Machine learning techniques implementation for detection of grape leaf disease

被引:3
作者
Shantkumari, M. [1 ]
Uma, S. V. [2 ]
机构
[1] Visvesvaraya Technol Univ, Dept CSE, RRC, Belagavi, Karnataka, India
[2] R N Shetty Inst Technol, Dept ECE, Bangalore, Karnataka, India
关键词
Classification; IKNN model; Grape leaf disease; Histogram gradient feature; Feature extraction; Convolutional neural network (CNN); NEURAL-NETWORK; IDENTIFICATION; CLASSIFICATION;
D O I
10.1007/s11042-023-14441-x
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Grape leaf diseases like Black Rot, Eska measles, Leaf Spot and Healthy are among the most common disease types of the grape crop. Accurate detection of grape leaf diseases in the initial stages can control the disease spread significantly and guarantee progressive development of the grape crop industry. The existing research provides several complex image processing algorithms and cannot assure high classification accuracy. Therefore, machine learning techniques are presented in this article to enhance leaf disease classification accuracy for efficiently detecting grape leaf diseases. Moreover, two classification models are introduced in which the simple Convolutional Neural Network based Classification (CNNC) Model is detailed. Then the improvised K-Nearest Neighbour (IKNN) model for precisely detecting grape leaf diseases is detailed. Moreover, pixel encoding methods are presented to obtain a histogram representation of extracted features. Training of simple CNNC and the proposed IKNN model is conducted on Plant-Village Dataset. Additionally, mathematical modeling is presented to formulate the problem in the feature extraction process. Moreover, Confined Intensity Directional Order Relation (CIDOR) operation ensures low dimensionality of histogram representation in the multiscale domain. Furthermore, Global Pixel Order Relation (GPOR) focuses on setting up a communication with long-reach pixels of an image outside of the central pixel neighborhood. Compared to the simple CNNC, the proposed IKNN model outperforms all the traditional leaf disease classification algorithms in terms of classification accuracy. However, the IKNN model provides superior results than CNNC comparatively in terms of classification accuracy.
引用
收藏
页码:30709 / 30731
页数:23
相关论文
共 39 条
  • [1] Diagnosis and recognition of grape leaf diseases: An automated system based on a novel saliency approach and canonical correlation analysis based multiple features fusion
    Adeel, Alishba
    Khan, Muhammad Attique
    Sharif, Muhammad
    Azam, Faisal
    Shah, Jamal Hussain
    Umer, Tariq
    Wan, Shaohua
    [J]. SUSTAINABLE COMPUTING-INFORMATICS & SYSTEMS, 2019, 24
  • [2] Algergawy A., 2017, DEEP LEARNING BASED, P79
  • [3] [Anonymous], PLANT VILLAGE DATASE
  • [4] [Anonymous], 2000, REPORT EXPERT CONSUL, V2000
  • [5] Vineyard computing: Sensor networks in agricultural production
    Burrell, J
    Tim, B
    Beckwith, R
    [J]. IEEE PERVASIVE COMPUTING, 2004, 3 (01) : 38 - 45
  • [6] CHOLLET F, 2017, PROC CVPR IEEE, P1800, DOI DOI 10.1109/CVPR.2017.195
  • [7] Bacterial Foraging Optimization Based Radial Basis Function Neural Network (BRBFNN) for Identification and Classification of Plant Leaf Diseases: An Automatic Approach Towards Plant Pathology
    Chouhan, Siddharth Singh
    Kaul, Ajay
    Singh, Uday Pratap
    Jain, Sanjeev
    [J]. IEEE ACCESS, 2018, 6 : 8852 - 8863
  • [8] Unmanned aerial systems for photogrammetry and remote sensing: A review
    Colomina, I.
    Molina, P.
    [J]. ISPRS JOURNAL OF PHOTOGRAMMETRY AND REMOTE SENSING, 2014, 92 : 79 - 97
  • [9] Crop Leaf Disease Image Super-Resolution and Identification With Dual Attention and Topology Fusion Generative Adversarial Network
    Dai, Qiang
    Cheng, Xi
    Qiao, Yan
    Zhang, Youhua
    [J]. IEEE ACCESS, 2020, 8 : 55724 - 55735
  • [10] Combination of multiple classifiers for automatic recognition of diseases and damages on plant leaves
    El Massi, Ismail
    Es-saady, Youssef
    El Yassa, Mostafa
    Mammass, Driss
    [J]. SIGNAL IMAGE AND VIDEO PROCESSING, 2021, 15 (04) : 789 - 796