Early Diagnosis and Severity Assessment of Weligama Coconut Leaf Wilt Disease and Coconut Caterpillar Infestation using Deep Learning-based Image Processing Techniques
SAMITHA VIDHANAARACHCHI 1,*, JANAKA L. WIJEKOON 2,3,*, W. A. SHANAKA P. ABEYSIRIWARDHANA 4, AND MALITHA WIJESUNDARA 1 · 2025
Read Early Diagnosis and Severity Assessment of Weligama Coconut Leaf Wilt Disease and Coconut Caterpillar Infestation using Deep Learning-based Image Processing Techniques by SAMITHA VIDHANAARACHCHI 1,*, JANAKA L. WIJEKOON 2,3,*, W. A. SHANAKA P. ABEYSIRIWARDHANA 4, AND MALITHA WIJESUNDARA 1 (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2501-18835.txt.
Introduction
Samitha Vidhanaarachchi and co-authors teach cameras to scout coconut plantations for two scourges: Weligama leaf wilt disease and caterpillar infestation. Both are still spotted by walking inspectors, slowly and late. Using image sets gathered in three Sri Lankan districts, the team trained transfer-learned convolutional networks for diagnosis and a mask network for severity.
The classifiers caught the wilt disease at about ninety percent accuracy and caterpillar damage at about ninety-five. A severity grader for the wilt hit about ninety-seven. YOLO detectors in three versions then counted individual caterpillars on leaflets at around ninety-six percent accuracy. The authors present the pipeline as an early-warning system for a crop whose losses are currently found by eye, after the fact.
Catalog record
| CREATOR | SAMITHA VIDHANAARACHCHI 1,*, JANAKA L. WIJEKOON 2,3,*, W. A. SHANAKA P. ABEYSIRIWARDHANA 4, AND MALITHA WIJESUNDARA 1 |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 52 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2501-18835
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Early Diagnosis and Severity Assessment of Weligama Coconut Leaf Wilt Disease and Coconut Caterpillar Infestation using Deep Learning-based Image Processing Techniques SAMITHA VIDHANAARACHCHI 1,*, JANAKA L. WIJEKOON 2,3,*, W. A. SHANAKA P. ABEYSIRIWARDHANA 4, AND MALITHA WIJESUNDARA 1 Abstract Global Coconut (Cocos nucifera (L.)) cultivation faces significant challenges, including yield loss, due to pest and disease outbreaks. In particular, Weligama Coconut Leaf Wilt Disease (WCWLD) and Coconut Caterpillar Infestation (CCI) damage coconut trees, causing severe coconut production loss in Sri Lanka and nearby coconut-producing countries. Currently, both WCWLD and CCI are detected through on-field human observations, a process that is not only time-consuming but also limits the early detection of infections. This paper presents a study conducted in Sri Lanka, demonstrating the effectiveness of employing transfer learning-based Convolutional Neural Network (CNN) and Mask Region-based-CNN (Mask R-CNN) to identify WCWLD and CCI at their early stages and to assess disease progression. Further, this paper presents the use of the You Only Look Once (YOLO) object detection model to count the number of caterpillars distributed on leaves with CCI. The introduced methods were tested and validated using datasets collected from Matara, Puttalam, and Makandura, Sri Lanka. The results show that the proposed methods identify WCWLD and CCI with an accuracy of 90% and 95%, respectively. In addition, the proposed WCWLD disease severity identification method classifies the severity with an accuracy of 97%. Furthermore, the accuracies of the object detection models for calculating the number of caterpillars in the leaflets were: YOLOv5-96.87%, YOLOv8-96.1%, and YOLO11-95.9%. † † history: Date of publication xxxx 00, 0000, date of current version xxxx 00, 0000. † † doi: 10.1109/ACCESS.2022.0092316 † † address: Sri Lanka Institute of Information Technology, New Kandy Rd, Malabe 10115, Sri Lanka † † address: Victorian Institute of Technology, Adelaide Campus, South Australia † † address: Department of System Design Engineering, Keio University, Yokohama, Japan † † address: ACSL Ltd 3-6-4 Rinkaicho Edogawa, Tokyo 134-0086 Japan † † corresponding: Corresponding author: Samitha Vidhanaarachchi (e-mail: [email protected]). I Introduction The Cocos nucifera (L.) (commonly known as coconut) tree is a member of the palm family, which has a 50-60-year economic lifespan and starts bearing fruits within 5-6 years after being planted. Usually, the tree grows up to 30 m with a crown of 30-40 leaves, making an ample amount of food and shelter for the survival and reproduction of many insects and pests [ 1 ] . Pests and diseases in coconut plants cause fatal damage to the palm and can result in outbreaks, causing heavy economic losses [ 1 ] . Therefore, efficient disease and pest-controlling methods are vital for coconut cultivation [ 2 ] . Recent studies [ 3 , 4 , 5 , 6 , 7 , 8 ] discuss that Weligama Coconut Leaf Wilt Disease (WCLWD) and Coconut Caterpillar Infestation (CCI) pose the greatest threats to coconut plantations in Sri Lanka and regional countries, causing rapid and severe damage to the coconut cultivation. According to [ 4 , 5 , 6 ] , WCLWD, which is an incurable disease, has been detected to be spreading at an alarming rate in Sri Lanka due to lack of early detection capabilities. During the expert survey, it was confirmed by principal entomologists from the Coconut Research Institute of Sri Lanka (CRISL) [ 9 ] WCLWD symptoms, specifically uneven yellowing [ 10 , 3 , 11 ] , are frequently misclassified as the yellowing caused by Magnesium (Mg) deficiency [ 10 , 11 ] , making it difficult for even trained individuals to differentiate the two [ 3 ] . Similarly, the dried appearance of leaves caused by CCI is also difficult to distinguish from Leaf Scorch Decline (LSD) [ 3 ] , which is a physiological disorder. For these reasons, the majority of coconut growers are incapable of identifying the WCLWD and CCI infestations at their early stages. The principal entomologists also confirmed that both WCLD and CCI are detected using on-field human observations, which is not only time-consuming but also limits the early detection of infections. ICT tools have been successfully incorporated into developing pest and disease management practices in the recent past to early identification and treatments [ 12 , 13 ] . Also, recent literature has shown that Machine Learning (ML) and artificial intelligence (AI) are thriving in agricultural research studies, significantly improving precision farming through automated analysis and decision-making [ 14 , 15 ] . Among these advancements, the use of computer vision and deep learning (DL) techniques, such as feature extraction to automate pest and disease classification, is gaining momentum as a valuable tool to modernize conventional observation methods [ 1 , 16 , 2 , 17 , 18 , 19 , 20 ] . Consequently, as an extension of the preliminary study [ 13 ] , this paper presents a study on the effectiveness of using DL techniques for the early identification of CCI, automating the detection of caterpillars, and the early identification of WCLWD. II Background Survey TABLE I : Summary of significant ML and DL applications in agriculture Ref Crop Objectives Methods Respective Accuracy’s Year [ 21 ] Pear, Cherry etc… Disease classification CNN-based models 96.3% 2016 [ 17 ] Black Gram Nutrient deficiency classification Transfer learning-based CNN 65.44% 2019 [ 19 ] Wheat Disease propagation prediction ANN, Logistic Models 93% & 90% 1997 [ 22 ] Tomato, Beans etc… Disease detection & dispersion CNN, Mask R-CNN, Gaussian Plume 90% - 94% 2022 [ 20 ] Tomato Disease detection Near-infrared spectroscopy 78% 2004 [ 23 ] Moth Orchids Disease detection Gray level co-occurrence matrix 89.6% 2007 [ 24 ] Tomato Disease detection Faster R-CNN, Mask R-CNN 88.53% & 99.64% 2021 [ 25 ] Cabbage Pest Detection CNN, Mask R-CNN 95.3% & 98.3% 2024 [ 26 ] Guava, Citrus etc… Nutrition deficiency classification CNN-based model 88% 2020 [ 27 ] Tea Disease severity assessment CNN, YOLOv8, Mask R-CNN 89.9%, 98% & 89% 2023 [ 28 ] Maize Pest Detection Mask RCNN 94.21% 2023 [ 16 ] Wheat, Maize etc… Pest classification & counting OpenCV, ANN 97.9% & 92.5% 2020 [ 2 ] Coconut Pest and disease identification DNN Only Proposed 2019 [ 29 ] Coconut Disease Classification Gated Recurrent Unit 97.75% 2022 [ 30 ] Coconut Infestation Classification CNN, Faster R-CNN 84.64% & 97.30% 2021 [ 31 ] Coconut Disease Classification CNN 96.94% 2021 This section explores the application of DL and Machine Learning (ML) in agriculture, focusing on disease and pest management. The section covers the use of DL models, CNNs, instance segmentation, and object detection techniques in identifying and classifying plant diseases, nutrient deficiencies, and pests. The review also highlights specific studies and their findings in these areas, emphasizing the potential and current limitations in the field. A comprehensive summary is presented in Table I . Several studies have explored the use of DL techniques for disease identification and management. For instance, Sladojevic et al. [ 21 ] and Han et al. [ 17 ] have highlighted the success of CNN-based models in classifying healthy and diseased leaves, as well as detecting nutrient deficiencies in black gram. The work of Francl et al. [ 19 ] , further underscores the application of Artificial Neural Network (ANN)