technology support for micro-leakage detection of underground gas storage or pipelines.

Spectral library of crops and discrimination of major vegetables grown in the eastern Himalayan ecosystem: A proximal hyperspectral remote sensing approach

Identifying, characterising and mapping, unique vegetable crops grown on small mixed lands in the eastern Himalayan mountain ecosystem (EHME) using traditional methods is a challenge. The exploration of high-resolution, multispectral optical satellite data for vegetable mapping is hindered by heavy rainfall, clouds, and year-round fog/snow cover. Hyperspectral imaging is restricted in the region that borders several international countries because of security concerns. In the global literature on spectroscopy, information on vegetable crops cultivated in the Himalayan ecosystem is marginal or never covered. Thus, we used a portable Field Spec Handheld2 (ASD) spectroradiometer for spectral characterization of 28 important vegetables, from 10 different families. During the summer and winter seasons of 2020–2022, they were grown without stress and in good health in the experimental field at Umiam, Meghalaya. The ground radiometer was able to achieve a spectral resolution of <3 nm in the 325 to 1075 nm wavelength range. At critical growth stages, data were acquired on cloudless days in 1100 and 1300 nm, with a height of 1 m above the crop and a field of view (FOV) <25°. Crop-specific sensitive wavebands were identified and estimated 12 vegetation indices (VI) that relate to biophysical and biochemical properties for crop discrimination. The near infrared region (NIR: 700–1070 nm) was effective for crop discrimination. Of the 10 families, the mean NDVI values varied in the following descending order: Umbelliferae(0.878) > Malvaceae(0.863) > Cucurbitaceae(0.802) > Lamiaceae(0.732) > Solanaceae(0.721) > Leguminosae(0.651) > Cruciferae (0.607) > Amarantha ceae(0.604) > Zingiberaceae(0.595) > Liliaceae(0.531). Among vegetables, the highest NDVI was reported for Fenugreek (0.922) and the lowest for Broccoli (0.421). The position on the red edge (700-750 nm) varied greatly. Similarly, crop families had varying VI values for chlorophyll, carotenoid, xanthophyll and anthocyanin pigments. The ranges of these indices were defined as healthy or stressed crops for the region. From the multi-season spectral signature, a spectral crop library (CSL) for all vegetables was developed under clearly defined conditions (e.g. growth stage, time and year, soil, cultivar type, and climate). The CSL generated from the study provides a reference spectral source for spatial mapping and quantitative spectral modelling to assess crop conditions and prescribe measures. This will be a crucial step in promoting precision agriculture in the eastern Himalayas and other similar ecosystems.

Applicability study of four atmospheric correction methods in the remote sensing of lake water color

Due to the complexity of spectral characteristics for inland lake water, high-precision atmospheric correction methods play an essential role in remote sensing. This research conducted atmospheric correction on the Tiande lake zone in Zhengzhou, China based on Zhuhai-1, Sentinel-2, Planet, and GF-2 images using four methods (QUAC, Dark Subtract, Fast Line-of-sight Atmospheric Analysis of Spectral Hypercubes (FLAASH) and 6S). In doing so, water reflectance can be extracted and compared with the data measured from a ground-based spectrometer. The analysis indicates that the FLAASH method has best applicability to Zhuhai-1 and Sentinel-2 images; while 6S method shows best applicability to Planet and GF-2 images. The comparison of single-band images indicates that the four atmospheric correction approaches used in different images are proven to be applicable to construction of sensitive wavebands of a water body inversion model in the Tiande lake zone. The correction results of four types of images prove that FLAASH and 6S are more applicable to lake water color remote sensing inversion.

An improved hyperspectral sensing approach for the rapid determination of copper ion concentrations in water environment using short-wavelength infrared spectroscopy

Copper ion is one of the hazardous pollutants often present in industrial wastewater or acid mine drainage that is regarded as a primary environmental challenge. Hyperspectral remote sensing has a long tradition in water quality monitoring. However, its application in heavy metal detection is relatively similar, and the detection is highly influenced by water turbidity or total suspended matter (TSM), requiring research efforts to improve accuracy and generalize the applicability of this technique. In this study, the use of simple filtration (pore size of 0.7 μm) for sample pretreatment to improve hyperspectral remote sensing of copper ion concentrations (Cu, 100–1000 mg/L) in water samples is proposed. A wide variety of water samples, including as-prepared and field (fish pond and river water) samples, were investigated to validate the developed method. Spectral data containing sensitive bands characterized in the range of 900–1100 nm were first preprocessed with logarithm transformation, followed by quantitative prediction model development using stepwise multivariate linear regression (SMLR) with the most sensitive wavebands at around 900 nm and 1080 nm. Satisfactory prediction performance for Cu ions was found for turbid water samples (TSM greater than approximately 200 mg/L) after simple filtration pretreatment, suggesting that pretreatment removed suspended solids in the mixtures and enhanced the spectral features of Cu ions in the model. Moreover, good agreement between the laboratory results and the field samples (adjusted R2 > 0.95 and NRMSE <0.15) highlights the suitability of the developed model and filtration pretreatment for obtaining relevant information for the rapid determination of Cu ion concentrations in complex water samples.

Detection of Defective Features in Cerasus Humilis Fruit Based on Hyperspectral Imaging Technology

Detection of skin defects in Cerasus humilis fruit is a critical process to guarantee its quality and price. This study presents a valid method for the detection of defective features in Cerasus humilis fruits based on hyperspectral imaging. A total of 420 sample images were acquired that included three types of natural defects and undamaged samples. After acquiring hyperspectral images of Cerasus humilis fruits, the spectral data were extracted from the region of interest (ROI). Five spectral preprocessing methods were used to preprocess the original spectral data, including Savitsky–Golay (S-G), standard normal variate (SNV), multiplicative scatter correction (MSC), baseline correction (BC), and de-trending (De-T). Regression coefficient (RC), successive projections algorithm (SPA), and competitive adaptive reweighed sampling (CARS) were conducted to select optimal sensitive wavelengths (SWs); as a result, 11 SWs, 17 SWs, and 13 SWs were selected, respectively. Then, the least squares-support vector machine (LS-SVM) discrimination model was established using the selected SWs. The results showed that the discriminate accuracy of the CARS-LS-SVM method was 91.43%. Based on the characteristics of image information, images corresponding to eight sensitive wavebands (950, 994, 1071, 1263, 1336, 1457, 1542, and 1628 nm) selected by CARS were subjected to principal component analysis (PCA). Then, an effective approach for detecting the defective features was exploited based on the imfill function, canny operator, region growing algorithm, bwareaopen function, and the images of PCA. The location and area of defect feature of 105 Cerasus humilis fruits could be recognized; the detect precision was 88.57%. This investigation demonstrated that hyperspectral imaging combined with an image processing technique could achieve the rapid identification of undamaged samples and natural defects in Cerasus humilis fruit. This provides a theoretical basis for the development of Cerasus humilis fruit grading and sorting equipment.

Monitoring Cadmium Content in the Leaves of Field Pepper and Eggplant in a Karst Area Using Hyperspectral Remote Sensing Data

The ability to quickly and non-destructively monitor the cadmium (Cd) content in agricultural crops is the basic premise of effective prevention and control of Cd contamination in agricultural products. Hyperspectral technology provides a solution for this issue. The potential capability for the spectral prediction of the Cd content in the leaves of pepper and eggplant in the field was explored, and a spectral prediction model of the Cd content in these leaves was established. In this study, based on the indoor spectrum, the sensitive wavebands for predicting the Cd content in leaves were determined preliminarily by correlation analysis. Partial least squares regression (PLSR) and support vector machine regression (SVMR) were used to establish spectral prediction models, and the final sensitive wavebands were determined by the size of the model index. The results show that the SVMR model exhibited higher prediction accuracy than the PLSR model. The RPDp (relative percent different of prediction set) values of the best SVMR prediction models for the pepper leaves and the eggplant leaves were 1.82 and 1.49, respectively. The values of Rp2 (coefficient of determination of prediction set), which can quantitatively estimate the Cd content in leaves, were 0.897 (p &lt; 0.01) and 0.726 (p &lt; 0.01), respectively. This study demonstrated that the leaf spectra of pepper and eggplant in the field can be used to predict the Cd content in leaves, providing a reference for monitoring the Cd content in the fruits of pepper and eggplant in the future.

Experimental Study on the Inversion of Coal Concentration in Mine Water by Visible-Near Infrared Spectroscopy

The coal concentration in mine water is the main indicator of mine water discharge. The accurate determination of coal concentration is of great significance for the purification and secondary utilization of mine water. In order to study the spectral inversion method of coal concentration in mine water, samples with different coal concentrations of 0mg/L-1000mg/L are prepared in this paper, and the ASD Field Spec 4 spectrometer is used for spectral collection (350-2500nm),It is found that the maximum influence of different coal content on spectral reflectance is 0.9. Based on this, a CKCNN (C-K -Convolutional Neural Networks) inversion model of coal content in mine water is proposed. This model uses CARS (Competitive Adapative Reweighted Sampling) algorithm to extract sensitive wave bands, and uses CNN (Convolutional Neural Networks) to establish spectral inversion model in sensitive wave bands, K-fold cross validation is used to optimize the model, the model inversion accuracy is R <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> =0.9994, RMSE=6.1401,RPD=41.9692. In this study, CKCNN was compared with five models: SPA+BF, CARS+BF, SPA+CNN, All Band +CNN and CARS+CNN. The results show that CKCNN model has the best effect. In addition, The concentration of water coal in Jiaozuo Zhongma Coal Mine is 18.75mg/L, the actual concentration measured in the laboratory is 18.92mg/L, and the inversion error is 0.17mg/L. The inversion results meet the requirements of laboratory measurement in GB11901-1989. The research results show that the hyperspectral remote sensing in the visible-near-infrared band can quickly detect the coal concentration in the mine water. The CKCNN model provides a new method for the determination of the coal content in the mine water. It is of great significance to promote the research on the influence of the coal concentration in the mine water on the visible-near-infrared spectrum.

Machine Learning-Based Ensemble Band Selection for Early Water Stress Identification in Groundnut Canopy Using UAV-Based Hyperspectral Imaging

This paper presents the early identification of water stress in groundnut canopy using unmanned aerial vehicle (UAV) based hyperspectral imaging (in 385-1020 nm) and machine learning (ML) techniques. An efficient hyperspectral imaging (HSI) data analysis pipeline was presented which includes image quality assessment, denoising, band selection, and classification. A novel ML-based ensemble feature selection (FS) algorithm has been proposed for optimal water stress sensitive waveband selection. The data analysis pipeline and the selected bands were validated on HSI data acquired at two different water stress levels. Wavelengths 515.05, 552.16, 711.92, 724.75, and 931.92 nm were identified as optimal water stress sensitive bands in groundnut canopy, using which we could identify early stress with 96.46% accuracy. The proposed data analysis pipeline and ensemble FS algorithm will benefit crop phenotyping applications such as early abiotic stress detection.

Hyperspectral imaging-based classification of rice leaf blast severity over multiple growth stages

BackgroundRice blast, which is prevalent worldwide, represents a serious threat to harvested crop yield and quality. Hyperspectral imaging, an emerging technology used in plant disease research, is a stable, repeatable method for disease grading. Current methods for assessing disease severity have mostly focused on individual growth stages rather than multiple ones. In this study, the spectral reflectance ratio (SRR) of whole leaves were calculated, the sensitive wave bands were selected using the successive projections algorithm (SPA) and the support vector machine (SVM) models were constructed to assess rice leaf blast severity over multiple growth stages.ResultsThe average accuracy, micro F1 values, and macro F1 values of the full-spectrum-based SVM model were respectively 94.75%, 0.869, and 0.883 in 2019; 92.92%, 0.823, and 0.808 in 2021; and 88.09%, 0.702, and 0.757 under the 2019–2021 combined model. The SRR–SVM model could be used to evaluate rice leaf blast disease during multiple growth stages and had good generalizability.ConclusionsThe proposed SRR data analysis method is able to eliminate differences among individuals to some extent, thus allowing for its application to assess rice leaf blast severity over multiple growth stages. Our approach, which can supplement single-stage disease-degree classification, provides a possible direction for future research on the assessment of plant disease severity during multiple growth stages.

Winter wheat and soil total nitrogen integrated monitoring based on canopy hyperspectral feature selection and fusion

Nitrogen, as a key element for crop reproductive and nutritional growth, plays an important role in overall agricultural production. Hyperspectral remote sensing (RS) offers possibility for monitoring farm nitrogen content by its non-destructive and high throughput advantages. However, the integrated high-precision acquisition of nitrogen content in crops and soils is still a technical challenge that needs to be solved urgently. This study proposed a novel approach combining feature selection based on improved grey wolf optimizer (IGWO) and feature fusion with projected gradient nonnegative matrix factorization and matrix cross recombination (PNMF-MCR) to realize the integrated monitoring of total nitrogen (TN) content in winter wheat and soil based on crop canopy hyperspectral data. Therein, the IGWO originated from GWO, additionally considered a comprehensive feature selection criterion which are the relevance with TN, the representative ability of the entire spectra, and the redundancy of the selected wavebands. The selected TN sensitive wavebands are 487, 611, 706, 817, 920 nm for winter wheat and 478, 587, 699, 812, 892 nm for soil, which could effectively represent the TN information of full spectrum from the aspects of nitrogen metabolism enzymes, chlorophyll photosynthesis, and the structure of cellular arrangements within the leaf. To further integrate the TN sensitive spectral information of crop and soil, PNMF-MCR was firstly proposed to generate a feature matrix VF by fusing the winter wheat and soil TN sensitive wavelengths. The TN inversion results based on the fused feature matrix VF achieved a high accuracy for both winter wheat and soil simultaneously. On the ground-based RS platform, the R2 and RMSE of the validation dataset for winter wheat are 0.7920 and 0.1431 g/kg. The validation results for soil are 0.7928 and 0.0037 g/kg. On the Airborne-based RS platform, the R2 and RMSE of the validation dataset for winter wheat are 0.6934 and 0.1490 g/kg. The validation results for soil are 0.6988 and 0.0040 g/kg. To verify the proposed method, predictive performances for TN of winter wheat and soil were validated in different time, at different locations for different winter wheat varieties on different platforms (ground-based platform and airborne-based platform). All above validations demonstrated the good universality and predictive ability of proposed hybrid method, which provide a feasible solution for integrated monitoring of crop and soil.

Newly-developed three-band hyperspectral vegetation index for estimating leaf relative chlorophyll content of mangrove under different severities of pest and disease

Hyperspectral imaging-derived vegetation indices (VIs) have rarely been developed to estimate leaf chlorophyll content of mangrove forests under pest and disease stress. Moreover, the optimal newly-developed hyperspectral VI is generally chosen through comparison of model accuracy alone with all possible VI combinations, which might render the conclusion one-sided. With SOC710 hyperspectral images of 119 mangrove leaf samples, this study aimed to develop a new hyperspectral VI sensitive to leaf relative chlorophyll content (SPAD value) by comprehensive comparison from five aspects (estimation accuracy, sensitivity, anti-noise performance, application to simulated EnMAP and PRISMA sensors, and spatial visualization quality). Eight types of newly-developed VIs were constructed from the sensitive wavebands selected by successive projection algorithm (SPA) method, and simple linear regression model was established using each VI. The results showed that the three-band VI ((λ757.9-λ709.4)/(λ709.4-λ708.1)) was the optimal for leaf SPAD estimation, because it had stronger correlation with SPAD, higher model accuracy of SPAD estimation using leaf and simulated hyperspectral imageries, stronger resistance to Gaussian noise, more sensitivity to extremely high chlorophyll content, and reasonable spatial visualization of SPAD. The four types of three-band VIs had higher model accuracy than the four types of two-band VIs, while two-band VIs had stronger resistance to higher Gaussian noise. Moreover, the wavelengths in the red edge region were efficient to develop hyperspectral VIs sensitive to leaf SPAD, and leaf SPAD could be more accurately estimated with pest and disease severity of 15–25%. We concluded that three-band VI consisting wavebands in the red edge region derived from leaf hyperspectral images is effective in capturing the changes of leaf chlorophyll content, which could provide great potentials for early warning of mangrove pest and disease with fine visualization details of chlorophyll content.

Extraction of reflectance spectra features for estimation of surface, subsurface, and profile soil properties

• Accuracy of reflectance spectroscopy estimates of soil properties varied by depth. • Number and location of important wavebands varied by depth and by soil property. • Some calibrations could be applied successfully across multiple depth intervals. • Results could guide development of discrete-waveband soil property sensors. Diffuse reflectance spectroscopy in the visible and near-infrared wavelength ranges has potential to provide high-resolution, pollution-free, and nondestructive estimation of soil chemical and physical properties for use in precision agriculture. Practical implementation of this approach would be facilitated if soil property sensors using a limited number of reflectance bands could maintain accuracy similar to more expensive and complex full-spectrum sensors. Studies identifying such bands are limited, especially for subsurface soils. Thus, in this study, an existing spectral database of 697 soil samples was used to compare results for three soil categories (profile, surface, and subsurface) and multiple waveband selection methods. Soil cores were obtained to approximately 1.2 m depth from ten fields, two each in Missouri, Illinois, Michigan, South Dakota, and Iowa, USA, then sieved and air-dried. Laboratory soil spectra were obtained from 350 to 2500 nm using a commercial spectrometer and soil properties (total nitrogen, soil organic carbon, total carbon, magnesium, calcium, potassium, soil texture (clay, silt, and sand) fractions, cation exchange capacity, and pH) were measured using standard laboratory analyses. The ability of ten spectral preprocessing techniques to improve analysis results was investigated. Backward interval partial least squares was used to identify those spectral regions most predictive of soil properties. Alternatively, specific characteristic wavelengths were identified by a combination genetic algorithm (GA)-back propagation neural network (BPNN) approach. Results were compared for three soil property estimation methods: (1) partial least squares regression (PLSR) models based on the full spectrum, (2) PLSR models based on sensitive regions, and (3) BPNN models based on characteristic wavelengths. The best results for profile and subsurface soils were obtained with absorbance preprocessing, but for the surface soils, the standard normal variate transformation was best. For some soil properties, the prediction R 2 of the PLSR models based on sensitive regions was better than that of the PLSR models based on the full spectrum, demonstrating that retaining only sensitive wavebands could improve estimates. However, in some cases, the reduction in wavebands decreased accuracy. Differences in prediction accuracy across all calibration models over profile and subsurface soils were relatively small but were larger for surface soils. Furthermore, application of characteristic wavelength calibrations to other soil datasets resulted in a lower accuracy than with the full spectrum calibration developed for that dataset. In general, this study shows that there are measurable differences in prediction accuracy across all calibration models over the three soil depth categories. The experimental results of this study illustrate the potential for a set of wavelengths optimized for one depth category to still provide acceptable estimates for other depth categories. Overall, these results provide important guidance for the development of DRS soil sensors based on discrete wavebands to reduce cost and increase the speed of in-field data collection.

A Sensitive Wavebands Identification System for Smart Farming

Sensing the content of macronutrients in the agricultural soil is an essential task in precision agriculture. It helps the farmers in the optimal use of fertilizers. It reduces the cost of food production and also the negative environmental impacts on atmosphere and water bodies due to indiscriminate dosage of fertilizers. The traditional chemical-based laboratory soil analysis methods do not serve the purpose as they are hardly suitable for site specific soil management. Moreover, the spectral range used in the chemical-based laboratory soil analysis may be of 350–2500 nm, which leads to redundancy and confusion. Developing sensors based on the discovery of spectral wavebands that respond to soil macronutrient concentrations, on the other hand, is an innovative and successful technology since the results are dependable and timely. The goal of this article is to use a supervised neuro-fuzzy based dimensionality reduction approach in the sensor development process to determine sensitive wavebands of soil macronutrients. Accordingly, the spectral signatures of the soil are collected in an outdoor environment and mapped with its macronutrient concentrations. In this spectral analysis, the spectral reflectance of 424 wavelengths has been obtained and these wavelengths are evaluated through combined and individual modes as well. Appropriate wavelengths are selected in each case by minimizing the fuzzy reflectance assessment index. The effectiveness of these selected wavelengths in each mode is validated by modeling the relation between the reduced reflectance space and each macronutrient concentration using Partial Least Squares Multi Variable Regression (PLS-MVR) method. Set of optimal wavebands are identified and the results are compared with the existing systems.

Optimized spectral index models for accurately retrieving soil moisture (SM) of winter wheat under water stress

Soil moisture (SM) is an important indicator of the photosynthetic rate and growth status of crops. A few related parameters, such as the red-edge parameters and spectral indices, have been adopted for retrieving the SM of winter wheat. To further study their abilities to detect the SM, field-scale water stress experiments on winter wheat were conducted during the 2018/19 growing season. The spectral ratio index in the near-infrared (NIR) shoulder region (NSRI) (700–1100 nm) was selected by comparing the correlations between the SM and the red edge parameters and spectral indices, and it was optimized using the partial least squares regression (PLSR) method. To assess the performance of the sensitive wavebands of the NSRI in retrieving the SM, three types of spectral index models were established using multiple linear regression (MLR) for the winter wheat from the jointing to the ripening stage. The results indicate that the red-edge parameters are more sensitive to the spectral variation during the jointing and flowering stages. The sensitivity decreased with increasing water stress. The red-edge area (SDr) of winter wheat irrigated in the flowering stage (D1 treatment) and irrigated in the jointing stage (D2 treatment) decreased by 20–30%, respectively. In general, all of the parameters and indices were correlated with the surface SM (0–40 cm depth), especially for the NSRI, with a significant coefficient of determination (R2) of 0.52 in the 10–20 cm depth interval (P < 0.01). Moreover, all of the spectral index models based on the optimized NSRI have good capabilities for retrieving the SM in the jointing stage. The model for one derivative of the logarithm of the NSRI (logarithmic NSRI)' performed best, with R2 and root mean square error (RMSE) values of 0.81–0.92 and 0.17–0.89%, respectively. Finally, the (logarithmic NSRI)' model was used to retrieve the SM in the flowering–ripening stage (R2 =0.85). Overall, the optimized spectral index models can accurately and quickly retrieve the SM and can assist in predicting the effect of drought on the crop yield in the future.

Extraction of Areas of Rice False Smut Infection Using UAV Hyperspectral Data

Rice false smut (RFS), caused by Ustilaginoidea virens, is a significant grain disease in rice that can lead to reduced yield and quality. In order to obtain spatiotemporal change information, multitemporal hyperspectral UAV data were used in this study to determine the sensitive wavebands for RFS identification, 665–685 and 705–880 nm. Then, two methods were used for the extraction of rice false smut-infected areas, one based on spectral similarity analysis and one based on spectral and temporal characteristics. The final overall accuracy of the two methods was 74.23 and 85.19%, respectively, showing that the second method had better prediction accuracy. In addition, the classification results of the two methods show that the areas of rice false smut infection had an expanding trend over time, which is consistent with the natural development law of rice false smut, and also shows the scientific nature of the two methods.

Application of a hyperspectral imaging system to quantify leaf-scale chlorophyll, nitrogen and chlorophyll fluorescence parameters in grapevine

Rapidly and accurately monitoring the physiological and biochemical parameters of grape leaves is the key to controlling the quality of wine grapes. In this study, a Pika L hyperspectral imaging system (400–1000 nm) was used to acquire hyperspectral image information from grape leaves. New vegetation indices were developed on the basis of the screened sensitive wavebands to quantitatively predict changes in these parameters (the leaf chlorophyll level (SPAD), leaf nitrogen content (LNC) and chlorophyll fluorescence parameters (ChlF parameters)). The results showed that SPAD reached its maximum at the grape turning stage and declined thereafter. The vegetation index (D735−D573)/(D735+D573) was able to predict SPAD fairly well (validation dataset R2 = 0.50). LNC reached its maximum at the grape maturity stage. D682/R525 was highly correlated with LNC. Except for NPQ, all ChlF parameters showed a decreasing trend from the fruiting to harvesting stages. Among the dark-adapted ChlF parameters, FV/Fm had the strongest correlation to the new vegetation index (D735−D544)/(D735+D544) (modelling dataset R2 = 0.68), and Fo had the weakest correlation. Among the light-adapted ChlF parameters, Y(II) had the strongest correlation to the new vegetation index D676/R571 (validation dataset R2 = 0.63); this index also had good predictive power for Fm' (validation dataset R2 = 0.52) but low predictive power for Fo'. All the calculated vegetation indices had weak relationships with NPQ. In addition, this study also verified the predictive abilities of vegetation indices developed in previous studies. This study can provide a technical basis for the nondestructive monitoring of the physiological and biochemical parameters of grape leaves with hyperspectral imaging systems.

Hyperspectral response and quantitative estimation on soil aggregate characters

Soil aggregate stability can be used to describe soil structure quality. In contrast, the conventional determination for soil aggregate stability, relying on dry-sieving and wet-sieving methods, is too resource consuming, especially for mass soil samples covering a large-scale field. Exploring the application of hyperspectral technology will provide a rapid, efficient and alternative method for the estimation of soil aggregate characteristics. In this study, we collected 199 soil samples (0–15 cm soil depth) from the Loess Plateau regions in the south areas of Shanxi province in China. Afterwards, we analyzed the relationship between spectra in region of 350–2500 nm and seven soil aggregate characters. The quantitative models for each soil aggregate character were built and the sensitive wavebands were determined by using the method of partial least squares. The results showed there was a negative relationship between soil spectra and soil aggregate stability, and the extreme correlation coefficient could reach −0.60 and 0.62, respectively, indicating that soil spectra could produce a significant response on soil aggregate characteristic. The significant waveband regions for soil aggregate characters were 350–870, 1250–1360, 1885–2020, and 2215–2450 nm. The important wavelengths were around 427, 599, 1326, 1890, 1926, 2208, 2311, and 2336 nm, which were also confirmed by the consistent correlation between soil spectra and each aggregate character. In addition, the reliable models (R2 = 0.612–0.714; RPD = 1.584–1.796) were produced for most of the soil aggregate indicators, while percentage of aggregate destruction (PAD) led to the worst predictive model (R2 = 0.599; RPD = 1.261). Aggregate fractal dimension (D) character was proved as the perfect indicator for spectral monitoring of soil aggregate stability, mainly due to the two-fold interpretation functions to soil structure distribution and soil adhesion with organic matter and soil moisture. It indicated that hyperspectral technology provides an alternate method for soil aggregate characters, especially for a large area.

Early detection of grapevine leafroll disease in a red-berried wine grape cultivar using hyperspectral imaging

This research was conducted to examine the potential use of hyperspectral imaging for non-destructive detection of Grapevine leafroll-associated virus 3 (GLRaV-3) during asymptomatic and symptomatic stages of grapevine leafroll disease (GLD) in a red-berried wine grape (Vitis vinifera) cultivar. Cabernet Sauvignon vines tested positive and negative for GLRaV-3 were used in this study. Leaves from infected and non-infected vines were detached at five phenological stages and individual leaf images were acquired by a hyperspectral imager in 2017, 2018, and 2019 seasons. Those images were then preprocessed using spectra normalization and Monte-Carlo method for eliminating spectral sample outliers. Least absolute shrinkage and selection operator was used to select feature wavelengths for each phenological stage within three-season datasets. The sensitivity of selected feature wavelengths was evaluated based on analysis of variance and linear regression. Six salient wavelengths (690, 715, 731, 1409, 1425 and 1582 nm) were determined as sensitive wavebands for detecting virus symptoms in leaf samples. The detectability of GLD using those six salient wavelengths was evaluated using least squares-support vector machine. The classification accuracy was found between 66.67 and 89.93% for test datasets collected at first asymptomatic stage over three seasons. These results indicated that the hyperspectral imaging technique has the potential for nondestructively detecting virus-infected grapevines during asymptomatic stages.

Hyperspectral Estimation of Soil Organic Matter Content using Different Spectral Preprocessing Techniques and PLSR Method

Soil organic matter (SOM) is the main source of soil nutrients, which are essential for the growth and development of agricultural crops. Hyperspectral remote sensing is one of the most efficient ways of estimating the SOM content. Visible, near infrared, and mid-infrared reflectance spectroscopy, combined with the partial least squares regression (PLSR) method is considered to be an effective way of determining soil properties. In this study, we used 54 different spectral pretreatments to preprocess soil spectral data. These spectral pretreatments were composed of three denoising methods, six data transformations, and three dimensionality reduction methods. The three denoising methods included no denoising (ND), Savitzky–Golay denoising (SGD), and wavelet packet denoising (WPD). The six data transformations included original spectral data, R; reciprocal, 1/R; logarithmic, log(R); reciprocal logarithmic, log(1/R); first derivative, R’; and first derivative of reciprocal, (1/R)’. The three dimensionality reduction methods included no dimensionality reduction (NDR), sensitive waveband dimensionality reduction (SWDR), and principal component analysis (PCA) dimensionality reduction (PCADR). The processed spectra were then employed to construct PLSR models for predicting the SOM content. The main results were as follows—(1) the wavelet packet denoising (WPD)-R’ and WPD-(1/R)’ data showed stronger correlations with the SOM content. Furthermore, these methods could effectively limit the correlation between the adjacent bands and, thus, prevent “overfitting”. (2) Of the 54 pretreatments investigated, WPD-(1/R)’-PCADR yielded the model with the highest accuracy and stability. (3) For the same denoising method and spectral transformation data, the accuracy of the SOM content estimation model based on SWDR was higher than that of the model based on NDR. Furthermore, the accuracy in the case of PCADR was higher than that for SWDR. (4) Dimensionality reduction was effective in preventing data overfitting. (5) The quality of the spectral data could be improved and the accuracy of the SOM content estimation model could be enhanced effectively, by using some appropriate preprocessing methods (one combining WPD and PCADR in this study).

A possible fractional order derivative and optimized spectral indices for assessing total nitrogen content in cotton

Nitrogen is the key biochemical component of chlorophyll, protein and enzymes, and it is widely used as an indicator of photosynthesis and plant nutrient levels. Hyper-spectral data-based estimation of nitrogen allows for a low-cost, effective and environmentally beneficial diagnosis of plant growth. In this paper, a novel approach to characterize the Total Nitrogen Content (TNC) from canopy spectral reflectance through a fractional order derivative (FOD) and optimized spectral indices (NDSI, RSI) is proposed. A total of 60 sampling plots designed in field experiments, canopy spectral data and total nitrogen content are tested for each plot. Optimized remote sensing indices derived from FOD spectra were applied to investigate sensitive wavebands; finally, a Support Vector Machine Regression model for estimating cotton TNC was generated. Our results showed that small FOD orders improved the spectral resolution and provided abundant absorption features; as the orders increased, the spectral strength decreased and the curves were smoothed gradually. The coefficient of correlation (R) peak appeared at the 1.25 order with a value of 0.652. The coefficient of determination (R2) between TNCs and optimized spectral indices peaked at NDSI beyond the 1.5 order (R2 = 0.592). Fourteen TNC estimation models were created via SVR methods using optimized spectral indices. Modeling results indicated that the optimal model was original reflectance-RSI, where the highest R2 was 0.784, the lowest root mean square error (RMSE) was 1.333, and the residual prediction deviation (RPD) was 1.80. Overall, FOD can potentially exploit spectral characteristics and eliminate spectral redundancy. However, original reflectance still shows a high potential for accurate predictions of the total nitrogen content in cotton.