A Comprehensive Review on Crop Disease Classification using Machine Learning Technique

Abstract

With migrating populations towards cities and global urbanization, there arises a need to automate certain sections of agriculture. Precision agriculture has emerged as an active areas of research in which computational and data processing techniques are being applied in the domain of agriculture. One of the most common and menacing problems which farmers face is crop diseases which result in malnourishing the actual crop and in some cases completely destroying the crop. Hence, machine learning and deep learning based techniques are being extensively studied and explored to design systems which can identify diseases among plants, automatically. Crop disease identification is a relatively challenging task keeping in mind the similarity among the textures of plants and the variations among plants of the same category, blurring and noise effects appearing in images typically captured by unmanned aerial vehicles (UAVs). This paper presents a comprehensive review on image pre-processing and machine learning based techniques for automated classification of crops diseases. The fundamental mechanism of noise removal along with noise sources in images have been cited and explained. The basics of machine learning based classifiers applied to image classification have also been discussed. Salient features of existing techniques along with the research gap have been clearly highlighted. The research gaps identified in existing work allows future researchers in leveraging the limitations of existing approaches and devising novel methods. Finally the evaluation metrics to evaluate the performance of existing work have been presented for comparative performance evaluation.

Introduction

I. INTRODUCTION

The agriculture sector has been witnessing tremendous technological advancements with the rise in digital infrastructure. The use of high end technology driven methods have been rampant in this domain. The crop quality is a major aspect of crop harvesting. But there are many factors that lead to the diminishing quality of crops. Disease is one such problem that is commonly faced while cultivation of crops. Precise detection of these diseases can help in preventing damage to the crops and disease free cultivation [1]. With the help of machine learning, the challenges in the agricultural domain can be easily mitigated and solved. There happen to be certain areas where the traditional ways of farming methods don’t work effectively. Crop Diseases provide major hindrance towards successful and healthy cultivation of crops. Manual manpower also fails in disease detection when the quantity of the crop is huge. This is where the machine learning methods can greatly contribute in accurate and proper classification of crop and disease [2]-[3].

The concept of automatic detection and classification of diseases can play a beneficial role in crop disease management. Lesser the disease, higher will be the yield of crops. One of the major challenges that exist in detection of diseases is crops and diseases tend to have similar appearance, colour and texture. Henceforth accurate classifications of disease have to be carried out for the precise disease detection [4].

Machine learning based approaches provide a vast range of methods for accurate classification with high accuracy. With the advancement in technology, the concept of precision agriculture has gained quite a good prominence [5]. The disease needs to effectively classified and separated from the crop production for quality yield of crops. Machine learning methods have emerged as robust solution for crop and disease classification. It involves the use of artificial intelligence and machine learning methods for precise classification of crop and disease. Disease detection and management has to be done accurately for its detection. The classification methods that are used in machine learning provide very high accuracy [6]. Manual man power can be less cost effective and also less accurate when the amount of crop disease detection is huge. 

The use of supervised learning methods as a part of implementing artificial intelligence concepts can go a long way in quality crop cultivation with high yields. The disease classification can be facilitated accurately as the identification of different types of disease can be easily done using machine learning where human intervention can take a long time. The growth of diseases is rampant on farm lands and not quite easily recognizable and they feed on the healthy crops [7]. So, automated methods can revolutionize this aspect of crop disease detection and aid in precision agriculture with the use of efficient machine learning backed models and methods. It would entail high yield, cost effectiveness and quality produce [8].

There have been several models which have bene used to classify crops and diseases. A detailed analysis of contemporary methods along with the findings has been presented in table I.

TABLE I

NOTEWORTHY CONTRIBUTION IN THE DOMAIN

II. THEORITICAL BACKGROUND

Automated crop-disease classification has gained prominence with advanced image processing techniques coupled with machine learning. Automated detection is generally based on three major steps which are image processing and data preparation, feature extraction and final classification.  Each of these processes are discussed in this section:

A. Image Processing

Prior to computing important parameters or feature of the fundus image, which lays the foundation for the final classification, it is necessary to process the image for the following reasons:

1) Illumination Correction: In this part, the inconsistencies in the image illumination are corrected so as to make the image background uniform and homogenous. Illumination inconsistencies occur due to capturing the image from different angles under inconsistent lighting [9]. Inconsistencies in the illumination can be caused due to the position and orientation of the source, the non-homogeneity of wavelengths of the source, the nature of the surface such as smoothness, orientation and material characteristics and finally the characteristics of the sensing device such as resolution, capturing capability and sensitivity [10]. Typically, illumination correction is done based on the computation of the correlation co-efficient given by: 

Conclusion

This paper introduces the need for precision agriculture along with its applications in the domain of automated classification of crop and diseases. The working of automated classifiers along with their attributed and dependence on feature extraction has been explained in detail. Different stages of the image processing and segmentation have been enlisted. The significance of different image features and extraction techniques have been clearly mentioned with their utility and physical significance. Various machine learning based classifiers and their pros and cons have been highlighted. The mathematical formulations for the feature extraction and classification gave been furnished. A comparative analysis of the work and results obtained has been cited in this paper. It can be concluded that image enhancement and feature extraction are as important as the effectiveness of the automated classifier, hence appropriate data processing should be applied to attain high accuracy of classification.

References

[1] S.Haug J.Ostermann, “Plant classification system for crop /disease discrimination without segmentation”, IEEE Winter Conference on Applications of Computer Vision, IEEE 2014, pp. 1142-1149. [2] A. Zhang, X.Wei X. “A review on disease detection using ground-based machine vision and image processing techniques. Computers and Electronics in Agriculture”, Elsevier 2019, 158, pp: 226-240 [3] AM Anter, AE Hassenian, D Oliva, “An improved fast fuzzy c-means using crow search optimization algorithm for crop identification in agricultural”, Expert Systems with Applications, Elsevier 2019, vol.118, pp.340-354. [4] A.Bakhshipour, A.Jafari A, S.Nassiri. D.Zare, “Disease segmentation using texture features extracted from wavelet sub-images”, Biosystems Engineering, Elsevier 2017, vol.157, pp.1-12. [5] J Torres-Sospedra, P Nebot, “Two-stage procedure based on smoothed ensembles of neural networks applied to disease detection in orange groves”, Biosystems Engineering, Elsevier 2014, vol.123, pp.40-55. [6] I Dasgupta, J Saha, P Venkatasubbu, “AI crop predictor and disease detector using wireless technologies: a smart application for farmers”, Arabian Journal for Science and Engineering, Springer 2020, vol.45, pp.11115– 11127. [7] U. Shruthi, V. Nagaveni and B. K. Raghavendra, \"A Review on Machine Learning Classification Techniques for Plant Disease Detection,\" 2019 5th International Conference on Advanced Computing & Communication Systems (ICACCS), 2019, pp. 281-284. [8] Y Chen, Z Wu, B Zhao, C Fan, S Shi, “Disease and corn seedling detection in field based on multi feature fusion and support vector machine” Sensors, MDP1 2021, vol.21, no.12, pp.1-18. [9] X. Jin, J. Che and Y. Chen, \"Disease Identification Using Deep Learning and Image Processing in Vegetable Plantation,\" in IEEE Access 2021, vol. 9, pp. 10940-10950. [10] S Abouzahir, M Sadik, E Sabir, “Bag-of-visual-words-augmented Histogram of Oriented Gradients for efficient disease detection”, Biosystems Engineering, Elsevier 2021, vol.202, pp. 179-194. [11] V. N. Thanh Le, G. Truong and K. Alameh, \"Detecting diseases from crops under complex field environments based on Faster RCNN,\" 2020 IEEE Eighth International Conference on Communications and Electronics (ICCE), 2021, pp. 350-355. [12] M. Alam, M. S. Alam, M. Roman, M. Tufail, M. U. Khan and M. T. Khan, \"Real-Time Machine-Learning Based Crop/Disease Detection and Classification for Variable-Rate Spraying in Precision Agriculture,\" 2020 7th International Conference on Electrical and Electronics Engineering (ICEEE), 2020, pp. 273-280. [13] S Sabzi, Y Abbaspour-Gilandeh, “Using video processing to classify potato plant and three types of disease using hybrid of artificial neural network and partincle swarm algorithm”, Measurement, Elsevier 218, vol.126, pp.Pages 22-36. [14] Tin-Yau Kwok and Dit-Yan Yeung, \"Constructive algorithms for structure learning in feedforward neural networks for regression problems,\" in IEEE Transactions on Neural Networks, vol. 8, no. 3, pp. 630-645. [15] F.Ahmed, Al-Mamun, A.Bari, E.Hossain, “Classification of crops and diseases from digital images. A support vector machine approach Journal of Crop Production”, Elsevier 2013, vol.40, pp: 98-104. [16] S. S. Athani and C. Tejeshwar, \"Support Vector Machine-Based Classification Scheme of Maize Crop,\" 2017 IEEE 7th International Advance Computing Conference (IACC), 2017, pp. 84-88. [17] P. Lottes, J. Behley, N. Chebrolu, A. Milioto and C. Stachniss, \"Joint Stem Detection and Crop-Disease Classification for Plant-Specific Treatment in Precision Farming,\" 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2018, pp. 8233-8238. [18] S. Haug, A. Michaels, P. Biber and J. Ostermann, \"Plant classification system for crop /disease discrimination without segmentation,\" IEEE Winter Conference on Applications of Computer Vision, 2014, pp. 1142-1149.

Copyright

Copyright © 2022 Priyanka Bande, Mr. Kranti Dewangan . This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.