0% found this document useful (0 votes)
69 views6 pages

Genetic Algorithms To Predict Problems in Crops

Genetic algorithms can be used to predict problems in crops by analyzing images of plant leaves. The document discusses several studies that used techniques like artificial intelligence, pattern recognition, and classification algorithms to automatically detect diseases in agricultural crops. These studies extracted features from images of leaves and then classified the diseases using algorithms like support vector machines, neural networks, naive Bayes, and others. Overall, the studies showed that genetic algorithms and machine learning techniques can successfully identify diseases in crops with 80-95% accuracy.

Uploaded by

raghuramu23456
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
69 views6 pages

Genetic Algorithms To Predict Problems in Crops

Genetic algorithms can be used to predict problems in crops by analyzing images of plant leaves. The document discusses several studies that used techniques like artificial intelligence, pattern recognition, and classification algorithms to automatically detect diseases in agricultural crops. These studies extracted features from images of leaves and then classified the diseases using algorithms like support vector machines, neural networks, naive Bayes, and others. Overall, the studies showed that genetic algorithms and machine learning techniques can successfully identify diseases in crops with 80-95% accuracy.

Uploaded by

raghuramu23456
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 6

Genetic algorithms to predict problems in crops

Abstract. There are methods for determining the diseases of any plant, such as taking
samples of vegetative tissue to a specialized laboratory or taking an expert agronomist
to the site of the crop. In either method, the disadvantage lies in the time needed to
obtain the results. For this reason, the use of artificial vision and pattern recognition
techniques have been considered, as well as some classification algorithms that
automatically determine the possible disease, facilitating the task of specialists to
develop their work and to find a timely diagnosis for treatment. This paper presents
several studies on the possibility of detecting diseases in agricultural crops, using
artificial intelligence techniques.

1. Introduction
It is well known that the prevention and timely diagnosis of any disease will bring the strategic
advantage over such condition and in agriculture is no exception, because knowing what ails a
crop or plant, increases the chances of success in treatment. In the developing world, more than
80% of agricultural production is generated by small farmers [1], and reports of yield losses of
more than 50% due to pests and diseases are common [2]. In addition, the largest proportion of
people with poverty and hunger problems (50%) live in these productive areas [3], making
small farmers a particularly vulnerable group to interruptions in the food supply caused by
pathogens. And as it is stated by [4], tools for the automatic recognition of plant diseases have
the potential to become a valuable source of information to assist decision-making in
agriculture. In the following section, the techniques used are shown, as well as the most
significant studies that address the challenge of detecting a disease, based on the analysis of the
characteristics present in the leaves of the crops.

2. Techniques and algorithms


In order to obtain a classification of diseases, the authors rely on proven methodologies shown
in Figure 1. Data acquisition, pre-processing, feature extraction and recognition are the steps or
procedures followed to obtain assertive results [5].
For the acquisition of the images, digital cameras are used to capture the leaves or the parts
where the damage caused by the disease is visible, as well as sets of images available on the
web and that have been made available to the whole public serving as a basis for the training of
the model [6]. For the images acquired by camera, the shots were taken in controlled
environments with acceptable resolutions, although it is important to mention that Bhange [6]
used a mobile phone to acquire the images and do an alternate experiment to see how much the
quality of the images affected [7].
Once the images are obtained, they are pre-processed and treated for scaling, noise removal,
color space transformation, histogram equalization and anything else that can be done to
maximize the features. By the time it goes to the next process, the image is cleaner, therefore,
when applying the segmentation techniques, they will separate the points of interest with better
precision, obtaining valuable data that will be more descriptive of the disease. Once the
characteristics are obtained, the next step is to classify them, using the algorithms, obtaining the
results and describing them [8]. It is in this step when the effectiveness of the procedure is
shown, determining if it has been classified according to what is expected.

Figure 1. Generalized methodology used in these articles.

Below are the algorithms with their particularities to be chosen as classifiers in these studies.

Table 1. Algorithms and their characteristics.

Algorithm Properties
Fuzzy logic It is based on heuristic rules, and used for highly non- linear
processes [9]. Easy implementation.
SVM Searches for a hyperplane that works as a separator. In
training and typical problems, it is very efficient [10].
It is very efficient where this type of environment of supervised
Bayes learning is used. No large amount of data is required for training
[11].
Searches for the closest observations to the one you're trying to
KNN predict and classify the point of interest based on most training
data [12].
ANN The classification is very efficient, at the cost of a
computationally expensive training [13].
The performance of convolutional neural networks in
object recognition and image classification has made great
CNN progress in recent years. They tend to be more accurate at the
expense of high computational cost. Training may require a
considerable number of images to produce reliable results [14].

In the following section, significant studies are shown, in relation to the detection of diseases in
several types of plants, using various automatic learning algorithms, but always following the
same methodology.

3. Literature Review
In [2], they used a fuzzy logic system and decision trees and with the help of a human expert,
they were able to make a recognition of the coffee diseases, where they obtained the
characteristics of the symptoms that are presented in the plant and thus be able to make their
decision tree. The results they obtained are within 85% accuracy. Although this research does
not deal with artificial vision techniques, it can be noted that a very important step has been
taken for the detection of diseases, using expert systems applying decision trees with fuzzy
logic.
In [5], they propose a web tool, where the farmers who grow the pomegranate will upload an
image of the fruit to be analyzed and with the trained model, they can check whether or not the
fruit is infected. The technique is based on extracting the characteristics of the images of the
pomegranate, such as color, morphology and color coherence vector and then make the
classification applying the SVM algorithm giving them an accuracy of 85% with a 10-
megapixel camera. The authors, knowing that the farmers will not always have the optimal
means of capture to upload the images to the system, conducted significant tests with the
cameras of the mobile devices with resolutions of 5 and 3 megapixels, and obtained results of
82% and 79% respectively.
In [7], they used an ANN, KNN, Naïve Bayes, a hybrid of self-organizing maps (SOM) and
radial base function (RBF), in order to determine the best algorithm to classify the rust, coffee
wilt (CWD) and CFD diseases affecting the coffee fruit. In their study, they explain that they
obtained 58.16% for the KNN, while for Naive Bayes they obtained 53.47%, for ANN 79.04%,
and for the combination of RBF and SOM they obtained 90.07% of accuracy, which shows a
great improvement with respect to the previous algorithms, although they note that the last one
takes more time in the training.
In [8], they tested for identifying diseases in the leaves. The studied crops were bananas, beans,
lemons and roses. After processing the images, they proposed to do the segmentation using
genetic algorithms, and later performing the grouping. For the extraction of characteristics, they
used the color concurrence method, since they consider better to use the color image than the
traditional gray scale. In order to make the classification, they used MDC with K-Mean
obtaining 86.54 %, MDC with an algorithm proposed by them, obtaining an improvement of
93.63% and SVM with proposed algorithm obtaining a significant improvement of 95.71%. All
these percentages describing a general average of the four studied crops.
In [10], they propose algorithms to identify multiple plant diseases, based on color analysis and
using a paired classification algorithm. According to them, their methodology allows them to
operate under uncontrolled conditions and thus be able to cover a large number of diseases. This
method was tested with a large set, without restrictions of the images of the leaves containing
the symptoms belonging to 74 diseases, 4 pests and 4 abiotic disorders, affecting 12 different
plant species. The results obtained were between 40% and 80% of accuracy.
The authors in [11] report that they found a feasible solution for the diagnosis and identification
of four alfalfa diseases. They extracted 129 texture, color and shape characteristics from the
1651 images using the ReliefF, 1R and CFS methods. To classify the diseases, they used SVM,
KNN and Random Forest. They found that the best classifier was SVM and the RelifF method
for obtaining characteristics, as they achieved 97.64% accuracy for the training set and 94.74%
for the test set.
12] They compared SVM and ANN algorithms to classify diseases in various crops that have
been attacked by fungi, bacteria, nematodes and nutrient deficiency. They mention that plant
disease symptoms exhibit different properties such as color, shape and texture and based on this,
the characteristics are obtained. They consider color as an important dimension, but applying
dimensionality reduction they discover by experimentation that, out of 24 only 8 characteristics
are significant for the classification of diseases. At the end, they test both algorithms and find
that with SVM they obtain 92.17 % of precision and that ANN only provides 87.48%, so they
demonstrate that, for this case, SVM is a better classifier.
In [12], they try to prove that it is not necessary to use a neural network or complex algorithm to
verify whether or not there is a characteristic disease in the strawberry leaves. For the authors, it
is better to use fuzzy logic algorithms for classification than any other method that consumes
more computational resources such as neural networks. The results they obtained are
encouraging, as they achieved 97% optimization in the segmentation and classification of
diseases with a processing time of 1.2 seconds. With this type of technique, they prove that it is
feasible to be implanted in an integrated circuit, adapted to a robot with artificial vision and that
could make an automated inspection in a modern botanical center.
The research in [13] conducted 60 experiments using deep convolutional neural networks to
identify 14 crop species and 26 diseases, using trained models such as AlexNet [17] and
GoogleNet [18]. They use the approach of [19], which demonstrates that supervised end-to-end
training with a CNN architecture is a possible option for a large number of classes, overcoming
the traditional approach of using hand-designed features. Within the PlantVillage dataset of
54,306 images containing 38 classes of 14 crop species and 26 diseases (or their absence), this
objective has been achieved, as demonstrated by the maximum accuracy of 99.35%. Therefore,
without any kind of feature engineering, the model correctly classifies the crops and diseases of
38 possible classes in 993 out of 1000 images. The authors noted that the training requires a lot
of computer work but, for the classification, it is less than a second, so they believe it could be
implemented on a cell phone.
Table 2 shows the summary of the results obtained by the above-mentioned investigations.

4. Discussion and results


One of the advantages that can be observe when [5] used the fuzzy logic classification
algorithm, is how easy and fast it could be implemented, because it does not require a large
collection of data for its training, and it presented more or less favorable results at the time of
classifying. Now, if the target class is reduced, it would have better results, as shown by the
study of [12], who achieved better accuracy just by predicting a single disease.

Table 2. Summary of the algorithms and their results

Reference. Cultivation Classifier Accuracy in the


Results (%)
[6] Coffee Fuzzy logic 84.55%
[7] Granada SVM 83.22%
KNN 59.14%
[8] Coffee ANN 80.14%
N. BAYES 54.52%
RBF and SOM 92.14%
Banana, Bean, Lemon, Rose MDC + K-Means 87.65%
[9] MDC + proposal SVM 94.95%
+ proposal 94.24%
Bean 51.24%
Sour Cassava 48.01%
Coconut tree Proposal Pair Sorting 54.02%
[10] Coffee Corn 72.35%
Cotton 54.74%
Grape 42.85%
77.24%
59.65%
Passion fruit 57.24%
Soya 57.36%
Sugar Cane Wheat 60.47%
72.95%
[11] Alfalfa SVM 95.65%
Wheat, sunflower, grape, ANN SVM 84.17%
[12] corn, cucumber, cotton, 93.62%
cabbage, tomato

Reference. Cultivation Classifier Accuracy in the


Results (%)
[13] Strawberry Fuzzy logic 98.02%
[14] Miscellaneous CNN 98.32%
[15] Tomato CNN 98.14%
Bean 96.00
Coconut Orange Cassava 84.03%
Corn Coffee Cotton 66.04%
Cashew Grape 98.14%
Kale Passion fruit Soya 64.65%
bean 79.75%
[16] Sugar cane Wheat 100%
Total 85.00%
83.00%
100%
94.33%
77.14%
100%
69.41%
84.65%
[17] Tomato CNN 38.75%

The proposal made in [7] using RBF and SOM seems to be quite good, since they are the ones
that obtained better results, compared to KNN, ANN, Bayes, and this is because their approach
to obtain the characteristics by their texture and color, made them obtain better results, since the
choice of the latter is the one that works best for these cases. SVM algorithms are a good
alternative for the resolution of these cases, although, at the beginning, they were not designed
to solve this kind of problems, with the passage of time they have been adapted and have
presented very good results, as those of [11], [8] and [6].

5. Conclusion
As it could be observed in this review of papers, the works that obtained better results are
convolutional neuronal networks, since their use is increasingly biased for this type of problems,
and diagnoses are obtained closer to what a human expert would determine. The only problem
would be that the training is computationally expensive and a large amount of data is required to
do it. If the study data were not sufficient, the recommendation would be decision trees with
fuzzy logic, since it is the algorithm with better results after convolutional networks. Although it
is not very precise, as technologies advance, more optimal values can be reached at a lower
computational cost. Another important point noted is that, with the use of new technologies
such as automatic learning and pattern recognition, crop diseases can be detected and a timely
diagnosis can be obtained, decreasing the risk of agricultural and economic losses, which would
bring a direct benefit to the farmers who implement it.

References

[1] Partel, V., Kim, J., Costa, L., Pardalos, P., & Ampatzidis, Y. (2019). Smart Sprayer for
Precision Weed Control Using Artificial Intelligence: Comparison of Deep Learning
Frameworks.
[2] Arora, J., & Agrawal, U. (2020). Classification of Maize leaf diseases from healthy
leaves using Deep Forest. Journal of Artificial Intelligence and Systems, 2(1), 14-26.
[3] Sufian, A., Chandra, U., Das, S., & Chettri, D. (2020). Advancements in Agriculture
Strategies and Environmental Impact: A Review. Available at SSRN 3516438.
[4] Patrício, D. I., & Rieder, R. (2018). Computer vision and artificial intelligence in
precision agriculture for grain crops: A systematic review. Computers and electronics in
agriculture, 153, 69-81.
[5] Bannerjee, G., Sarkar, U., Das, S., & Ghosh, I. (2018). Artificial Intelligence in
Agriculture: A Literature Survey. International Journal of Scientific Research in
Computer Science Applications and Management Studies, 7(3), 1-6.
[6] Fuentes, A., Im, D. H., Yoon, S., & Park, D. S. (2017, June). Spectral analysis of CNN
for tomato disease identification. In International Conference on Artificial Intelligence
and Soft Computing (pp. 40-51). Springer, Cham.
[7] Sufian, A., Chandra, U., Das, S., & Chettri, D. (2020). Advancements in Agriculture
Strategies and Environmental Impact: A Review. Available at SSRN 3516438.
[8] Joshi, B. M., & Bhavsar, H. (2020). Plant leaf disease detection and control: A survey.
Journal of Information and Optimization Sciences, 41(2), 475-487.
[9] Vivek, S., & Jesma, V. A. (2019). The changing phase of agriculture: Artificial
intelligence. AgricINTERNATIONAL, 6(1), 33-38.
[10] Drury, B., Valverde-Rebaza, J., Moura, M. F., & de Andrade Lopes, A. (2017). A survey
of the applications of Bayesian networks in agriculture. Engineering Applications of
Artificial Intelligence, 65, 29-42.
[11] Joshi, B. M., & Bhavsar, H. (2020). Plant leaf disease detection and control: A survey.
Journal of Information and Optimization Sciences, 41(2), 475-487.
[12] Bharti, V. K., & Bhan, S. (2018). Impact of artificial intelligence for agricultural
sustainability. Journal of Soil and Water Conservation, 17(4), 393-399.
[13] Joseph, R. B., Lakshmi, M. B., Suresh, S., & Sunder, R. (2020, March). Innovative
Analysis of Precision Farming Techniques with Artificial Intelligence. In 2020 2nd
International Conference on Innovative Mechanisms for Industry Applications (ICIMIA)
(pp. 353-358). IEEE.
[14] Dharmaraj, V., & Vijayanand, C. (2018). Artificial Intelligence (AI) in Agriculture. Int. J.
Curr. Microbiol. App. Sci, 7(12), 2122-2128.
[15] Izquierdo, N.V., Lezama, O.B.P., Dorta, R.G., Viloria, A., Deras, I., Hernández-Fernández,
L.: Fuzzy logic applied to the performance evaluation. honduran coffee sector case. In:
Tan, Y., Shi, Y., Tang, Q. (eds.) Advances in Swarm Intelligence. ICSI 2018. Lecture
Notes in Computer Science, vol. 10942 (1), pp. 1–12. Springer, Cham (2018)
[16] Amelec, V.: Increased efficiency in a company of development of technological solutions
in the areas commercial and of consultancy. Adv. Sci. Lett. 21(5), 1406–1408 (2015)
[17] Eli-Chukwu, N. C. (2019). Applications of artificial intelligence in agriculture: A review.
Engineering, Technology & Applied Science Research, 9(4), 4377-4383.
[18] Chilwal, B., & Mishra, P. K. (2020). A Survey of Fuzzy Logic Inference System and Other
Computing Techniques for Agricultural Diseases. In International Conference on
Intelligent Computing and Smart Communication 2019 (pp. 1-6). Springer, Singapore.
[19] Ampatzidis, Y., & Partel, V. (2019). UAV-based high throughput phenotyping in specialty
crops utilizing artificial intelligence. In 2019 ASABE Annual International Meeting (p.
1). American Society of Agricultural and Biological Engineers.
[20] Viloria A. et al. (2020) Optimization of Flow Shop Scheduling Through a Hybrid Genetic
Algorithm for Manufacturing Companies. In: Pandian A., Ntalianis K., Palanisamy R.
(eds) Intelligent Computing, Information and Control Systems. ICICCS 2019. Advances
in Intelligent Systems and Computing, vol 1039. Springer, Cham

You might also like