Development of a chickpea disease detection and classification model using deep learning
Ethiopia is the largest producer of chickpeas in Africa. Crop production and yield in Ethiopia is greatly affected by plant diseases which cause loss of agricultural products every year. One of these plant diseases is chickpea disease which is a fungal disease. Ascochyta blight and Fusarium wilt are...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2022-01-01
|
| Series: | Informatics in Medicine Unlocked |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2352914822001150 |
| _version_ | 1818474952060305408 |
|---|---|
| author | Abebech Jenber Belay Ayodeji Olalekan Salau Minale Ashagrie Melaku Bitew Haile |
| author_facet | Abebech Jenber Belay Ayodeji Olalekan Salau Minale Ashagrie Melaku Bitew Haile |
| author_sort | Abebech Jenber Belay |
| collection | DOAJ |
| description | Ethiopia is the largest producer of chickpeas in Africa. Crop production and yield in Ethiopia is greatly affected by plant diseases which cause loss of agricultural products every year. One of these plant diseases is chickpea disease which is a fungal disease. Ascochyta blight and Fusarium wilt are the most common chickpea diseases in Ethiopia that affect crop production quality and quantity. The identification of these diseases requires experienced experts or systems. Although numerous methods have been previously adopted in literature, deep learning (DL) is suggested as an efficient approach for disease recognition and classification since it can automatically learn features from the input image. In this paper, a chickpea disease detection model was developed using deep learning techniques by combining Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) for feature extraction and Softmax for classification. To develop the proposed model, various image preprocessing stages such as image resizing, normalization, and noise filtering using a combination of Gaussian filter (GF) and Median filter (MF) were performed. To prevent the problem of overfitting, augmentation was applied, while to train and test the effectiveness of the developed model, 8391 images were used. From the acquired images, 80% of the dataset was used for training, 20% of the dataset was used for testing and out of the 80% training data, 20% was used for validation. The proposed CNN-LSTM performed well in identifying chickpea disease, with an accuracy of 92.55%. According to the study's findings, the proposed CNN-LSTM outperforms existing methods. |
| first_indexed | 2024-04-14T04:43:52Z |
| format | Article |
| id | doaj.art-ab5613648f9a4b39a513c9a797b3b302 |
| institution | Directory Open Access Journal |
| issn | 2352-9148 |
| language | English |
| last_indexed | 2024-04-14T04:43:52Z |
| publishDate | 2022-01-01 |
| publisher | Elsevier |
| record_format | Article |
| series | Informatics in Medicine Unlocked |
| spelling | doaj.art-ab5613648f9a4b39a513c9a797b3b3022022-12-22T02:11:32ZengElsevierInformatics in Medicine Unlocked2352-91482022-01-0131100970Development of a chickpea disease detection and classification model using deep learningAbebech Jenber Belay0Ayodeji Olalekan Salau1Minale Ashagrie2Melaku Bitew Haile3Department of Information Technology, College of Informatics, University of Gondar, EthiopiaDepartment of Electrical/Electronics and Computer Engineering, Afe Babalola University, Nigeria; Saveetha School of Engineering, Saveetha Institute of Medical and Technical Sciences, India; Corresponding author. Department of Electrical/Electronics and Computer Engineering, Afe Babalola University, Nigeria.Department of Information Technology, College of Informatics, University of Gondar, EthiopiaDepartment of Information Technology, College of Informatics, University of Gondar, EthiopiaEthiopia is the largest producer of chickpeas in Africa. Crop production and yield in Ethiopia is greatly affected by plant diseases which cause loss of agricultural products every year. One of these plant diseases is chickpea disease which is a fungal disease. Ascochyta blight and Fusarium wilt are the most common chickpea diseases in Ethiopia that affect crop production quality and quantity. The identification of these diseases requires experienced experts or systems. Although numerous methods have been previously adopted in literature, deep learning (DL) is suggested as an efficient approach for disease recognition and classification since it can automatically learn features from the input image. In this paper, a chickpea disease detection model was developed using deep learning techniques by combining Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) for feature extraction and Softmax for classification. To develop the proposed model, various image preprocessing stages such as image resizing, normalization, and noise filtering using a combination of Gaussian filter (GF) and Median filter (MF) were performed. To prevent the problem of overfitting, augmentation was applied, while to train and test the effectiveness of the developed model, 8391 images were used. From the acquired images, 80% of the dataset was used for training, 20% of the dataset was used for testing and out of the 80% training data, 20% was used for validation. The proposed CNN-LSTM performed well in identifying chickpea disease, with an accuracy of 92.55%. According to the study's findings, the proposed CNN-LSTM outperforms existing methods.http://www.sciencedirect.com/science/article/pii/S2352914822001150ChickpeaAscochyta blightFusarium wiltCNN-LSTMDeep learning |
| spellingShingle | Abebech Jenber Belay Ayodeji Olalekan Salau Minale Ashagrie Melaku Bitew Haile Development of a chickpea disease detection and classification model using deep learning Informatics in Medicine Unlocked Chickpea Ascochyta blight Fusarium wilt CNN-LSTM Deep learning |
| title | Development of a chickpea disease detection and classification model using deep learning |
| title_full | Development of a chickpea disease detection and classification model using deep learning |
| title_fullStr | Development of a chickpea disease detection and classification model using deep learning |
| title_full_unstemmed | Development of a chickpea disease detection and classification model using deep learning |
| title_short | Development of a chickpea disease detection and classification model using deep learning |
| title_sort | development of a chickpea disease detection and classification model using deep learning |
| topic | Chickpea Ascochyta blight Fusarium wilt CNN-LSTM Deep learning |
| url | http://www.sciencedirect.com/science/article/pii/S2352914822001150 |
| work_keys_str_mv | AT abebechjenberbelay developmentofachickpeadiseasedetectionandclassificationmodelusingdeeplearning AT ayodejiolalekansalau developmentofachickpeadiseasedetectionandclassificationmodelusingdeeplearning AT minaleashagrie developmentofachickpeadiseasedetectionandclassificationmodelusingdeeplearning AT melakubitewhaile developmentofachickpeadiseasedetectionandclassificationmodelusingdeeplearning |