Classification of Leaf Disease in Potato Plants Based on Color and

Texture Features Using the K Nearest Neighbor (KNN) Method

Andi Sri Bulan

and Ingrid Nurtanio

Department of Informatics Hasanuddin University Makassar, Indonesia

Keywords: Image Processing, Image Classification, KNN, Color and Texture Feature Extraction, Potato Leaf Disease.

Abstract: Potato (Solanum tuberosum) is an essential food commodity after rice and wheat, widely cultivated in

Indonesia’s highlands due to its high economic value. However, production is frequently threatened by leaf

diseases such as early blight (Alternaria Solani), late blight (Phytophthora infestans), leaf spot, mosaic virus

(PVY), leaf roll, blackleg, soft rot, fusarium wilt, bacterial wilt, and rhizoctonia canker. These diseases reduce

yield and quality, leading to economic losses. This study proposes a potato leaf disease classification system

by combining color and texture features with Principal Component Analysis (PCA) for dimensionality

reduction and K-Nearest Neighbor (KNN) as the classifier. The dataset consisted of 1,055 leaf images

collected from Malino, South Sulawesi, expanded through augmentation to 1,155 images across 12 classes

(healthy + 11 diseases). Preprocessing included resizing, color space conversion, and segmentation. The

optimized KNN model with PCA (k=1, PCA=14) achieved 97.78% accuracy on test data, outperforming

Support Vector Machine (SVM) and Random Forest. These results confirm that simple handcrafted features

with lightweight classifiers can achieve competitive performance compared to deep learning approaches,

making them suitable for mobile-based agricultural applications in resource-limited environments to support

early disease detection.

1 INTRODUCTION

Potatoes (Solanum tuberosum) are the third most

important food commodity after rice and wheat,

widely cultivated in the highlands of Indonesia (Y.

Alkhalifi et al., 2021). Despite their high economic

value, potato productivity often decreases due to

various leaf diseases that significantly reduce crop

quality and quantity (L.C.Ngugi et al., 2021).

Common potato leaf diseases include early blight

(Alternaria Solani), late blight (Phytophthora

infestans), leaf spot, mosaic virus (PVY), leaf roll,

blackleg, soft rot, Fusarium wilt, bacterial wilt, and

rhizoctonia canker (C. Hou et al., 2021) (S. Song et

al., 2024). These diseases often exhibit similar

symptoms, such as necrotic spots, discoloration, leaf

rolling, and wilting (S. Sattar et al., 2024), making

manual identification difficult, particularly for

farmers without specialized expertise (H. A Santoso

et al., 2024).

https://orcid.org/0009-0006-3365-1848

https://orcid.org/0000-0003-2026-1809

Potato (Solanum tuberosum L.), the fourth most

important food crop in the world, is affected by

several viral pathogens with potato virus Y (PVY)

having the greatest economic impact. At least nine

biologically distinct variants of PVY are known to

infect potato. These include the relatively new

recombinant types named PVY-NTN and PVYN-Wi,

which induce tuber necrosis in susceptible cultivars.

To date, the molecular plant-virus interactions

underlying this pathogenicity have not been fully

characterized. We hypothesized that this necrotic

behavior is supported by transcriptional and

functional signatures that are unique to PVY-NTN

and PVYN-Wi (Richard Manasseh et al., 2024).

Potato Y virus (PVY, genus Potyvirus), which has

RNA as genetic material, is generally detected among

potatoes cultivated around the world. The virus is

differentiated into strains O, C, and Z according to

specific hypersensitive responses to three potato

genes: Ny, Nc, and Nz, respectively. PVY has been

associated as one of the most important plant viruses

Bulan, A. S. and Nurtanio, I.

Classiﬁcation of Leaf Disease in Potato Plants Based on Color and Texture Features Using the K Nearest Neighbor (KNN) Method.

DOI: 10.5220/0014272800004928

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 1st International Conference on Research and Innovations in Information and Engineering Technology (RITECH 2025), pages 61-67

ISBN: 978-989-758-784-9

due to its wide range of hosts, distribution, and large

economic impact. Recent estimates estimate losses of

EUR 187 million per year due to PVY infection in

potatoes in Europe (Adyatma Irawan Santosa et al.,

2025).

Manual identification by agricultural experts also

has limitations, such as subjectivity, time constraints,

and reliance on individual experience an automated

system based on leaf imagery is required to identify

diseases quickly, accurately, and consistently (S.

Qian et al., 2021). Several previous studies have

developed plant disease detection systems using

image-based approaches with deep learning methods

such as CNNs and transfer learning (I. Harfian et al.,

2020).

Although these approaches provide high accuracy

(A.J. Rozaki et al., 2020), they generally require large

datasets and significant computational resources (B.

Rahmat et al., 2022), which makes them less suitable

for implementation on devices with limited capacity,

such as mobile applications in agricultural settings.

This paper proposes an alternative approach using a

combination of simple but representative features,

namely color features (mean RGB) and texture

features (GLCM) (H. Habaragamuwa et al., 2021).

and, dimensionality reduction with PCA, and

classification using the K-Nearest Neighbor (KNN)

algorithm. This approach has several advantages:

computational efficiency (F. W. Siddhi et al., 2022).

Potato crops and their salability are influenced by

potato pests in that both crop yield and quality are

reduced. This in turn reduces the income for potato

farmers due to lower prices for the crop, lower crop

yield, trade restriction and reduced market access.

Agricultural viability over the long run therefore

depends on sustainable pest management. In order to

efficiently detect potato pests, a dataset was

constructed which contains eight prevalent potato

species that were taken from several sources. Image

pre-processing techniques were employed enhance

image quality for compatibility with deep learning

models (Amir Sohel et al., 2024).

Various potato species with different sizes, colors,

and shapes have different benefits and can be

developed in a variety of climates and are a crucial

food source for many nations. However, global potato

production faces significant challenges from several

diseases and disorders during their cultivation period.

Around 20-40% of overall food crops are lost due to

diseases and pest attacks globally as such, worldwide,

32% of potatoes are lost annually. Numerous factors

are responsible for the spread of infections in crops.

The disease triangle model is taken into consideration

as the principal behind the disease development. As

for this model, three predominant components

including a plant, favorable environmental

conditions, and a pathogen cause disease. An

infection evolves when these three factors co-exist,

which further causes damage to the plants and

reduces crop (Avneet Kaur et al., 2024).

Plant pests and diseases are a significant threat to

almost all major types of plants and global food

security. Traditional inspection across different plant

fields is time-consuming and impractical for a wider

plantation size, thus reducing crop production.

Therefore, many smart agricultural practices are

deployed to control plant diseases and pests. Most of

these approaches, for example, use vision-based

artificial intelligence (AI), machine learning (ML), or

deep learning (DL) methods and models to provide

disease detection solutions. Plant pathogens and pests

cause substantial reduction in plant production

depending on adverse seasonal and environmental

conditions leading to economic and social losses.

Contemporary pests and pathogen management

depend profoundly on pesticide application, for

example, herbicides, fungicides, and insecticides

(Wasswa Shafik et al., 2023)

Plant diseases are the primary cause of quality and

quantity loss in plants/crops. Bacteria, fungi, and

viruses are responsible for the majority of plant

disease. Each year, plant diseases cause 10%-16%

losses in agricultural yields worldwide, costing the

global economy $220 billion. To feed an expanding

population, agricultural output must be increased by

70%. Chemicals used to control plant diseases, such

as bactericide and fungicide, have a negative effect on

the agroecosystem. Effective early disease detection

strategies are necessary for food security and

agroecosystem sustainability. Bacterial wilt is caused

by Ralstonia solanacearum. These bacteria may

penetrate roots via natural wounds produced in

secondary root emergence, man-made wounds

generated during cultivation. Humidity and heat

promote illness growth. A bacterial slime fills the

plant's water conducting tissue by quickly

proliferating. The plant's vascular system is affected,

although the leaves may remain green. Infected plant

stems look brown in cross section with yellowish

stuff pouring out (Siva Prasad Patnayakuni et al.,

2022).

Disease detection from leaf images has been

among the popular studies in recent years. Classifying

leaf diseases using computational methods provides

great convenience for farming. In the studies carried

out in this field, systems that work with high accuracy

and are least affected by environmental factors that

can be used in agricultural lands come to the fore.

RITECH 2025 - The International Conference on Research and Innovations in Information and Engineering Technology

This study investigates the application of deep

learning architectures for accurate and efficient plant

disease detection within the context of the ongoing

digital transformation of the agricultural sector.

Recognizing the critical role of AI in modernizing

agriculture, this research focuses on enhancing the

accuracy of the classification of plant diseases.

Applying artificial intelligence to detect and classify

plant diseases enables farmers to intervene early.

Expert laboratory study of plant leaves is a protracted

and expensive endeavor. Farmers may swiftly and

consistently make decisions through an easily

accessible artificial intelligence system, facilitating

early disease intervention and cost reduction. The

application of artificial intelligence in agriculture is

more significant due to advancing technologies. The

adoption of artificial intelligence in agriculture has

accelerated due to advancements in image processing

and big data. Deep learning and machine learning

research assist farmers in making educated decisions

by swiftly processing agricultural data (T. Ozcan et

al., 2025).

2 MATERIAL AND METHODS

This study uses an experimental approach with the

following stages:

2.1 Data and Pre-Processing

The dataset was obtained through personal data

collection from several potato plants located in

Malino, Gowa Regency, South Sulawesi. The

imagery was taken using a drone device. The image

is divided into 12 classes, namely: normal, early

blight, late blight, leaf spot, mosaic virus (PVY), leaf

roll, blackleg, soft rot, fusarium wilt, bacterial /

mucus wilt, and rhizoctonia canker (scurf). The

dataset used in this study came from potato leaf

images with an original number of 1,055 images. The

data is then shared through the Roboflow platform

with a proportion of around 72% designated for

training, 15% for validation, and 14% for testing. In

the division process, the training dataset was applied

in the form of a grayscale transformation of 15%,

resulting in a doubling of images, which caused the

number of training data to increase to 1,155 images.

The dataset consisted of 1,055 original potato leaf

images collected from Malino, Gowa Regency, South

Sulawesi, using drone devices. These images were

divided into 12 classes, including healthy leaves and

eleven disease categories: early blight, late blight, leaf

spot, mosaic virus (PVY), leaf roll, blackleg, soft rot,

fusarium wilt, bacterial wilt, and rhizoctonia canker

as shown in Figure 1. To improve the diversity and

robustness of the model, data augmentation

techniques such as rotation, flipping, and zooming

were applied. This process increased the total

number of training images to 1,155. All images were

resized to a uniform resolution of 640 × 640 pixels

before being processed to ensure consistency and

compatibility with the classification system.

Segmentation was performed using the K-Means

clustering algorithm to separate diseased areas from

the background leaf area. This process grouped

pixels into clusters based on color similarity, where

pixels representing infected spots were assigned to

specific clusters. This step helped to focus the feature

extraction process on relevant regions of the image,

improving the accuracy of subsequent classification.

Meanwhile, the number of validation data remains at

235 images, and the test data (test set) remains at 220

images. All images were re-resolved to 640×640

pixels to be uniform before being processed to the

classification stage. Once the images are collected,

they are transformed from the RGB color space to

L*a*b for segmentation purposes. Segmentation was

carried out using the K-Means algorithm to separate

disease spots from the leaf.

Figure 1: Image of Potato Leaves.

2.2 Texture Feature Extraction

(GLCM)

The process of extracting texture features is done

using the extract glcm features function. This

function begins by ensuring that the input image is

formatted with three color channels (BGR), then the

dimage is converted to grayscale. To improve the

local contrast quality, the Contrast Limited Adaptive

Classiﬁcation of Leaf Disease in Potato Plants Based on Color and Texture Features Using the K Nearest Neighbor (KNN) Method

Histogram Equalization (CLAHE) is used. Next, the

Gray Level Co-occurrence Matrix (GLCM) is

calculated at pixel distances of 1, 2, and 3 and at

angles 0°, 45°, 90°, and 135°. From the GLCM

matrix, the values of texture features in the form of

contrast, homogeneity, energy, correlation, and

texture are extracted. Angular Second Moment

(ASM). All of the texture feature values are then

combined into a single feature vector. Color feature

extraction is also done through the

extract_color_features function. This function

calculates the average intensity on each color channel

Red, Green, and Blue, from formatted input images

BGR. The result is three average values that represent

the dominant color features in the image.

2.3 Feature Processing, Normalization,

and Model Training

The feature extraction process is done on each image

in the dataset by resizing the image to the target size,

extracting texture features using GLCM, calculating

color features (Mean RGB), and adding flattened raw

pixel features so that all information is combined into

one comprehensive feature vector and stored in a

processed images list, while class labels are stored in

processed labels. Once the feature vector is obtained,

normalization is performed using the Min Max Scaler

of scikit-learn to adjust the scale to the range 0–1 so

that all features have an equal contribution and

minimize bias. The normalized data was then

dimensioned with the Principal Component Analysis

(PCA) method according to the best parameter (best

n) of optimization, then trained with KNN model

using the best number of Neighbors (best k) and

distance-based weighting for image classification

relying on texture, hue, and pixel information.

2.4 Classification of Potato Leaf

Disease Using K-Nearest Neighbor

(KNN)

The results of feature extraction are used as inputs for

the KNN algorithm. The process of classification is

performed. by calculating the Euclidean distance

between the sample data and the training data. The

optimal k-value is determined based on the results of

cross-validation to obtain the best accuracy. The

model was developed to recognize eleven classes of

diseases and tested with separate test data.

Figure 2: Image Processing and Classification Process

Diagram.

This Figure 2 illustrates the main stages in the

potato leaf disease classification system, starting from

image input, color conversion, spot segmentation,

extraction of color and texture features, to the final

classification process using the KNN algorithm.

2.5 Performance Evaluation

The evaluation of system performance is carried out

using a confusion matrix to calculate metrics such as

accuracy, precision, and recall for each class of

disease as shown in Figure 3. Cross-validation

techniques are used to ensure stable and unbiased

results.

Figure 3: Confusion Matrix.

Confusion Matrix Analysis: Most classes

achieved more than 95% accuracy, showing the

robustness of the model. However, minor

misclassifications occurred between visually similar

diseases, such as early blight and late blight, due to

overlapping features. This indicates challenges

related to class imbalance and similarity of disease

RITECH 2025 - The International Conference on Research and Innovations in Information and Engineering Technology

symptoms. Future research should focus on balancing

the dataset and incorporating advanced feature

selection or sampling techniques to mitigate this

issue. Based on the image above, it can be seen that

the confusion matrix shows the classification

performance of the KNN model with dimension

reduction using Principal Component Analysis

(PCA), which has been optimized through grid

search. Most of the data in each class was correctly

classified, which is characterized by the high diagonal

values in the matrix. For example, in class 1 there are

162 data points that are correctly classified out of a

total of 167 data points, while in class 7 there are 633

data points that are correctly classified out of a total

of 636 data. Some misclassification still occurs, such

as in class 1 which has 5 data points that are

misclassified to class 6, and class 3, which has been

misclassified to class 1 and class 11. This shows that

although the accuracy of the model is relatively high

with a value of 97.78%, there is still an overlap of

data in certain classes that cause false positives and

false negatives. Overall, these results indicate that the

PCA method is able to reduce the data dimension

without losing important information, this supporting

the performance of KNN in recognizing patterns

between classes well. This high level of accuracy is also

supported by the distribution of data on the confusion

matrix which is predominantly on target, so that the

model can be relied upon to classify test data with

consistent performance.

3 RESULTS AND DISCUSSION

3.1 Grid MAP

Figure 4: Grid MAP.

This heatmap (Figure 4) shows the grid search results

to find the best combination of the number of PCA

components and the K-value in the KNN that results

in the highest classification accuracy. The horizontal

axis represents the value of K Neighbors, while the

vertical axis indicates the number of PCA

components. Each cell contains a validation accuracy,

where lighter colors indicate higher accuracy. From

the heatmap pattern, it can be seen that the more PCA

components (about 10 to 14) and the smaller K values

(1 or 3) provide the highest accuracy, reaching 98%,

while combinations with larger Ks tend to decrease

the accuracy slightly.

The grid map illustrates the results of the grid

search process used to find the best combination of

the number of Principal Component Analysis (PCA)

components and the number of neighbors (k) in the

K-Nearest Neighbor (KNN) algorithm. The

horizontal axis represents the k values, while the

vertical axis represents the number of PCA

components tested. Lighter colors indicate higher

validation accuracy. From the grid map, it is clear that

using a higher number of PCA components (10 to 14)

with a smaller k value (1 or 3) provides the highest

accuracy, reaching up to 98%, which confirms.

3.2 KNN Classification Results

Table 1: Classification result.

Class

Table

Precision Recall

F1-

score

Support

0 0.67 1.00 0.80 2

1 0.95 0.97 0.96 167

2 1.00 0.71 0.83 14

3 0,92 0.73 0.81 15

4 1.00 1.00 1.00 2

5 1.00 1.00 1.00 1

6 1.00 0.50 0.67 6

7 0.99 1.00 0.99 636

8 1.00 1.00 1.00 1

9 1.00 1.00 1.00 1

10 1.00 1.00 1.00 2

11 0.82 1.00 0.90 9

accuracy 0.98 855

Marco

avg

0.94 0.90 0.91 855

Weighted

avg

0.98 0.98 0.98 855

The Table 1 shows the results of the classification

model performance evaluation based on Precision,

Recall, and F1-Score metrics for each class (0–10)

with a total of 855 test data samples. The model

achieved an accuracy of 98%, indicating highly

Classiﬁcation of Leaf Disease in Potato Plants Based on Color and Texture Features Using the K Nearest Neighbor (KNN) Method

accurate predictions. The Weighted Average value

for precision, recall, and F1-score was 0.98 each,

indicating consistent and stable model performance,

especially for classes with large data sets. Although

there were several classes with lower F1-scores, such

as class 6 (0.67) and class 3 (0.81), due to the small

amount of data, overall, the model was able to classify

the data well and had high generalization capabilities.

Table 2: Algorithm comparison.

Algorith

Accuracy (%)

KNN + PCA (Proposed) 97.78

SVM (RBF Kennel) 95.43

Random Fores

94.87

Based on the results in Table 2, the KNN + PCA

(Proposed) algorithm showed the highest accuracy of

97.78%, compared to SVM (RBF Kernel) at 95.43%

and Random Forest at 94.87%. These results indicate

that the application of Principal Component Analysis

(PCA) in the KNN algorithm is able to improve

classification accuracy by reducing excess feature

dimensions, so that the data recognition process

becomes more efficient. Thus, the combination of

KNN and PCA is proven to provide the best

performance compared to the other two algorithms in

this study.

Figure 5: Classification Result Graph.

The graph shows the effect of varying the K value

on the accuracy of the K-Nearest Neighbor (KNN)

model. It can be seen that increasing the K value from

4 to 14 increases the model accuracy from

approximately 0.93 to 0.98. This indicates that the

more neighbors considered in the classification

process, the more stable and accurate the prediction

results become. However, after the K value reaches

approximately 10, the accuracy increase begins to

slow and tends to stabilize, indicating that the model

has reached its optimal point.

4 CONCLUSIONS

This study successfully developed a multiclass

classification system for 11 types of potato leaf

conditions using color and texture features combined

with the K-Nearest Neighbor (KNN) algorithm

optimized by PCA. The resulting system achieved a

high accuracy of 97.78%, demonstrating that it is a

lightweight yet effective solution, suitable for use on

devices with limited computational resources.

The model has the potential to be implemented in

mobile applications to support early detection of plant

diseases in agricultural environments. However,

further testing with independent datasets is needed to

fully validate its performance before large- scale

implementation.

ACKNOWLEDGEMENTS

The authors would like to express their sincere

appreciation to the AIMP Thematic Research Group,

Faculty of Engineering, Hasanuddin University, for

the guidance, support, and facilities that enabled this

research to be successfully conducted from its initial

stages to completion. Such contributions were

invaluable in ensuring the smooth progress of this

study. Furthermore, the authors acknowledge the use

of Generative AI tools during the preparation of this

manuscript. These tools were employed in a limited

manner to enhance readability, refine language

structure, and maintain consistency in writing.

However, all ideas, analyses, data interpretations, and

conclusions presented in this study remain entirely

the original work and responsibility of the authors.

REFERENCES

F. W. Siddhi, B. Rahmat, and S. N. Hertiana, "Cattle Health

Monitoring System with Waterfall Method via the

Internet Of Things," EProceeding of Engineering, vol.

8, no. 6, pp. 3952–3961, 2022.

Hou, C., Zhuang, J., Tang, Y., He, Y., Miao, A., Huang, H.,

& Luo, S. (2021). Recognition of early blight and late

blight diseases on potato leaves based on graph cut

segmentation. Journal of Agriculture and Food

Research, 5, 100154.

I. Harfian, N. Fadhilah, and A. F. Amalia, "Technique of

using drones with RGB camera sensors and VARI

algorithms to identify stress levels of maize crops,"

Agricultural Engineering Bulletin, vol. 25, no. 2, pp.

85–88,2020.

Kaur, A., Randhawa, G. S., Abbas, F., Ali, M., Esau, T. J.,

Farooque, A. A., & Singh, R. (2024). Artificial

RITECH 2025 - The International Conference on Research and Innovations in Information and Engineering Technology

intelligence driven smart farming for accurate detection

of potato diseases: a systematic review. IEEE Access.

Manasseh, R., Sathuvalli, V., & Pappu, H. R. (2024).

Transcriptional and functional predictors of potato

virus Y-induced tuber necrosis in potato (Solanum

Ngugi, L. C., Abelwahab, M., & Abo-Zahhad, M. (2021).

Recent advances in image processing techniques for

automated leaf pest and disease recognition–A review.

Information processing in agriculture, 8(1), 27-51.

Oishi, Y., Habaragamuwa, H., Zhang, Y., Sugiura, R.,

Asano, K., Akai, K., ... & Fujimoto, T. (2021).

Automated abnormal potato plant detection system

using deep learning models and portable video

cameras. International Journal of Applied Earth

Observation and Geoinformation, 104, 102509.

Ozcan, T., & Polat, E. (2025). BorB: A Novel Image

Segmentation Technique for Improving Plant Disease

Classification with Deep Learning Models. IEEE

Access.

Patnayakuni, S. P. (2022). Tomato: Different leaf disease

detection using transfer learning, based

network. Journalof Mobile Multimedia, 18(3), 743-

756.

Qian, S., Shuo, Y., Xiaofan, G., Siting, W., Xintong, J.,

Shuang, L., & Yuanhu, X. (2021). RAVL1 activates

IDD3 to negatively regulate rice resistance to sheath

blight disease. Rice Science, 28(2), 146-155.

Rozaqi, A. J., & Sunyoto, A. (2020, November).

Identification of disease in potato leaves using

Convolutional Neural Network (CNN) algorithm.

In 2020 3rd International Conference on Information

and Communications Technology (ICOIACT) (pp. 72-

76). IEEE.

Santosa, A. I., Wulandari, R., Vadilah, M. N., Sabila, E.,

Kusuma, A. F., Mulyadi, D., ... & Çelik, A. (2025).

Survey of Potyviruses, Carlaviruses, and

Begomoviruses in Potato Cultivation Centers of West,

Central, and East Java Provinces,

Indonesia. International Journal of Plant

Biology, 16(2), 65.

Santoso, H. A., Fandhi Safsalta, B., Febrianto, N.,

Wilujeng Saraswati, G., & Haw, S. C. (2024).

Comparative analysis of convolutional neural network

and DenseNet121 transfer learning in agriculture

focusing on crop leaf disease identification. Applied

Computing and Informatics.

Sattar, S., & Khalid, N. (2024). Selection of processed and

packaged potato-based snacks among university

students: a cross-sectional study regarding food

environment and dietary behavior. Arab Gulf Journal

of Scientific Research, 42(2), 306-317.

Shafik, W., Tufail, A., Namoun, A., De Silva, L. C., &

Apong, R. A. A. H. M. (2023). A systematic literature

review on plant disease detection: Motivations,

classification techniques, datasets, challenges, and

future trends. Ieee Access, 11, 59174-59203.

Sohel, A., Shakil, M. S., Siddiquee, S. M. T., Al Marouf,

A., Rokne, J. G., & Alhajj, R. (2024). Enhanced Potato

Pest Identification: A Deep learning approach for

identifying potato pests. IEEE Access.

Y. Alkhalifi, "Detection of Capsicum Plant Diseases Based

on Leaf Images Using FineTuned Transfer Learning,"

Thesis, 2021.

Zhuo, L. Ü., Jing, W. A. N. G., Jing, Z. H. U., Mei-ying,

G. U., Qi-yong, T. A. N. G., Wei, W. A. N. G., & Bo,

W. A. N. G. (2020). First report of a new potato disease

caused by Galactomyces candidum F12 in

China. Journal of Integrative Agriculture, 19(10),

2470-2476.

Classiﬁcation of Leaf Disease in Potato Plants Based on Color and Texture Features Using the K Nearest Neighbor (KNN) Method