Improved Segmentation Approach for Plant Disease Detection

The wide scale prevalence of diseases in agricultural crops affects both the production quality and quantity of agricultural products at local to regional scale. More often than not, the diseases remain unidentified causing huge distress to the farmers while threatening national food security. In order to circumvent this problem, early diagnosis of diseases using a fast and reliable method is beneficial. Plant disease identification from images captured by digital cameras is an area of active research. Use of various machine learning algorithms for plant disease classification and the evolution of deep convolutional neural network (CNN) based architectures have further enhanced the plant disease classification accuracy. In this context, an automated computer vision-based plant disease detection and classification scheme from plant and leaf’s photographs will be highly desirable. Although, there exist a few techniques currently used in an adhoc fashion for plant disease detection and/or classification, a systematic study to evaluate their usage and efficacy on actual plant data has largely remained unexplored.The aim of this paper is to evaluate various CNN based state-of-the-art transfer learning architectures like GoogLeNet, AlexNet, VGG16 and ResNet50V2 models for plant disease detection and classification. The models were tested on popular publicly available three plant disease benchmark database such as PlantVillage Dataset, New Plant Disease Dataset and Plant Pathology Dataset. Various validation metrics such as Precision, Recall, F1 score and overall accuracy were used to evaluate the final results of the experiments, which revealed that VGG16 rendered highest accuracy of 96.6%, 98.5% and 89% on the three dataset respectively, outperforming all other state-of-the-art models.

Subsistence farmers and global food security depend on sufficient food production, which aligns with the UN's “Zero Hunger,” “Climate Action,” and “Responsible Consumption and Production” sustainable development goals. In addition to already available methods for early disease detection and classification facing overfitting and fine feature extraction complexities during the training process, how early signs of green attacks can be identified or classified remains uncertain. Most pests and disease symptoms are seen in plant leaves and fruits, yet their diagnosis by experts in the laboratory is expensive, tedious, labor-intensive, and time-consuming. Notably, how plant pests and diseases can be appropriately detected and timely prevented is a hotspot paradigm in smart, sustainable agriculture remains unknown. In recent years, deep transfer learning has demonstrated tremendous advances in the recognition accuracy of object detection and image classification systems since these frameworks utilize previously acquired knowledge to solve similar problems more effectively and quickly. Therefore, in this research, we introduce two plant disease detection (PDDNet) models of early fusion (AE) and the lead voting ensemble (LVE) integrated with nine pre-trained convolutional neural networks (CNNs) and fine-tuned by deep feature extraction for efficient plant disease identification and classification. The experiments were carried out on 15 classes of the popular PlantVillage dataset, which has 54,305 image samples of different plant disease species in 38 categories. Hyperparameter fine-tuning was done with popular pre-trained models, including DenseNet201, ResNet101, ResNet50, GoogleNet, AlexNet, ResNet18, EfficientNetB7, NASNetMobile, and ConvNeXtSmall. We test these CNNs on the stated plant disease detection and classification problem, both independently and as part of an ensemble. In the final phase, a logistic regression (LR) classifier is utilized to determine the performance of various CNN model combinations. A comparative analysis was also performed on classifiers, deep learning, the proposed model, and similar state-of-the-art studies. The experiments demonstrated that PDDNet-AE and PDDNet-LVE achieved 96.74% and 97.79%, respectively, compared to current CNNs when tested on several plant diseases, depicting its exceptional robustness and generalization capabilities and mitigating current concerns in plant disease detection and classification.

Plant disease detection is a critical task in agriculture, essential for ensuring crop health and productivity. Traditional methods in this context are often labor-intensive and prone to errors, highlighting the need for automated solutions. While computer vision-based solutions have been successfully deployed in recent years for plant disease identification and localization tasks, these often operate independently, leading to suboptimal performance. It is essential to develop an integrated solution combining these two tasks for improved efficiency and accuracy. This research proposes the innovative Plant Disease Localization and Classification model based on Vision Transformer (PDLC-ViT), which integrates co-scale, co-attention, and cross-attention mechanisms and a ViT, within a Multi-Task Learning (MTL) framework. The model was trained and evaluated on the Plant Village dataset. Key hyperparameters, including learning rate, batch size, dropout ratio, and regularization factor, were optimized through a thorough grid search. Early stopping based on validation loss was employed to prevent overfitting. The PDLC-ViT model demonstrated significant improvements in plant disease localization and classification tasks. The integration of co-scale, co-attention, and cross-attention mechanisms allowed the model to capture multi-scale dependencies and enhance feature learning, leading to superior performance compared to existing models. The PDLC-ViT model evaluated on two public datasets achieved an accuracy of 99.97%, a Mean Average Precision (MAP) of 99.18%, and a Mean Average Recall (MAR) of 99.11%. These results underscore the model's exceptional precision and recall, highlighting its robustness and reliability in detecting and classifying plant diseases. The PDLC-ViT model sets a new benchmark in plant disease detection, offering a reliable and advanced tool for agricultural applications. Its ability to integrate localization and classification tasks within an MTL framework promotes timely and accurate disease management, contributing to sustainable agriculture and food security.

The detection and classification of plant diseases are crucial to ensuring food security and maximizing agricultural productivity. Traditional methods of plant disease identification are time-consuming and labor-intensive, necessitating the adoption of more efficient and accurate techniques. In recent years, advancements in machine learning have led to the development of robust methods for automated plant disease detection. This research paper presents a novel approach for plantdisease detection using Convolutional Neural Networks (CNNs). CNNs have demonstrated exceptional performance in image recognition tasks, making them a promising choice for detecting diseases in plant images. The proposed system utilizes a pre- processed dataset of plant images, comprising both healthy and diseased samples, to train the CNN model. Keywords—Plant disease and detection, convolutional neural networks, deep learning architectures, feature extraction, classifier methods

Our project presents an innovative approach to address the crucial issue of plant disease identification through the utilization of deep learning techniques. Agricultural productivity is significantly affected by plant diseases, leading to economic losses and food security concerns. In this research, a comprehensive dataset comprising images of healthy and diseased plants is collected and pre-processed for training deep convolutional neural networks (CNNs). The proposed system harnesses the power of deep learning to automatically learn intricate patterns and features from plant images, enabling accurate classification between healthy and diseased states. The trained model is evaluated on an independent dataset to assess its classification performance. Various deep learning architectures, such as VGG, ResNet, and Inception, are experimented with to identify the most suitable architecture for achieving high accuracy and robustness. To enhance the model's generalization capability, data augmentation techniques are employed during training. Transfer learning is also explored, allowing the pre-trained models to be fine-tuned for the specific task of plant disease classification. The performance metrics, including accuracy, precision, recall, and F1-score, are thoroughly evaluated to quantify the model's effectiveness. The proposed deep learning-based plant disease classification system holds great promise for real-world agricultural applications. Its automated and accurate nature has the potential to revolutionize plant disease management by enabling early detection and timely intervention. As a result, this research contributes to the advancement of precision agriculture practices, helping to mitigate the adverse impacts of plant diseases and promote sustainable agricultural production.

A review on machine learning and deep learning image-based plant disease classification for industrial farming systems

Smart agriculture utilizes modern technologies to increase crop productivity. The productivity of agricultural crops has to be elevated for food security for the ever-increasing global population. The production of agricultural crops is largely affected due to increasing infestation of diseases and pests in addition to abiotic stresses. Early detection and management of diseases hold key to tackling the challenge. The disease pest infestation can be controlled by applying pesticides and insecticides. But numerous negative health effects that have been associated with chemical pesticides have been well documented. The increasing computational technology and recent advances in deep learning have paved the way for rapid disease diagnosis and management. Here we have discussed the automatic detection and classification of plant diseases as well as their severity through Image Processing. The detection of disease and its degree of severity from images is based on colour, texture and shape and gives a fast and accurate solution through the use of smart computational tools. DNNs (Deep Neural Networks) and CNNs (Convolutional Neural Networks) are effective in the detection, recognition and classification of plant diseases towards an automated solution for the large-scale agricultural industry.

Abstract—Everyone’s daily existence in the twenty-first century involves a large amount of machine learning.These days, a wide range of applications, including object identification, object categorisation, and medicinal uses, use it.This seeks to identify illnesses in plant leaves using deep convolutional neural networks. Farmers typically use manual disease detection techniques since they are ignorant of illnesses on plant leaves. They frequently become less abundant as the virus multiplies.However, in many parts of the world, rapid identification has to be enhanced due to a lack of necessary infrastructure.The efficient handling of plant diseases is essential for both food security and agricultural production.The proposed application uses CNNs for plant disease detection and classification. It attempts to identify illnesses from photos and offers disease descriptions and preventive steps and suggests suitable pesticides along with product link for further action and other features of the application are crop recommendation and fertilizer prediction based on some values(nitrogen,phosphorous,pottasium).The CNN model archi- tecture is optimized for accurate disease classification,taking into account various types of plant diseases and leaf variations.This aims to enhance agricultural practices by providing a scal- able and efficient tool for early disease detection,classification, and treatment recommendation,crop recommendation,thereby contributing to improved crop yield and sustainable farming practices. Index Terms—Plant Disease Detection, CNN (Convolutional Neural Network), Automated Disease Description,Pesticide Sug- gestion,Precision Farming.

With the rapid development of precision agriculture and smart agriculture, the need to build an automatic identification and detection system for diseases and insect pests is increasing. Using computers to correctly label plant diseases and insect pests is an important prerequisite for achieving accurate classification of plant diseases and insect pests and ensuring system performance. In order to improve the accuracy of computer classification of plant pests and diseases, this paper proposes an automatic pest identification method based on the Vision Transformer (ViT). In order to avoid training overfitting, the plant diseases and insect pests data sets are enhanced by methods such as Histogram Equalization, Laplacian, Gamma Transformation, CLAHE, Retinex-SSR, and Retinex-MSR. Then use the enhanced data set to train the constructed ViT neural network, so as to realize the automatic classification of plant diseases and insect pests. The simulation results show that the constructed ViT network has a test recognition accuracy rate of 96.71% on the plant disease and insect pest public data set Plant_Village, which is about 1.00% higher than the Plant disease and pest identification method based on traditional convolutional neural networks such as GoogleNet and EfficentNetV2.

The majority of agricultural yields are affected by plant diseases leading to economic losses. Hence, it is necessary to evaluate plant diseases at the very beginning to enhance crop production by making effective control actions. Therefore, automatic plant disease detection models are required to reduce the huge laborious tasks. In recent times, deep learning and computer vision models have been established for the automated detection of plant disease which functions by analyzing the symptoms on plant leaves. Thus, this paper reviews a set of articles that utilizes deep learning and image processing in the field of detecting plant and crop diseases. This survey ensures a comprehensive view of all research works regarding plant disease detection for exploring the effectiveness of the existing models for detecting diseases in diverse crops. This paper aims to analyze various machine learning and deep learning approaches regarding crop disease identification. This review gives an introduction to plant leaf diseases, the conventional methods for detecting plant diseases, the dataset used, the performance metrics utilized for analyzing the existing detection frameworks, and the implementation platforms. Finally, this review provides the limitations in this field of plant leaf detection and the future research trends for further enhancement.

Traditional plant disease diagnosis methods are mostly based on expert diagnosis, which easily leads to the backwardness of crop disease control and field management. In this paper, to improve the speed and accuracy of disease classification, a plant disease detection and classification method based on the optimized lightweight YOLOv5 model is proposed. We propose an IASM mechanism to improve the accuracy and efficiency of the model, to achieve model weight reduction through Ghostnet and WBF structure, and to combine BiFPN and fast normalization fusion for weighted feature fusion to speed up the learning efficiency of each feature layer. To verify the effect of the optimized model, we conducted a performance comparison test and ablation test between the optimized model and other mainstream models. The results show that the operation time and accuracy of the optimized model are 11.8% and 3.98% higher than the original model, respectively, while F1 score reaches 92.65%, which highlight statistical metrics better than the current mainstream models. Moreover, the classification accuracy rate on the self-made dataset reaches 92.57%, indicating the effectiveness of the plant disease classification model proposed in this paper, and the transfer learning ability of the model can be used to expand the application scope in the future.

In recent years, the rapid advancement of computer vision and machine learning techniques has revolutionized the field of agriculture by enabling automated detection and classification of plant diseases. This work presents a comprehensive approach for the efficient segmentation and multi-class classification of heterogeneous plant disease and non-disease images. The proposed methodology combines multi-level thresholding-based filtering, feature ranking, leaf feature segmentation, and ensemble classification to achieve high accuracy in disease recognition. In the initial stage, multi-level thresholding-based filtering is applied to enhance image quality and reduce noise, facilitating more precise disease pattern extraction. Subsequently, feature ranking methods are employed to select the most discriminative features from the pre-processed images. These features play a pivotal role in distinguishing between various plant diseases and healthy leaves. Experimental results on a multi-modal plant disease and non- disease images demonstrate the superiority of our proposed method in terms of accuracy and efficiency.

Presently, rapid and precise disease identification process plays a vital role to increase agricultural productivity in a sustainable manner. Conventionally, human experts identify the existence of anomaly in plants occurred due to disease, pest, nutrient deficient, weather conditions. Since manual diagnosis process is a tedious and time consuming task, computer vision approaches have begun to automatically detect and classify the plant diseases. The general image processing tasks involved in plant disease detection are preprocessing, segmentation, feature extraction and classification. This paper performs a review of computer vision based plant disease detection and classification techniques. The existing plant disease detection approaches including segmentation and feature extraction techniques have been reviewed. Additionally, a brief survey of machine learning (ML) and deep learning (DL) models to identify plant diseases also takes place. Furthermore, a set of recently developed DL based tomato plant leaf disease detection and classification models are surveyed under diverse aspects. To further understand the reviewed methodologies, a detailed comparative study also takes place to recognize the unique characteristics of the reviewed models.

In recent years, intelligent plant disease detection has become an essential task in agriculture for ensuring crop health and improving yield. Detection and classification of plant diseases remain a challenging problem due to the vast numbers of plants species worldwide and the numerous diseases that negatively affect crop production. However, existing Deep Learning (DL) methods are used for plant disease detection and classification suffer from overfitting, limited robustness, and lack of interpretability. To overcome these issues, this paper proposes a Hybrid Vision Graph Neural Network with Ontology-Driven Knowledge Graph Embedding (HV-GNN-OKGE) method for effective plant disease detection and classification. The HV-GNN method captures both local and global structural features through isotropic and pyramid layers. While the ontology-driven embedding integrates semantic relationships among plant species, symptoms, and diseases to enrich classification with domain knowledge. The performance of the proposed HV-GNN-OKGE method is evaluated using two publicly available datasets: Plant Leaf Dataset (4,590 images, 22 classes) and PlantVillage (2,052 images, 2 classes) are used for training and evaluation, with preprocessing techniques including augmentation, resizing, and normalization to enhance image quality. Experimental results demonstrate that the proposed HV-GNN-OKGE achieves an accuracy of 99.95%, significantly outperforming existing deep learning models in plant disease detection and classification.

Agriculture accepts a basic part by virtue of the quick improvement of the general population and extended interest in food in India. Hence, it is required to increase harvest yield. One serious cause of low collect yield is an infection brought about by microorganisms, infection, and organisms. Plant disease investigation is one of the major and essential tasks in the part of cultivating. It tends to be forestalled by utilizing plant disease detection techniques. To monitor, observe or take care of plant diseases manually is a very complex task. It requires gigantic proportions of work, and moreover needs outrageous planning time; consequently, image processing is utilized to distinguish diseases of plants. Plant disease classification can be done by using machine learning algorithms which include steps like dataset creation, load pictures, pre-preparing, segmentation, feature extraction, training classifier, and classification. The main objective of this research is to construct one model, which classifies the healthy and diseased harvest leaves and predicts diseases of plants. In this paper, the researchers have trained a model to recognize some unique harvests and 26 diseases from the public dataset which contains 54,306 images of the diseases and healthy plant leaves that are collected under controlled conditions. This paper worked on the ResNets algorithm. A residual neural network (ResNet) is a subpart of the artificial neural network (ANN). ResNet algorithm contains a residual block that can be used to solve the problem of vanishing/exploding gradient. ResNet algorithm is also used for creating Residual Network. For the image classification, ResNets achieve a much well result. The ResNets techniques applied some of the parameters like scheduling learning rate, gradient clipping, and weight decay. Using the ResNet algorithm, the researchers expect high accuracy results and detecting more diseases from the various harvests.