Plant Disease Classification using Deep Neural Network

DOI : 10.17577/IJERTCONV10IS11071

Download Full-Text PDF Cite this Publication

Text Only Version

Plant Disease Classification using Deep Neural Network

Prashantha G R Deepa N

Deparment of Computer science and Engineering Department of Computer Science andEngineering Jain Institute of Technology Jain Institute of Technology

Davanagere,Karnataka,India Davanagere,Karnataka,India

Jahnavi V V Harshita N S

Department of Computer Science and Engineering Deparment of Computer science and Engineering Jain Institute of Technology Jain Institute of Technology

Davanagere,Karnataka,India Davanagere,Karnataka,India

Bhavana K

Department of Computer and Engineering Jain Institute of Technology Davanagere,Karnataka,India

AbstractThe identification and prevention of diseases of crop are essential for improving crop prediction. Plants play a vital role in survival of all organisms on earth. It is important to take measures in detecting and mitigate diseases in plants. This author includes deep cnn models to detect problems in plant samples on their leaf. CNN gives more accuracy than other algorithm, CNN have achieved accurate results in identifying the diseases. The key to increasing the quality and eliminating agricultural goods is to identify plant illnesses. Plant disease is a biggest challenge for farmers, which impacts on income and food security. This model is capable of recognizing plant diseases of healthy and unhealthy leaf images. Due to fundamental qualities such as extracting the features, Artificial Neural Networks had made a significant advance in the area of computer vision. When opposed to computer vision, achieve effective outcomes with data sets inside a short amount of time.

KeywordsMachine learning;Deep Learning;Convolutional neural network;

  1. INTRODUCTION

    Agriculture is the backbone in the advancement of the human civilizations in Indian. In recent times, agribusiness has taken on a whole new role. People are researching production methods. Increase product activity, use pesticides and reduce environmental impact are important. In the world of farming, the automatic detection of phytopathogens derived It is a significant step forward. Moreover, plant disease identification that is timely has a positive impact on agriculture output and quality. As a result, employing machine learning to recognize and identify plant diseases offers a rapid and effective way. Pictures of crop infections signs are used to detect diseases in plants as well as for research, education, and evaluation. Utilizing image processing and deep learning To acquire a fast and precise identification, use this procedure. Full examination is the main means of illness detection in emerging economies' remote regions Growers from remote regions may have to travel hundreds of miles seeing a specialist, that's both expensive and time consuming.With its message passing and

    reliability, computerized types of instruments for the diagnosing of plant pathogens increase yields and agriculturists. Works have presented There are numerous techniques of dealing with the makes citizens. Biological agents might be classified via automated based on a range different factors. Traditionally handicraft is the most preferred typically subject among these qualities. For accurately features extracted, preparation is required, such as computer vision, colour modification, and segmented. Classification methods might be employed after image retrieval. K-nearest neighbour (KNN), machine(svm), reptree, randomized forest (RF), naïve bayes (NB), regression analysis (LR), increased regulation, convolutional neural networks (ANNs), and Deep CNN are some of the most often used classifiers. In order to create good image processors for diagnosing plant diseases The capability of computer technologies to anticipate the better crop illnesses is how system evaluate their success. Profound approaches, especially CNNs, are most promising method of learning a most exclusionary and conclusive characteristics manually.

  2. LITERATURE SURVEY

    The use of proper methodologies to detect good and diseased leaf aids in plant management, seed loss reduction, and performance improvement. Various machine learning approaches for detecting plant illnesses are included in the leaf tissue.

    Merecelin et al. [1] presented a detailed research of detection of disease in plants (apple and tomato leaf) utilising CNN ideas. The model has been trained using a digital image collection comprising 3663 photos of cherry and tomato plant leaves, and it achieved accuracy of 87%.

    Eftekhar Hossain et al. [2] devised a technique for recognising leaf disease detection using the K-nearest neighbour predictor (KNN). The classification was carried out using the features taken from the photos of sick images. The method KNN classifier was used to classify illnesses often observed in plants, such as abiotic stresses, late blight,

    microbial spot, and septoria leaf, in the study. The efficiency of this approach was 96.76 percent.

    Ch Usha Kumari et al. [3] built a system that uses the K- Means clustered and Knn methods to compute various parameters such as difference, connection, efficiency, average, confidence interval, and variation. The study's main flaw was that it only looked at the reliability of four unique conditions, and the precision value was quite low.

    Robert G et al. [4] proposed a CNN-based system for detecting the type of tomato leaf spot. According to this study, the F-RCNN training set had a reliability of 95.75 percent, while the Transferring Teaching method had an accuracy of 80 percent. The automatic photograph grabbing method had a precision of 91.67 percent.

    Alternia leaf spot, Brown spot, Mosaic, Grey spot, and Rust are five common apples infections than drastically diminish apples yield. However, current study does not have a viable and rapid means of detecting apples disease to maintain this fruit industry's long-term viability. There are 2 kinds of computer vision algorithms, such as Ssds, DSSD, and R-SSD: This beginning stage is the pre-network structure, which is used as a fundamental features generator. It's an added piece it identifies with the help of non – self characteristics.

  3. METHODOLOGY

    The plant identification method passes through five stages, as indicated in the diagram above.

    Fig 2: Image Acquisition

    Image Preprocessing:

    It refers to actions taken to organize images prior to their usage in models training and testing. This covers downsizing, positioning, and colour adjustments, among other things. Pre – processing is being used to enrich visual information by suppressing unwanted distortions or enhancing certain visual qualities that are useful for future analysis. Input picture, resize picture, remove noise, recognition, and typology are the stages in image processing techniques.

    Fig 3: Image Preprocessing

    Image Segmentation:

    Through simplifying a photo's portrayal, method focuses on making it more understandable and easy. Because it is the foundation of extracting data, this element is a fundamental technique for image assessment. A range of techniques, such as k-means grouping, Otsu's algorithm, and the thresholding, can be used to distinguish images.. The knn classifier divides items or pictures into K clusters simply a set of characteristics. The items and their accompanying cluster are classified by reducing the sum of the squared of radii among them.

    Fig1: Plant disease classification using deep neural network

    Image Acquisition:

    Feature Extraction:

    Fig 4: Image Segmentation

    Photographs of leaf tissue are collected using electrnic channels such as cameras, cell devices, and during such a stage, you'll use other gadgets that are of sufficient size and design. Images were available immediately access. Its construction of an imaging net database is completely the responsibility of the programme systems engineer. Inside the final piece of the recognizing programme, the image information is critical for improving its classifier's performance.

    As a consequence, within that step, the characteristics from

    such a research area should be obtained. Those qualities are required in order to derive a photo's connotation. For move up, you can just use colour, height, or substance. The majority of scientists have recently indicated an enthusiasm in employing surface features for diagnose plant diseases. Many of the extracted characteristics ways that could be utilised for developing the software are really the gray-level co-occurrence matrix (GLCM), colour frequency technique, geographical severely worsened dependant grids, or segmentation techniques edges. The Texture – based test is a scientific method for categorising materials.

    Classification:

    Its categorizing process involves assessing if the digital image is acceptable or otherwise. When a sick image was detected, past examinations have classified this as one of several ailments. In categorise, a computerized method, also known as a predictor, must be written in Tms. In current history, scientists have used a range different classifications, including Support vector machines (SVM), artificial neural networks (ANN), back propagation neural networks (BPNN), Nave Bayes, and Decision tree classifier were examples of k-nearest neighbour (KNN) classifier. This SVM classifier has been demonstrated to be among the most extensively utilised categorization. Any prediction includes benefits and disadvantages, but Svm is a straightforward and dependable technique.

    Fig 5: Classification Image

  4. RESULT

    The text message called click below to choose picture for testing so that a user can understand to click the below button. It has a button called choose file which can be used to browse the images on the systems hard disk. the popup which appears when user clicks on 'predict' button. The popup window will be having number of input images to be selected, this action should be confirmed with a double click or an open button.

    Fig 6: Healthy Image

    The selected image from the system directory, there will be a button provided called as Analyze Image for analyzing the input image to detect the condition of the leaf

    Fig 7: Disease Image

    It displays the status of leaf therefore Healthy or Unhealthy, if it is unhealthy it displays the particular disease name.

  5. CONCLUSION

Many established techniques for detecting and recognizing phytopathogens using infected plant leaves known. However, there's is no accurate and expense commercial technology for diagnosing illnesses. If illnesses are still not adequately recognised, they have an impact on agricultural productivity and can cause long difficulties like climate change and sometimes even poverty. In our research, System to deetect diseases in plants, researchers evaluated normal and infected leaf photos as well as infected plant names. The applied deep- learning model exhibits greater prediction ability in terms of both efficiency and damage when comparison to other deep- learning approaches. A deep neural network developed for detecting and recognising plant leaf disease was able to accurately categorise and existing evidence – based for about all photos with few exceptions, achieving accuracy. The model's training process was substantially less than those of previous machine-learning methods. Furthermore, the framework is a deep convolutional neural network that has been tuned to restrict the amount of parameters and processes more than allowed, and can run on any device.

REFERENCES

[1] Loey, A. ElSawy, and M. Afify, Deep learning in plant diseases detection for agricultural crops, International Journal of Service Science, Management, Engineering, and Technology, vol. 11, no. 2, pp. 4158, 2020.

[2] Sembiring, Y. Away, F. Arnia, and R. Muharar, Development of concise convolutional neural network for tomato plant disease classi fication based on leaf images, Journal of Physics: Conference Series, vol. 1845, article 012009, 2021.

[3] Ü. Atila, M. Uçar, K. Akyol, and E. Uçar, Plant leaf disease classification using EfficientNet deep learning model, Ecological Informatics, vol. 61, article 101182, 2021.

[4] J. G. M. Esgario, R. A. Krohling, and J. A. Ventura, Deep learning for classification and severity estimation of coffee leaf biotic stress, Computers and Electronics in Agriculture, vol. 169, article 105162, 2020.

[5] V. V. Adit, C. V. Rubesh, S. S. Bharathi, G. Santhiya, and R. Anuradha, A Comparison of Deep Learning Algorithms for Plant Disease Classification, in Advances in Cybernetics, Cognition, and Machine Learning for Communication Technologies, Lecture Notes in Electrical Engineering, vol. 643, pp. 153161, Springer, Singapore, 2020.

[6] Peng Jiang , Yuehan Chen ,Bin Liu , Dongjian He , Chunquan Liang

, Real-Time Detection of Apple Leaf Diseases Using Deep Learning Approach Based on Improved Convolutional Neural Networks, ( Volume: 7 ), pp. 06 May 2019.

[7] Zulkifli Bin Husin, Abdul Hallis Bin Abdul Aziz, Ali Yeon Bin Md Shakaff Rohani Binti S Mohamed Farook, Feasibility Study on Plant Chili Disease Detection Using Image Processing Techniques, 2012 Third International Conference on Intelligent Systems Modelling and Simulation.

[8] Tolstikhin, I.; Houlsby, N.; Kolesnikov, A.; Beyer, L.; Zhai, X.; Unterthiner, T.; Yung, J.; Keysers, D.; Uszkoreit, J.; Lucic, M.; et al. Mlp-mixer: An all-mlp architecture for vision. arXiv 2021, arXiv:2105.01601.

[9] Hu, X.; Yang, K.; Fei, L.; Wang, K. Acnet: Attention based network to exploit complementary features for rgbd semantic segmentation. In Proceedings of the 2019 IEEE International Conference on Image Processing (ICIP), Taipei, Taiwan, 2225 September 2019; pp. 1440

1444. 58.

[10] Dosovitskiy, A.; Beyer, L.; Kolesnikov, A.; Weissenborn, D.; Zhai, X.; Unterthiner, T.; Dehghani, M.; Minderer, M.; Heigold, G.; Gelly, S.; et al. An image is worth 16×16 words: Transformers for image recognition at scale. arXiv 2020, arXiv:2010.11929. 59.

[11] Kwon, H.; Yoon, H.; Park, K.W. Acoustic-decoy: Detection of adversarial examples through audio modification on speech recognition system. Neurocomputing 2020, 417, 357370. [CrossRef]

56.

[12] Kwon, H.; Yoon, H.; Choi, D. Restricted evasion attack: Generation of restricted-area adversarial example. IEEE Access 2019, 7, 60908

60919. [CrossRef] 57.

[13] Zhao, Y.; Chen, Z.; Gao, X.; Song, W.; Xiong, Q.; Hu, J.; Zhang, Z. Plant Disease Detection using Generated Leaves Based on DoubleGAN. IEEE/ACM Trans. Comput. Biol. Bioinform. 2021. [CrossRef] [PubMed] 47.

[14] S. R. G. Reddy, G. P. S. Varma, and R. L. Davuluri, Optimized convolutional neural network model for plant species identi fi – cation from leaf images using computer vision, International Journal of Speech Technology, 2021.

[15] A. Sembiring, Y. Away, F. Arnia, and R. Muharar, Development of concise convolutional neural network for tomato plant disease classi fication based on leaf images Journal of Physics: Conference Series, vol. 1845, article 012009, 2021.

[16] Ch. Usha Kumari, S. Jeevan Prasad, G. Mounika, Leaf Disease Detection: Feature Extraction with K-means clustering and Classification with ANN Proceedings of the 3rd International Conference on Computing Methodologies and Communication (ICCMC), Erode, India, 2019, pp 1095-1098 4.

[17] Mercelin Francis, C. Deisy, Disease Detection and Classification in Agricultural Plants Using Convolutional Neural Networks A Visual Understanding, Proceeding of the 6 th International Conference on ignal processing and Integrated network, Noida, India, 2019, pp 1063-1068. 5. Jiayue Zhao,

[18] Jianhua Qu, A Detection Method for Tomato Fruit Common Physiological Diseases Based on YOLOv2 Proceeding of the 10th International Conference on Information Technology in Medicine and Education, Qingdao, China, China, 2019 6.

[19] Y. Guo, J. Zhang, C. Yin et al., Plant disease identification based on deep learning algorithm in smart farming, Discrete Dynamics in Nature and Society, vol. 2020, Article ID 2479172, 11 pages, 2020.

[20] Ü. Atila, M. Uçar, K. Akyol, and E. Uçar, Plant leaf disease classification using EfficientNet deep learning model, Ecological Informatics, vol. 61, article 101182, 2021. [33] G.

[21] M. Loey, A. ElSawy, and M. Afify, Deep learning in plant diseases detection for agricultural crops, International Journal of Service Science, Management, Engineering, and Technology, vol. 11, no. 2, pp. 4158, 2020.

[22] Azizkhan F Pathan and Chetana Prakash, Attention-based position- aware framework for aspect-based opinion mining using bidirectional long short-term memory,Journal of King Saud University-Computer and Information Sciences,2021,ISSN 1319

1578,https://doi.org/10.1016/j.jksuci.2021.09.011.