Yong He

Zhejiang University, Hang-hsien, Zhejiang Sheng, China

Are you Yong He?

Claim your profile

Publications (447)404.99 Total impact

  • 11/2015; DOI:10.1007/s11434-015-0922-9
  • Xiao-Li Li · Chan-Jun Sun · Liu-Bin Luo · Yong He ·

    Scientific Reports 10/2015; 5:15729. DOI:10.1038/srep15729 · 5.58 Impact Factor
  • Source
    Yan-Ru Zhao · Ke-Qiang Yu · Yong He ·
    [Show abstract] [Hide abstract]
    ABSTRACT: Chemometrics methods coupled with hyperspectral imaging technology in visible and near infrared (Vis/NIR) region (380-1030 nm) were introduced to assess total soluble solids (TSS) in mulberries. Hyperspectral images of 310 mulberries were acquired by hyperspectral reflectance imaging system (512 bands) and their corresponding TSS contents were measured by a Brix meter. Random frog (RF) method was used to select important wavelengths from the full wavelengths. TSS values in mulberry fruits were predicted by partial least squares regression (PLSR) and least-square support vector machine (LS-SVM) models based on full wavelengths and the selected important wavelengths. The optimal PLSR model with 23 important wavelengths was employed to visualise the spatial distribution of TSS in tested samples, and TSS concentrations in mulberries were revealed through the TSS spatial distribution. The results declared that hyperspectral imaging is promising for determining the spatial distribution of TSS content in mulberry fruits, which provides a reference for detecting the internal quality of fruits.
    Journal of Analytical Methods in Chemistry 10/2015; 2015(2):343782. DOI:10.1155/2015/343782 · 0.79 Impact Factor
  • Shuiguang Deng · Yifei Xu · Xiaoli Li · Yong He ·
    [Show abstract] [Hide abstract]
    ABSTRACT: Near infrared (NIR) hyperspectral imaging has been used as a rapid non-destructive technique to predict moisture content of tea. To improve the performance of predicting, we first find and validate the fact that the texture near the veins is continues and directional. And then we propose Three-Dimension Gabor Filter (TDGF) and its corresponding filterbank to describe the textures of tealeaf. After that we construct two types of models based on partial least squares (PLS) regression. Experiments are conducted to predict the moisture content of Longjing tea, and different regression models based on different types of features are built for comparison. The results show that the proposed filterbank is able to detect the optimal direction of water flow and the model combining the spectrum and TDGF textures outperform the other comparative models.
    Computers and Electronics in Agriculture 10/2015; 118:38-46. DOI:10.1016/j.compag.2015.08.014 · 1.76 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Hyperspectral imaging combined with feature extraction methods were applied to determine soluble sugar content (SSC) in mature and scatheless strawberry. Hyperspectral images of 154 strawberries covering the spectral range of 874-1,734 nm were captured and the spectral data were extracted from the hyperspectral images, and the spectra of 941~1,612 nm were preprocessed by moving average (MA). Nineteen samples were defined as outliers by the residual method, and the remaining 135 samples were divided into the calibration set (n = 90) and the prediction set (n = 45). Successive projections algorithm (SPA), genetic algorithm partial least squares (GAPLS) combined with SPA, weighted regression coefficient (Bw) and competitive adaptive reweighted sampling (CARS) were applied to select 14, 17, 24 and 25 effective wavelengths, respectively. Principal component analysis (PCA) and wavelet transform (WT) were applied to extract feature information with 20 and 58 features, respectively. PLS models were built based on the full spectra, the effective wavelengths and the features, respectively. All PLS models obtained good results. PLS models using full-spectra and features extracted by WT obtained the best results with correlation coefficient of calibration (r(c)) and correlation coefficient of prediction (r(p)) over 0.9. The overall results indicated that hyperspectral imaging combined with feature extraction methods could be used for detection of SSC in strawberry.
    Guang pu xue yu guang pu fen xi = Guang pu 07/2015; 35(4):1020-4. DOI:10.3964/j.issn.1000-0593(2015)04-1020-05 · 0.29 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: The pyrolysis characteristics of two Chinese coals, two biomass materials, and their blends were investigated by both experimental and numerical methods. Single particles of the coal and biomass were prepared for the pyrolysis experiment through grinding and pressing, while the blended particles were made by mixing the coal and biomass powder with different ratios before the pressing. Sample particles pyrolyzed in a single-particle reactor system, with the time history of the particle temperature and mass recorded. The analysis of the measured pyrolysis data of the coal, biomass, and coal–biomass blends indicate the absence of a synergistic effect between the coal and biomass pyrolysis. A numerical method coupling the chemical percolation devolatilization (CPD) model with a particle energy equation was employed to analyze the pyrolysis process. The model prediction agreed well with the experimental data for different particle diameters, fuel types, and blend mixing conditions. The fact that the co-pyrolysis of blended coal–biomass particles is well-predicted by the simple addition of the individual pyrolysis characteristics of its components also corroborates the lack of synergistic interactions. These findings will be useful for the co-combustion modeling of coal–biomass blends.
    Energy & Fuels 07/2015; 29(8):150713125709001. DOI:10.1021/acs.energyfuels.5b00646 · 2.79 Impact Factor
  • Feng-le Zhu · Yong He · Yong-ni Shao ·
    [Show abstract] [Hide abstract]
    ABSTRACT: Near-infrared hyperspectral imaging technique was employed in the present study to determine water contents in salmon flesh rapidly and nondestructively. Altogether 90 samples from different positions of salmon fish were collected for hyperspectral image scanning, and mean spectra were extracted from the region of interest (ROI) inside each image. Sixty samples were randomly selected as calibration set, and the remaining 30 samples formed prediction set. The full-spectrum and water contents were correlated using partial least squares regression (PLSR) and least-squares support vector machines (LS-SVM), which were then applied to predict water contents for prediction samples. A novel variable extraction method called random frog was applied to select effective wavelengths (EWs) from the full-spectrum. PLSR and LS-SVM calibration models were established respectively to detect water contents in salmon based on the EWs. Though the performances of EWs-based models were worse than models using full-spectrum, only 12 wavelengths were used to substitute for the original 151 wavelengths, thus models were greatly simplified and more suitable for practical application. For EWs-based PLSR and LS-SVM models, satisfactory results were achieved with correlation coefficient of prediction (Rp) of 0. 92 and 0. 93 respectively, and root mean square error of prediction (RMSEP) of 1. 31% and 1. 18% respectively. The results indicated that near-infrared hyperspectral imaging combined with chemometrics allows accurate prediction of water contents in salmon flesh, providing important reference for the rapid inspection of fish quality.
    Guang pu xue yu guang pu fen xi = Guang pu 05/2015; 35(1):113-7. DOI:10.3964/j.issn.1000-0593(2015)01-0113-05 · 0.29 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Classification is a critical step to make full use of the hyperspectral data. The most current approaches perform well for analyzing the macro texture, but they often fail to deal with the micro texture. Thus, this study proposes a general framework for the material with micro texture based on Hyperspectral Image (HSI) technique. In this framework, Local Response Pattern (LRP) is firstly proposed to describe 2D image texture to preserve more structural information and keep less sensitive to image conditions. Then, LRP is extended to represent HSI with Texture Enhancement (TE) by considering opponent relationships between pairs of bands. After that, Discriminated Locality Preserving Projection (DLPP) is proposed to reduce data dimension in a linearizing nonlinear manifold way. Finally, experiments on the hyperspectral images of fresh and frozen-thawed fish fillets are conducted. The results demonstrate that the proposed framework is efficient in terms of both recognition rates and robustness.
    Information Sciences 04/2015; 299. DOI:10.1016/j.ins.2014.12.025 · 4.04 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: NaOH pretreatment is a convenient and effective method which is widely used in rice straw anaerobic digestion. But the mechanism of the alkaline (NaOH) hydrolysis of biopolymers compositions and polymeric cross-linked network structures ofrice straw cell wall need further study. This paper firstly studied the effect and mechanism of alkali pretreatment on anaerobic digestion and biogas production of rice straw by using a combination of confocal Raman microscopy and transmission electron microscope. First, the original rice straw and the rice straw pretreated by NaOH were taken for mapping scanning by confocal Raman microscopy withmicron-scale spatial resolution. Then principal component analysis was adoptedto extract main information of Raman spectra, it could be found that the two types of samples were respectively presented with ray-like distribution in the first two principal component space, which were with cumulative contribution of 99%. And there was a clear boundary between the two types of samples without any overlapping, indicating that there was a significant difference of Raman spectralcharacteristic between original rice leaf and rice leaf pretreated by NaOH. Further analysis of the loading weights of the first two principal components showed that the Raman peaks at 1 739, 1 508 and 1 094 cm-1 were the important bands, and these three Raman peaks were attributed to the scattering of hemicellulose, cellulose and lignin respectively. Following, chemical imaging analysisof hemicellulose, cellulose and lignin were achieved by combining these Raman peaks and microscopic image information. It could be found that the NaOH pretreatment resulted in a loss of dense spatial uniformity structure of tissue and great decreases of the contents of these three ingredients, particularly lignin. It can be concluded that it is feasible to non-destructively measure hemicellulose, lignin and cellulose in rice straw tissue by confocal Raman microscopy, and toachieve chemical imaging analysis of the three ingredients in tissue, and this research will be much help for revealing the promotion mechanism of NaOH pretreatment for the rice straw fermentation and biogas production.
    Guang pu xue yu guang pu fen xi = Guang pu 03/2015; 35(3). DOI:10.3964/j.issn.1000-0593(2015)03-0657-06 · 0.29 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: In order to estimate pepper plant growth rapidly and accurately, hyperspectral imaging technology combined with chemometrics methods were employed to realize visualization of nitrogen content (NC) distribution. First, pepper leaves were picked up with the leaf number based on different leaf positions, and hyperspectraldata of these leaves were acquired. Then, SPAD and NC value of leaves were measured, respectively. After acquirement of pepper leaves’ spectral information, random-frog (RF) algorithm was chosen to extract characteristic wavelengths. Finally, five characteristic wavelengths were selected respectively, and then thosecharacteristic wavelengths and full spectra were used to establish partial least squares regression (PLSR) models, respectively. As a result, SPAD predicted model had an excellent performance of R C=0.970, R CV=0.965, R P=0.934, meanwhile evaluation parameters of NC predicted model were R C=0.857, R CV=0.806, R P=0.839. Lastly, according to the optimal models, SPAD and NC of each pixel in hyperspectral images of pepper leaves were calculated and their distribution was mapped. In fact, SPAD in plant can reflectthe NC. In this research, the change trend of both was similar, so the conclusions of this research were proved to be corrected. The results revealed that it was feasible to apply hyperspectral imaging technology for mapping SPAD and NC inpepper leaf, which provided a theoretical foundation for monitoring plant growth and distribution of nutrients.
    Guang pu xue yu guang pu fen xi = Guang pu 03/2015; 35(3). DOI:10.3964/j.issn.1000-0593(2015)03-0746-05 · 0.29 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Nitrogen is a necessary and important element for the growth and development of fruit orchards. Timely, accurate and nondestructive monitoring of nitrogen status in fruit orchards would help maintain the fruit quality and efficient production of the orchard, and mitigate the pollution of water resources caused by excessive nitrogen fertilization. This study investigated the capability of hyperspectral imagery for estimating and visualizing the nitrogen content in citrus canopy. Hyperspectral images were obtained for leaf samples in laboratory as well as for the whole canopy in the field with ImSpector V10E (Spectral Imaging Ltd., Oulu, Finland). The spectral datas for each leaf sample were represented by the average spectral data extracted from the selected region of interest (ROI) in the hyperspectral images with the aid of ENVI software. The nitrogen content in each leaf sample was measured by the Dumas combustion method with the rapid N cube (Elementar Analytical, Germany). Simple correlation analysis and the two band vegetation index (TBVI) were then used to develop the spectra data-based nitrogen content prediction models. Results obtained through the formula calculation indicated that the model with the two band vegetation index (TBVI) based on the wavelengths 811 and 856 nm achieved the optimal estimation of nitrogen content in citrus leaves (R 2=0.607 1). Furthermore, the canopy image for the identified TBVI was calculated, and the nitrogen content of the canopy was visualized by incorporating the model into the TBVI image. The tender leaves, middle-aged leaves and elder leaves showed distinct nitrogen status from highto low-levels in the canopy image. The results suggested the potential of hyperspectral imagery for the nondestructive detection and diagnosis of nitrogen status in citrus canopy in real time. Different from previous studies focused on nitrogen content prediction at leaf level, this study succeeded in predicting and visualizing the nutrient content of fruit trees at canopy level. This would provide valuable information for the implementation of individual tree-based fertilization schemes in precision orchard management practices.
    Guang pu xue yu guang pu fen xi = Guang pu 03/2015; 35(3). DOI:10.3964/j.issn.1000-0593(2015)03-0715-04 · 0.29 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Simultaneous two-species imaging using single-shot planar laser-induced fluorescence have been performed to record high quality image pairs of CH/OH, CH/CH2O and OH/CH2O to visualize the flame front structures in swirl-stabilized lean premixed methane/air flames. The results show that the investigated flames exhibit various flame front structures distinctly in space, which covers: (1) the corrugated flamelet at the leading front; (2) the thin reaction-zone layer with distorted preheat zone in the shear-layer downstream; and (3) quenching, re-ignition and distributed reactions further downstream. The large variation of the flame characteristics in space stems from the entrainment of ambient cold air to the flame that results in flame quenching at the trailing edge of the flame. Thereafter, the unburned fuel/air mixture in the downstream region mixes with the entrained air and the hot combustion products from the upstream leading flame front, leading to reignition with distributed reactions. The current results provide a direct experimental evidence that distributed reactions can be a common combustion mode along with the results (Ref. [1], Zhou et al., 2015) recently reported in the highly turbulent premixed jet flames.
    Combustion and Flame 03/2015; 162(7). DOI:10.1016/j.combustflame.2015.02.018 · 3.08 Impact Factor
  • Xi-bin Ding · Fei Liu · Chu Zhang · Yong He ·
    [Show abstract] [Hide abstract]
    ABSTRACT: In the present work, prediction models of SPAD value (Soil and Plant Analyzer Development, often used as a parameter to indicate chlorophyll content) in oilseed rape leaves were successfully built using hyperspectral imaging technique. The hy perspectral images of 160 oilseed rape leaf samples in the spectral range of 380-1030 nm were acquired. Average spectrum was extracted from the region of interest (ROI) of each sample. We chose spectral data in the spectral range of 500-900 nm for analysis. Using Monte Carlo partial least squares(MC-PLS) algorithm, 13 samples were identified as outliers and eliminated. Based on the spectral information and measured SPAD values of the rest 147 samples, several estimation models have been built based on different parameters using different algorithms for comparison, including: (1) a SPAD value estimation model based on partial least squares(PLS) in the whole wavelength region of 500-900 nm; (2) a SPAD value estimation model based on successive projections algorithmcombined with PLS(SPA-PLS); (3) 4 kind of simple experience SPAD value estimation models in which red edge position was used as an argument; (4) 4 kind of simple experience SPAD value estimation models in which three vegetation indexes R710/R760, (R750-R705)/(R750-R705) and R860/(R550 x R708), which all have been proved to have a good relevance with chlorophyll content, were used as an argument respectively; (5) a SPAD value estimation model based on PLS using the 3 vegetation indexes mentioned above. The results indicate that the optimal prediction performance is achieved by PLS model in the whole wavelength region of 500-900 nm, which has a correlation coefficient(r(p)) of 0.8339 and a root mean squares error of predicted (RMSEP) of 1.52. The SPA-PLS model can provide avery close prediction result while the calibration computation has been significantly reduced and the calibration speed has been accelerated sharply. For simple experience models based on red edge parameters and vegetation indexes, although there is a slight gap between theprediction performance and that of the PLS model in the whole wavelength region of 500-900 nm, they also have their own unique advantages which should be thought highly of: these models are much simpler and thus the calibration computation is reduced significantly, they can perform an important function under circumstances in which increasing modeling speed and reducing calibration computation operand are more important than improving the prediction accuracy, such as the development of portable devices.
    Guang pu xue yu guang pu fen xi = Guang pu 02/2015; 35(2):486-91. DOI:10.3964/j.issn.1000-0593(2015)02-0486-06 · 0.29 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: This paper reports a combined experimental and numerical investigation of the pyrolysis characteristics of coal, biomass, and coal–biomass blends. Coal and straw were grounded and pressed into spherical particles with diameter of 8 mm, and blended coal–straw particles were prepared through mixing pulverized coal and straw before pressed into particles. Sample particles were suspended in the center of a single-particle reactor system and devolatilized under different temperatures. The analysis of the time history of the residual mass of particles of coal, straw, and coal–straw blends suggested an absence of synergistic effect between the coal and the straw. In addition, a one-dimensional, time-dependent particle model; based on the chemical percolation devolatilization (CPD) and bio-CPD models; was developed to simulate the pyrolysis of coal and straw particles. The model predictions agree will with the measured data. An extended CPD model was proposed to explain the co-pyrolysis characteristics of coal–biomass blends. Encouraging agreement was found between the predicted and the experimental results of pyrolysis of coal–straw blends.
    Fuel 01/2015; 139:356–364. DOI:10.1016/j.fuel.2014.08.069 · 3.52 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Visible/near-infrared (Vis/NIR) hyperspectral imaging was employed to determine the spatial distribution of total nitrogen in pepper plant. Hyperspectral images of samples (leaves, stems, and roots of pepper plants) were acquired and their total nitrogen contents (TNCs) were measured using Dumas combustion method. Mean spectra of all samples were extracted from regions of interest (ROIs) in hyperspectral images. Random frog (RF) algorithm was implemented to select important wavelengths which carried effective information for predicting the TNCs in leaf, stem, root, and whole-plant (leaf-stem-root), respectively. Based on full spectra and the selected important wavelengths, the quantitative relationships between spectral data and the corresponding TNCs in organs (leaf, stem, and root) and whole-plant (leaf-stem-root) were separately developed using partial least-squares regression (PLSR). As a result, the PLSR model built by the important wavelengths for predicting TNCs in whole-plant (leaf-stem-root) offered a promising result of correlation coefficient (R) for prediction (R-P=0.876) and root mean square error (RMSE) for prediction (RMSEP=0.426%). Finally, the TNC of each pixel within ROI of the sample was estimated to generate the spatial distribution map of TNC in pepper plant. The achievements of the research indicated that hyperspectral imaging is promising and presents a powerful potential to determine nitrogen contents spatial distribution in pepper plant.
    PLoS ONE 12/2014; 9(12):e116205. DOI:10.1371/journal.pone.0116205 · 3.23 Impact Factor
  • Source
    Ke-Qiang Yu · Yan-Ru Zhao · Zi-Yi Liu · Xiao-Li Li · Fei Liu · Yong He ·
    [Show abstract] [Hide abstract]
    ABSTRACT: The intent of present work was to develop a valid method for detection of defective features in loquat fruits based on hyperspectral imaging. A laboratorial hyperspectral imaging device covering the visible and near-infrared region of 380–1,030 nm was utilized to acquire the loquat hyperspectral images. The corresponding spectral data were extracted from the region of interests of loquat hyperspectral images. The dummy grades were assigned to the defective and normal group of loquats, separately. Competitive adaptive reweighted sampling (CARS) was conducted to elect optimal sensitive wavelengths (SWs) which carried the most important spectral information on identifying defective and normal samples. As a result, 12 SWs at 433, 469, 519, 555, 575, 619, 899, 912, 938, 945, 970, and 998 nm were selected, respectively. Then, the partial least squares discriminant analysis (PLS-DA) model was established using the selected SWs. The results demonstrated that the CARS-PLS-DA model with the discrimination accuracy of 98.51 % had a capability of classifying two groups of loquats. Based on the characteristics of image information, minimum noise fraction (MNF) rotation was implemented on the hyperspectral images at SWs. Finally, an effective approach for detecting the defective features was exploited based on the images of MNF bands with “region growing” algorithm. For all investigated loquat samples, the developed program led to an overall detection accuracy of 92.3 %. The research revealed that the hyperspectral imaging technique is a promising tool for detecting defective features in loquat, which could provide a theoretical reference and basis for designing classification system of fruits in further work.
    Food and Bioprocess Technology 11/2014; 7(11). DOI:10.1007/s11947-014-1357-z · 2.69 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Background Acute rejection (AR) remains a life-threatening complication after orthotopic liver transplantation (OLT) and there are few available diagnostic biomarkers clinically for AR. This study aims to identify intestinal microbial profile and explore potential application of microbial profile as a biomarker for AR after OLT. Methods The OLT models in rats were established. Hepatic graft histology, ultrastructure, function, and intestinal barrier function were tested. Ileocecal contents were collected for intestinal microbial analysis. Results Hepatic graft suffered from the ischemia-reperfusion (I/R) injury on day 1, initial AR on day 3, and severe AR on day 7 after OLT. Real-time quantitative polymerase chain reaction results showed that genus Faecalibacterium prausnitzii and Lactobacillus were decreased, whereas Clostridium bolteae was increased during AR. Notably, cluster analysis of denaturing gradient gel electrophoresis (DGGE) profiles showed the 7AR and 3AR groups clustered together with 73.4% similarity, suggesting that intestinal microbiota was more sensitive than hepatic function in responding to AR. Microbial diversity and species richness were decreased during AR. Phylogenetic tree analysis showed that most of the decreased key bacteria belonged to phylum Firmicutes, whereas increased key bacteria belonged to phylum Bacteroidetes. Moreover, intestinal microvilli loss and tight junction damage were noted, and intestinal barrier dysfunction during AR presented a decrease of fecal secretory immunoglobulin A (sIgA) and increase of blood bacteremia, endotoxin, and tumor necrosis factor-α. Conclusion We dynamically detail intestinal microbial characterization and find a high sensitivity of microbial change during AR after OLT, suggesting that intestinal microbial variation may predict AR in early phase and become an assistant therapeutic target to improve rejection after OLT.
    Transplantation 10/2014; 98(8):844-852. DOI:10.1097/TP.0000000000000334 · 3.83 Impact Factor
  • Shuiguang Deng · Yifei Xu · Xiaoli Li · Yong He ·
    [Show abstract] [Hide abstract]
    ABSTRACT: Spectral unmixing is a critical issue in multi-spectral data processing, which has the ability to identify the constituent components of a pixel. Most of the hyperspectral unmixing current methods are based on Linear Mixture Model (LMM) and have been widely used in many scenarios. However, both the noise contained in the LMM and the requirement of essential prior knowledge strongly limit their practical applications. In order to address these issues, this paper proposes an iterative approach named CBIGMM. It utilizes infinite Gaussian mixture model to describe the hyperspectral data, which is robust to the noise due to the intrinsic randomness of the Gaussian components; and extracts the endmembers and their corresponding abundance in a fully unsupervised way without prior knowledge. A set of experiment is conducted on both synthetic and real data set from pesticide-contaminated vegetables. The results and analyses show CBIGMM outperforms other methods in addressing hyperspectral unmixing problem.
    Expert Systems with Applications 10/2014; 42(4). DOI:10.1016/j.eswa.2014.09.059 · 2.24 Impact Factor
  • Shu-Xi Cheng · Wen-Wen Kong · Chu Zhang · Fei Liu · Yong He ·
    [Show abstract] [Hide abstract]
    ABSTRACT: The variety of Chinese cabbage seeds were recognized using hyperspectral imaging with 256 bands from 874 to 1,734 nm in the present paper. A total of 239 Chinese cabbage seed samples including 8 varieties were acquired by hyperspectral image system, 158 for calibration and the rest 81 for validation. A region of 15 pixel x 15 pixel was selected as region of interest (ROI) and the average spectral information of ROI was obtained as sample spectral information. Multiplicative scatter correction was selected as pretreatment method to reduce the noise of spectrum. The performance of four classification algorithms including Ada-boost algorithm, extreme learning machine (ELM), random forest (RF) and support vector machine (SVM) were examined in this study. In order to simplify the input variables, 10 effective wavelengths (EMS) including 1,002, 1,005, 1,015, 1,019, 1,022, 1,103, 1,106, 1,167, 1,237 and 1,409 nm were selected by analysis of variable load distribution in PLS model. The reflectance of effective wavelengths was taken as the input variables to build effective wavelengths based models. The results indicated that the classification accuracy of the four models based on full-spectral were over 90%, the optimal models were extreme learning machine and random forest, and the classification accuracy achieved 100%. The classification accuracy of effective wavelengths based models declined slightly but the input variables compressed greatly, the efficiency of data processing was improved, and the classification accuracy of EW-ELM model achieved 100%. ELM performed well both in full-spectral model and in effective wavelength based model in this study, it was proven to be a useful tool for spectral analysis. So rapid and nondestructive recognition of Chinese cabbage seeds by hyperspectral imaging combined with machine learning is feasible, and it provides a new method for on line batch variety recognition of Chinese cabbage seeds.
    Guang pu xue yu guang pu fen xi = Guang pu 09/2014; 34(9):2519-22. DOI:10.3964/j.issn.1000-0593(2014)09-2519-04 · 0.29 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Visible and near infrared (Vis-NIR) hyperspectral imaging system was carried out to rapidly determinate the content and estimate the distribution of nitrogen (N) in oilseed rape leaves. Hyperspectral images of 420 leaf samples were acquired at seedling, flowering and pod stages. The spectral data of rape leaves were extracted from the region of interest (ROI) in the wave- length range of 380-1,030 nm. Different spectra preprocessing including Savitzky-Golay smoothing (SG), standard normal variate (SNV), multiplicative scatter correction (MSC), first and second derivatives were applied to improve the signal to noise ratio. Among 471 wavelengths, only twelve wavelengths (467, 557, 665, 686, 706, 752, 874, 879, 886, 900, 978 and 995 nm) were selected by successive projections algorithm(SPA) as the effective wavelengths for N prediction. Based on these effective wavelengths, partial least squares(PLS) and least-squares support vector machines (LS-SVM) calibration models were established for the determination of N content. Reasonable estimation accuracy was obtained, with Rp of 0.807 and RMSEP of 0.387 by PLS and Rp of 0.836 and RMSEP of 0.358 by LS-SVM, respectively. Considering the simple structure and satisfying results of PLS model, SPA-PLS model was used to generate the distribution maps of N content in rape leaves. The concentrations of N were calculated at each pixel of hyperspectral images at the selected effective wavelengths by inputting its correspond- ing spectrum into the established SPA-PLS model. Different colour represented the change in N content in the rape leaves under different fertilizer treatments. By including all pixels within the selected ROI, the average N status can be displayed in more detail and visualised. The visualization of N distribution could be helpful to understanding the change in N content in rape leaves during rape growth period and facilitate discovering the difference of N content within one sample as well as among the samples from different fertilising plots. The overall results revealed that hyperspectral imaging is a promising technique to detect N content and distribution within oilseed rape leaves rapidly and nondestructively.
    Guang pu xue yu guang pu fen xi = Guang pu 09/2014; 34(9):2513-8. DOI:10.3964/j.issn.1000-0593(2014)09-2513-06 · 0.29 Impact Factor

Publication Stats

4k Citations
404.99 Total Impact Points


  • 1970-2015
    • Zhejiang University
      • • State Key Lab of Clean Energy Utilization
      • • School of Biosystems Engineering and Food Science
      • • School of Medicine
      • • Department of Mathematics
      Hang-hsien, Zhejiang Sheng, China
  • 2013
    • Zhejiang Medical University
      • First Affiliated Hospital
      Hang-hsien, Zhejiang Sheng, China
    • Lund University
      • Division of Combustion Physics
      Lund, Skåne, Sweden
  • 2009
    • Hangzhou Normal University
      Hang-hsien, Zhejiang Sheng, China
  • 2008
    • Zhejiang Technical Institute of Economics
      Hang-hsien, Zhejiang Sheng, China
    • Xiamen University
      • Department of Physics
      Amoy, Fujian, China