PlantDet: A Robust Multi-Model Ensemble Method Based on Deep Learning For Plant Disease Detection

Abstract

Plant disease is a significant health concern among all living creatures. Early diagnosis can help farmers takenecessary steps to cure the disease and accelerate the production rate efficiently. Our research has beenconducted with five most common rice leaf diseases, such as bacterial leaf blight, brown spot, leaf blast,leaf scald, and narrow brown spot, including healthy class, and two categories of betel leaf, such as healthyand unhealthy class. A robust new deep ensemble model, based on InceptionResNetV2, EfficientNetV2L,and Xception, has been proposed, known as, PlantDet, in this research. PlantDet solves not only underfitting problems but also leverage nourished performances simultaneously for scarce dataset of the sparse number of different background image dataset. PlantDet integrates efficient data augmentation, preprocessing, Global Average Pooling layer, Dropout mechanism, L2 regularizers, PRelu activation function, Batch Normalization layers, and more Dense layers that make the model more robust compared to all existing models and help to handle underfitting and overfitting problems while maintaining high performance. PlantDet exceeds the previous state-of-art model for the Rice Leaf dataset with an Accuracy of 98.53%, a Precision of 98.50%, a Recall of 98.35%, a F1 of 98.42% and a Specificity of 99.71%. In addition, for the Betel Leaf dataset, PlantDet also surpassed all existing base models, including several robust ensemble models. Finally, Grad-CAM and Score-CAM have been accomplished with the Xception method to explain the model performances particularly to elaborate how the Deep Learning (DL) models works for this complex dataset. Score-CAM slightly outperformed Grad-CAM++ in terms of localizing the predicted area.

Phenotyping and predicting wheat spike characteristics using image analysis and machine learning

“…The ResNet152V2 model was selected due to its prevalence in image classification literature in a variety of different fields from disease ratings in agriculture (Kanchanadevi & Sandhia, 2023;Nigam et al, 2023) to medical research (Sulaiman et al, 2023). The final model was the pre-trained EfficientNetV2L model (Tan & Le, 2021), which was selected due to its recent use in plant disease detection (Shovon et al, 2023;Ulutaş & Aslantaş, 2023). Mean squared error (MSE) was used as the loss function for each model, which evaluates differences between values predicted by the model and their true value.…”

A Comprehensive Dataset of 10 Exotic Vegetable Plants Across Three Growth Stages

“…In the broader landscape of agricultural research, this dataset stands as a testament to the synergy of technology and agriculture [1][2][3] . While several studies have explored the potential of hydroponics, deep learning, and precision agriculture individually, the availability of a comprehensive image dataset bridges the gap between these domains, setting the stage for integrated and holistic research 10,15,18,19 .…”

An Effective approach for Plant Disease Detection Using Assessment-Based Convolutional Neural Networks (A-CNN)

“…The author proposed InceptionResNetV2, EfficientNetV2L and Exception based on a new powerful deep ensemble model, then already submitted and called PlantDet. Their performance can solve the planet mismatch problem when applied to sparse datasets with diverse background image datasets [9]. The author proposed a lightweight transfer learning-based approach to detect tomato plant leaf disease.…”