A classification method for hyperspectral imaging of Fusarium head blight disease symptom based on deep convolutional neural network

doi:10.3969/j.issn.1004-1524.2019.02.18

›› 2019, Vol. 31 ›› Issue (2): 315-325.DOI: 10.3969/j.issn.1004-1524.2019.02.18

• Biosystems Engineering • Previous Articles Next Articles

A classification method for hyperspectral imaging of Fusarium head blight disease symptom based on deep convolutional neural network

JIN Xiu^a, LU Jie^b, FU Yunzhi^a, WANG Shuai^a, XU Gaojian^a, LI Shaowen^{a, *}

a. School of Information and Computer, Anhui Agricultural University, Hefei 230036, China;
b. College of Agronomy, Anhui Agricultural University, Hefei 230036, China

Received:2018-06-14 Online:2019-02-25 Published:2019-03-06

Abstract

Abstract: In order to realize rapid and early diagnosis of wheat Fusarium head blight disease via hyperspectral imaging, the correlation between the convolutional layer and the spectrum feature of disease symptom was analyzed, and the classification modeling of hyperspectral image was studied. Two typical neural network, Visual Geometry Group (VGG) and residual neural network (ResNet), were introduced to construct the convolutional neural network with different depth. By comparing the training and testing results of the hyperspectral data set for wheat Fusarium head blight disease, it was shown that the loss value decreased with the increased depth of VGG structure, yet the loss value of validation set was not significantly decreased with the increased ResNet depth. According to the evaluation results for testing set, the VGG network of 4 basic units with 22 layers showed the best performance, as its accuracy of training, validation and testing was 0.846, 0843 and 0.742, respectively. Therefore, the VGG network could effectively extract the spectrum feature of Fusarium head blight disease. These results would provide theoretical basis for the intellectual diagnosis of wheat Fusarium head blight disease by the remote sensing in a large scale.

Key words: hyperspectral imaging, deep convolution neural network, Fusarium head blight disease, classification modeling

CLC Number:

S127

JIN Xiu, LU Jie, FU Yunzhi, WANG Shuai, XU Gaojian, LI Shaowen. A classification method for hyperspectral imaging of Fusarium head blight disease symptom based on deep convolutional neural network[J]. , 2019, 31(2): 315-325.

References

[1] 李迎春, 马进. 冬小麦赤霉病发生规律及防治技术[J]. 陕西农业科学, 2016, 62(5):125-126.
LI Y C, MA J.Occurance and prevention and control technology of winter wheat scab[J]. Shaanxi Journal of Agricultural Sciences, 2016, 62(5): 125-126. (in Chinese)
[2] PALACIOS S A, ERAZO J G, CIASCA B, et al.Occurrence of deoxynivalenol and deoxynivalenol-3-glucoside in durum wheat from Argentina[J]. Food Chemistry, 2017, 230: 728-734.
[3] MALLMANN C A, DILKIN P, MALLMANN A O, et al.Prevalence and levels of deoxynivalenol and zearalenone in commercial barley and wheat grain produced in Southern Brazil: an eight-year (2008 to 2015) summary[J]. Tropical Plant Pathology, 2017, 42(3): 146-152.
[4] PEIRIS K H S, DONG Y, DAVIS M A, et al. Estimation of the deoxynivalenol and moisture contents of bulk wheat grain samples by FT-NIR spectroscopy[J]. Cereal Chemistry, 2017, 94(4): 677-682.
[5] BAURIEGEL E, GIEBEL A, GEYER M, et al.Early detection of Fusarium infection in wheat using hyper-spectral imaging[J]. Computers and Electronics in Agriculture, 2011, 75(2): 304-312.
[6] DAMMER K H, MÖLLER B, RODEMANN B, et al. Detection of head blight (Fusarium ssp.) in winter wheat by color and multispectral image analyses[J]. Crop Protection, 2011, 30(4): 420-428.
[7] 梁琨, 杜莹莹, 卢伟, 等. 基于高光谱成像技术的小麦籽粒赤霉病识别[J]. 农业机械学报, 2016, 47(2):309-315.
LIANG K, DU Y Y, LU W, et al.Identification of Fusarium head blight wheat based on hyperspectral imaging technology[J]. Transactions of the Chinese Society for Agricultural Machinery, 2016, 47(2): 309-315. (in Chinese with English abstract)
[8] WANG H, QIN F, RUAN L, et al.Identification and severity determination of wheat stripe rust and wheat leaf rust based on hyperspectral data acquired using a black-paper-based measuring method[J]. PLoS One, 2016, 11(4): e0154648.
[9] 刘占宇, 祝增荣, 赵敏, 等. 基于主成分分析和人工神经网络的稻穗健康状态的高光谱识别[J]. 浙江农业学报, 2011, 23(3):607-616.
LIU Z Y, ZHU Z R, ZHAO M, et al.Hyperspectral discrimination of different health conditions in rice panicles based on principal component analysis and artificial neural network[J]. Acta Agriculturae Zhejiangensis, 2011, 23(3): 607-616. (in Chinese with English abstract)
[10] 齐浩, 王振锡, 岳俊, 等. 基于叶片光谱特征的南疆盆地主栽果树树种遥感识别[J]. 浙江农业学报, 2015, 27(12): 2142-2146.
QI H, WANG Z X, YUE J, et al.Remote sensing identification of main fruit tree species based on leaf spectral feature in southern Xinjiang basin[J]. Acta Agriculturae Zhejiangensis, 2015, 27(12): 2142-2146. (in Chinese with English abstract)
[11] MOHANTY S P, HUGHES D P, SALATHÉ M.Using deep learning for image-based plant disease detection[J]. Frontiers in Plant Science, 2016, 7: 1419.
[12] PICON A, ALVAREZ-GILA A, SEITZ M, et al. Deep convolutional neural networks for mobile capture device-based crop disease classification in the wild[EB/OL]. (2018-04-21) [2018-06-14]. https://www.sciencedirect.com/science/article/pii/S0168169917312619.
[13] YUE J, ZHAO W, MAO S, et al.Spectral-spatial classification of hyperspectral images using deep convolutional neural networks[J]. Remote Sensing Letters, 2015, 6(6): 468-477.
[14] HU W, HUANG Y, WEI L, et al.Deep convolutional neural networks for hyperspectral image classification[J]. Journal of Sensors, 2015, 2015: 258619.
[15] MAKANTASIS K, KARANTZALOS K, DOULAMIS A, et al.Deep supervised learning for hyperspectral data classification through convolutional neural networks[C]//2015 IEEE International Geoscience and Remote Sensing Symposium (IGARSS), 2015: 4959-4962.
[16] HAN S S, PARK G H, LIM W, et al.Deep neural networks show an equivalent and often superior performance to dermatologists in onychomycosis diagnosis: automatic construction of onychomycosis datasets by region-based convolutional deep neural network[J]. PLoS One, 2018, 13(1): e0191493.
[17] TANG Z, LUO L, PENG H, et al.A joint residual network with paired ReLUs activation for image super-resolution[J]. Neurocomputing, 2018, 273: 37-46.
[18] LE T H N, DUONG C N, HAN L, et al. Deep contextual recurrent residual networks for scene labeling[J]. Pattern Recognition, 2018, 80: 32-41.
[19] GÜÇLÜTÜRK Y, GÜÇLÜ U, BARO X, et al. Multimodal first impression analysis with deep residual networks[J]. IEEE Transactions on Affective Computing, 2017, 9(3): 316-329.
[20] YU L, CHEN H, DOU Q, et al.Automated melanoma recognition in dermoscopy images via very deep residual networks[J]. IEEE Transactions on Medical Imaging, 2017, 36(4): 994-1004.
[21] LIU Z, WU Q, ZHANG Y, et al.Adaptive least squares support vector machines filter for hand tremor canceling in microsurgery[J]. International Journal of Machine Learning and Cybernetics, 2011, 2(1): 37-47.
[22] CLEVERT D A, UNTERTHINER T, HOCHREITER S. Fast and accurate deep network learning by exponential linear units (ELUs)[EB/OL]. (2015-12-03)[2018-06-14]. https://arxiv.org/pdf/1511.07289v2.pdf.
[23] BOURDÈS V, BONNEVAY S, LISBOA P, et al. Comparison of artificial neural network with logistic regression as classification models for variable selection for prediction of breast cancer patient outcomes[J]. Advances in Artificial Neural Systems, 2010, 2010: 309841.
[24] VINCENT P, LAROCHELLE H, BENGIO Y, et al.Extracting and composing robust features with denoising autoencoders[C]// Proceedings of the 25th International Conference on Machine Learning, 2008: 1096-1103.
[25] LEE Y M, KIM J H.Robust and reliable feature extractor training by using unsupervised pre-training with self-organization map[M]//Robot Intelligence Technology and Applications. Cham, Switzerland: Springer, 2015: 159-170.
[26] CHIN W S, ZHUANG Y, JUAN Y C, et al.A learning-rate schedule for stochastic gradient methods to matrix factorization[C]//Pacific-Asia Conference on Knowledge Discovery and Data Mining. Cham, Switzerland: Springer, 2015: 442-455.
[27] SRIVASTAVA N, HINTON G, KRIZHEVSKY A, et al.Dropout: a simple way to prevent neural networks from overfitting[J]. The Journal of Machine Learning Research, 2014, 15(1): 1929-1958.

A classification method for hyperspectral imaging of Fusarium head blight disease symptom based on deep convolutional neural network

RichHTML

PDF (PC)

Knowledge

Abstract

Cite this article

share this article

References

Related Articles 4

Recommended Articles

Metrics

Comments

[1]	YU Wenjie, WANG Caixia, QIAO Lu, WANG Songlei, HE Xiaoguang. Construction of PLSR prediction model for detecting color of Jingyuan yellow beef by hyperspectral technique [J]. , 2020, 32(3): 527-533.
[2]	MENG Qinglong, ZHANG Yan, SHANG Jing. Nondestructive recognition of surface defect on kiwifruits using hyperspectral imaging technology [J]. , 2019, 31(8): 1372-1378.
[3]	WANG Yanxiang, ZHANG Yan, YANG Chengya, MENG Qinglong, SHANG Jing. Advances in new nondestructive detection and identification techniques of crop diseases based on deep learning [J]. , 2019, 31(4): 669-676.
[4]	MENG Youqing, WENG Haiyong, CEN Haiyan, LI Hongye, HE Yong. Carbohydrate metabolism and near-infrared spectral characteristics in asymptomatic Huang-longbing infected leaves [J]. , 2019, 31(3): 428-435.