Investigation on Data Fusion of Multisource Spectral Data for Rice Leaf Diseases Identification Using Machine Learning Methods

Liu, Feng; Wu, Bangyu; Zhu, Susu; Wang, Junmin; Su, Zhen-Zhu; Liu, Fei; He, Yong

doi:10.3389/fpls.2020.577063

Cited by 62 publications

(33 citation statements)

References 45 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In Figures 5A–D present healthy and disease-stressed leaves of the four rice cultivars, corresponding to rice varieties 01, 02, 03, and 04, respectively. The change tendency of these four varieties was similar to the spectral profile of the other two varieties of rice in the previous study (Feng et al, 2020 ).…”

Section: Resultssupporting

confidence: 85%

Hyperspectral Imaging Combined With Deep Transfer Learning for Rice Disease Detection

Liu

et al. 2021

Front. Plant Sci.

Self Cite

View full text Add to dashboard Cite

Various rice diseases threaten the growth of rice. It is of great importance to achieve the rapid and accurate detection of rice diseases for precise disease prevention and control. Hyperspectral imaging (HSI) was performed to detect rice leaf diseases in four different varieties of rice. Considering that it costs much time and energy to develop a classifier for each variety of rice, deep transfer learning was firstly introduced to rice disease detection across different rice varieties. Three deep transfer learning methods were adapted for 12 transfer tasks, namely, fine-tuning, deep CORrelation ALignment (CORAL), and deep domain confusion (DDC). A self-designed convolutional neural network (CNN) was set as the basic network of the deep transfer learning methods. Fine-tuning achieved the best transferable performance with an accuracy of over 88% for the test set of the target domain in the majority of transfer tasks. Deep CORAL obtained an accuracy of over 80% in four of all the transfer tasks, which was superior to that of DDC. A multi-task transfer strategy has been explored with good results, indicating the potential of both pair-wise, and multi-task transfers. A saliency map was used for the visualization of the key wavelength range captured by CNN with and without transfer learning. The results indicated that the wavelength range with and without transfer learning was overlapped to some extent. Overall, the results suggested that deep transfer learning methods could perform rice disease detection across different rice varieties. Hyperspectral imaging, in combination with the deep transfer learning method, is a promising possibility for the efficient and cost-saving field detection of rice diseases among different rice varieties.

show abstract

Section: Resultssupporting

confidence: 85%

Hyperspectral Imaging Combined With Deep Transfer Learning for Rice Disease Detection

Liu

et al. 2021

Front. Plant Sci.

Self Cite

View full text Add to dashboard Cite

show abstract

“…This approach utilizes a system much like a multilayer perceptron that has been designed for reduced processing requirements. CNN consists of an output layer, a hidden layer, multiple convolutional layers, pooling layers, fully connected layers, and normalization layers to automatically extract abstracted shallow and deep features of the input [50].…”

Section: H Convolutional Neural Network (Cnn)mentioning

confidence: 99%

Discovering the Ganoderma Boninense Detection Methods Using Machine Learning: A Review of Manual, Laboratory, and Remote Approaches

et al. 2021

View full text Add to dashboard Cite

Ganoderma disease is a kind of infection that actuates oil palm death. Early detection of Ganoderma disease is the most recommended strategy for proper treatment and disease control plan to be taken promptly. In this paper, the detection methods for Ganoderma disease were reviewed and categorized accordingly. It was found that the combination of remote sensors and machine learning techniques could identify the disease up to four severity levels, including the early stage of infection. It also significantly reduced the labor and time costs compared to the traditional visual inspection and lab-based approaches. In terms of machine learning, support vector machine (SVM) using the idea of finding a hyperplane was suggested as the best classifier in several studies. Despite only one research was done on ANN and no research evaluating CNN and GAN in Ganoderma disease detection; ANN, CNN and GAN were recognized as the potential machine learning techniques that could enhance the detection system.INDEX TERMS Basal stem rot, Ganoderma, machine learning, oil palm, remote sensors.

show abstract

“…The hyperspectral sensing approach has been used for the detection of more than 12 fungal, bacterial, and nematode diseases in 15 field crops [12]. In rice, sheath blight, rice blast and bacterial leaf blight can be identified with an accuracy of more than 93% using hyperspectral imaging data through machine leaning methods [44]. Most recently, Lin et al [45] analyzed and compared the spectral responses to rice leaf and sheath tissue infected with sheath blight with healthy tissue and found that the hyperspectral sensing approach performed very well on the identification of sheath blight with an accuracy of more than 95%.…”

Section: Remote Sensorsmentioning

confidence: 99%

UAV Remote Sensing: An Innovative Tool for Detection and Management of Rice Diseases

Zhou¹,

Zhang²,

Lin³

2021

Diagnostics of Plant Diseases

View full text Add to dashboard Cite

Unmanned aerial vehicle (UAV) remote sensing is a new alternative to traditional diagnosis and detection of rice diseases by visual symptoms, providing quick, accurate and large coverage disease detection. UAV remote sensing offers an unprecedented spectral, spatial, and temporal resolution that can distinguish diseased plant tissue from healthy tissue based on the characteristics of disease symptoms. Research has been conducted on using RGB sensor, multispectral sensor, and hyperspectral sensor for successful detection and quantification of sheath blight (Rhizoctonia solani), using multispectral sensor to accurately detect narrow brown leaf spot (Cercospora janseana), and using infrared thermal sensor for detecting the occurrence of rice blast (Magnaporthe oryzae). UAV can also be used for aerial application, and UAV spraying has become a new means for control of rice sheath blight and other crop diseases in many countries, especially China and Japan. UAV spraying can operate at low altitudes and various speeds, making it suitable for situations where arial and ground applications are unavailable or infeasible and where precision applications are needed. Along with advances in digitalization and artificial intelligence for precision application across fertilizer, pest and crop management needs, this UAV technology will become a core tool in a farmer’s precision equipment mix in the future.

show abstract

Investigation on Data Fusion of Multisource Spectral Data for Rice Leaf Diseases Identification Using Machine Learning Methods

Cited by 62 publications

References 45 publications

Hyperspectral Imaging Combined With Deep Transfer Learning for Rice Disease Detection

Hyperspectral Imaging Combined With Deep Transfer Learning for Rice Disease Detection

Discovering the Ganoderma Boninense Detection Methods Using Machine Learning: A Review of Manual, Laboratory, and Remote Approaches

UAV Remote Sensing: An Innovative Tool for Detection and Management of Rice Diseases

Contact Info

Product

Resources

About