Coffee Leaf Disease Recognition Based on Deep Learning and Texture Attributes

Sorte, Lucas Ximenes Boa; Ferraz, Carolina Toledo; Fambrini, Francisco; Goulart, Roseli dos Reis; Saito, José Hiroki

doi:10.1016/j.procs.2019.09.168

Cited by 43 publications

(27 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…From these above observations, the authors concluded that Deep learning-based CNN is better for the classification of different plant diseases. Tetila et al (2019) proven for coffee plants that Deep learning is useful in preprocessing and dimensionality reduction (Sorte et al, 2019). Barbedo (2013) also told Deep network has various models for the classification of complex traits.…”

Section: Techniques and Methods Dependent Data Analysis For Disease Detectionmentioning

confidence: 99%

Rice Crop Disease Prediction Using Machine Learning Technique

Patel

Sharaff

2021

International Journal of Agricultural and Environmental Information Systems

View full text Add to dashboard Cite

Crop yields are affected at large scale due to spread of unchecked diseases. The spread of these diseases is similar to the spreading of cancer in human body. But, unlike cancer these diseases can be identified at early stages through plant phenotyping traits analysis. In order to effectively identify these diseases, effective segmentation, feature extraction, feature selection and classification processes must be followed. Selection of the best combination for the given methods is very complex due to the presence of a large number of the aforementioned methods. Thereby disease prediction models are generally found to be ineffective. This paper proposes a highly effective machine learning-based formulation approach to select a proper classification process which improves the overall accuracy of crop disease detection with different dimensionality of plant dataset and included maximum features also. Hence, the proposed adaptive learning algorithm gives 99.2% accuracy compared to other techniques like Back-propagation Neural Network (BPNN), Convolutional Neural Network (CNN), and SVM.

show abstract

Section: Techniques and Methods Dependent Data Analysis For Disease Detectionmentioning

confidence: 99%

Rice Crop Disease Prediction Using Machine Learning Technique

Patel

Sharaff

2021

International Journal of Agricultural and Environmental Information Systems

View full text Add to dashboard Cite

show abstract

“…However, it is difficult when the infected region engages only a small segment of the image. Sorte et al [9] proposed a coffee leaf infection detection method depending on the deep learner and texture features. More texture features like spectral features need to be extracted and enhanced to enhance the performance of this recognition method.…”

Section: Literature Surveymentioning

confidence: 99%

Multi-Channel Convolutional Neural Network for Prediction of Leaf Disease and Soil Properties

2022

IJIES

View full text Add to dashboard Cite

Prediction of leaf disease and soil property helps farmers improve crop production quality through maintaining the soil property and taking proper actions for leaf disease. Various techniques have been developed to predict leaf disease and soil properties. Support Vector Machine (SVM) is one of the machine learning techniques that was used to predict leaf disease and soil properties. To predict the leaf disease and soil property, SVM processed the extracted features from leaf images and soil images. Deep learning can be used for prediction, which has the advantage of machine learning is that one does not need to be concerned about domain knowledge as no feature engineering is required in this, unlike SVM-based prediction. In addition to this, SVM-based prediction is not very effective for handling multiple inputs. A Convolutional Neural Network (CNN) is a deep learner which was applied for the prediction of leaf disease and soil property. Even though this method has better performance, the information from leaf and soil images is mixed together, which may affect the prediction accuracy. So, in this paper, a Multichannel CNN (MCNN) method is introduced in which individual channels are used for leaf and soil images. In MCNN, the feature learning using MCNN for leaf images is kept distinct from the soil image to avoid data fusion between the leaf and soil images. The features related to leaf and soil images are paired and transferred over the corresponding channels for the prediction of leaf disease and soil property. After the prediction of leaf disease and soil property, the correlation between leaf disease and soil property is identified using the Pearson correlation coefficient and it is sent to the farmers using mobile phones to improve their crop production. Finally, these methods are validated by using different leaf infections and soil images for 3 types of crops. The experimental results show that the MCNN method achieves an average accuracy of 87.77% for leaf disease prediction and 90.38% for soil property prediction compared to the classical methods.

show abstract

“…Many state-of-the-art methods already exist for plant disease classification and detection [ 23 , 24 , 25 , 26 , 27 ] and defect detection in general [ 28 , 29 , 30 ]. However, to the best of our knowledge, there are only a few studies on coffee disease detection and they focus only on classifying healthy and non-healthy leaves using transfer learning [ 31 ] or using an annotated bounding box for detection [ 32 ]. Nevertheless, these deep models are seen as “black box” approaches, and researchers face a lot of trial and error when developing a satisfactory CNN model.…”

Section: Introductionmentioning

confidence: 99%

Coffee Disease Visualization and Classification

Yebasse

Shimelis²,

Warku

et al. 2021

Plants

View full text Add to dashboard Cite

Deep learning architectures are widely used in state-of-the-art image classification tasks. Deep learning has enhanced the ability to automatically detect and classify plant diseases. However, in practice, disease classification problems are treated as black-box methods. Thus, it is difficult to trust the model that it truly identifies the region of the disease in the image; it may simply use unrelated surroundings for classification. Visualization techniques can help determine important areas for the model by highlighting the region responsible for the classification. In this study, we present a methodology for visualizing coffee diseases using different visualization approaches. Our goal is to visualize aspects of a coffee disease to obtain insight into what the model “sees” as it learns to classify healthy and non-healthy images. In addition, visualization helped us identify misclassifications and led us to propose a guided approach for coffee disease classification. The guided approach achieved a classification accuracy of 98% compared to the 77% of naïve approach on the Robusta coffee leaf image dataset. The visualization methods considered in this study were Grad-CAM, Grad-CAM++, and Score-CAM. We also provided a visual comparison of the visualization methods.

show abstract

Coffee Leaf Disease Recognition Based on Deep Learning and Texture Attributes

Cited by 43 publications

References 11 publications

Rice Crop Disease Prediction Using Machine Learning Technique

Rice Crop Disease Prediction Using Machine Learning Technique

Multi-Channel Convolutional Neural Network for Prediction of Leaf Disease and Soil Properties

Coffee Disease Visualization and Classification

Contact Info

Product

Resources

About