Image-Based Hot Pepper Disease and Pest Diagnosis Using Transfer Learning and Fine-Tuning

Gu, Yeong Hyeon; Yin, Helin; Dong, Jin; Park, Jonghan; Yoo, Seong Joon

doi:10.3389/fpls.2021.724487

Cited by 23 publications

(11 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To compare the proposed YOLO V5s-CAcT3 with other advanced methods, six well-known CNNs, such as Faster RCNN, SSD, YOLO v3, YOLO V4, and YOLO V5s, were selected as baseline methods for comparison experiments. By applying transfer learning methods, pre-trained weights are obtained on ImageNet (Gu et al, 2021 ) to initialize the weight parameters, and Softmax is embedded into the network for classification. The hyperparameters assigned to the network are a learning rate of 0.001, a momentum of 0.9, a batch size of 64, and a stochastic gradient descent (SGD) solver with unrestricted epochs for each model and multiple fine-tuning to ensure optimal convergence.…”

Section: Resultsmentioning

confidence: 99%

An Industrial-Grade Solution for Crop Disease Image Detection Tasks

Dai

Fan

2022

Front. Plant Sci.

View full text Add to dashboard Cite

Crop leaf diseases can reflect the current health status of the crop, and the rapid and automatic detection of field diseases has become one of the difficulties in the process of industrialization of agriculture. In the widespread application of various machine learning techniques, recognition time consumption and accuracy remain the main challenges in moving agriculture toward industrialization. This article proposes a novel network architecture called YOLO V5-CAcT to identify crop diseases. The fast and efficient lightweight YOLO V5 is chosen as the base network. Repeated Augmentation, FocalLoss, and SmoothBCE strategies improve the model robustness and combat the positive and negative sample ratio imbalance problem. Early Stopping is used to improve the convergence of the model. We use two technical routes of model pruning, knowledge distillation and memory activation parameter compression ActNN for model training and identification under different hardware conditions. Finally, we use simplified operators with INT8 quantization for further optimization and deployment in the deep learning inference platform NCNN to form an industrial-grade solution. In addition, some samples from the Plant Village and AI Challenger datasets were applied to build our dataset. The average recognition accuracy of 94.24% was achieved in images of 59 crop disease categories for 10 crop species, with an average inference time of 1.563 ms per sample and model size of only 2 MB, reducing the model size by 88% and the inference time by 72% compared with the original model, with significant performance advantages. Therefore, this study can provide a solid theoretical basis for solving the common problems in current agricultural disease image detection. At the same time, the advantages in terms of accuracy and computational cost can meet the needs of agricultural industrialization.

show abstract

Section: Resultsmentioning

confidence: 99%

An Industrial-Grade Solution for Crop Disease Image Detection Tasks

Dai

Fan

2022

Front. Plant Sci.

View full text Add to dashboard Cite

show abstract

“…Traditional methods of identification such as support vector machines, Naive Bayes and BP neural networks are not suitable for large area disease-pest identification in the field due to low recognition rate and weak generalization. In contrast, deep learning methods based on convolutional neural networks (CNN) have shown remarkable results and have strong generalization ( Gu et al, 2021 ). Pre-trained VGG and ResNet 50 architectures based on the ImageNet dataset are commonly used due to the scarcity of images of crop disease-pests.…”

Section: Crop Disease-pest Controlmentioning

confidence: 99%

Editorial: Machine learning and artificial intelligence for smart agriculture, volume II

Zhang

Park

et al. 2023

Front. Plant Sci.

View full text Add to dashboard Cite

Editorial on the Research TopicMachine learning and artificial intelligence for smart agriculture, volume II 1 Introduction Currently, AI is being widely used in various agricultural scenarios, including intelligent perception, real-time field monitoring, intelligent early warning, disease and pest detection, and intelligent decision-making for crop production environments. With the help of AI, farmers can now detect whether there are any diseases and pests, whether they need to use pesticides, and whether their plant protection practices are effective. This special edition focuses on several issues that still require further study and discussion, such as agricultural unmanned aerial vehicles, crop type mapping, crop phenotypic analysis, and the identification of crop diseases and pests in sustainable and intelligent phytoprotection. Agricultural unmanned aerial vehicleAgricultural unmanned aerial vehicles (AUAVs) integrate robots, AI, big data, and the Internet of Things. They have been widely applied to various agricultural operations, such as seed sowing, land monitoring, crop disease and pest detection, and pesticide and fertilizer spraying. AUAVs greatly improve agricultural production efficiency and liberate the labor force (Kim et al., 2019). They are becoming a new force in the field of precision agricultural aviation (Wang et al., 2019). Compared to traditional agricultural machinery, they are small, lightweight, and easy to transport, and have flexible flight control. AUAVs are characterized by precision operation, high efficiency, environmental friendliness, intelligence, and ease of use. However, in many cases, real-time changes in the AUAV load during flight can affect its speed, accuracy, and flight path stability. Xu et al. (Wei et al.) proposed a flight dynamics model for achieving AUAV flight trajectory stability using a PID controller and robust T-S fuzzy control method. This model can achieve certain stability in the flight path against load perturbations for different Frontiers in Plant Science frontiersin.org 01

show abstract

“…In the leaf disease recognition domain, researchers have employed enhanced deep learning network architecture through various models [9][10] and applied them to chili disease recognition. This research employs two types of deep learning models: CNN and ResNet-18, which are developed from scratch using the Deep Network Designer [19].…”

Section: Developments Of Deep Learning Modelmentioning

confidence: 99%

“…With small collection of datasets, several research [7][8][9][10] in the chili agricultural field has used the data augmentation method to increase the volume of datasets. The method generates data artificially via adding augmented images to the existing dataset through either oversampling or warping [11].…”

Section: Introductionmentioning

confidence: 99%

An Improved Deep Learning Model of Chili Disease Recognition with Small Dataset

Aminuddin¹,

Tukiran²,

Joret³

et al. 2022

IJACSA

View full text Add to dashboard Cite

Due to its tasty and spicy fruit with nutritional qualities, chili is a demanding crop widely farmed around the world. Hence, it is essential to accurately determine the health status of chili for agricultural productivity. Recent years have seen impressive results in recognition fields due to deep learning approaches. However, deep learning models' networks need an abundant data to perform well and collecting enormous data for the networks is time-consuming and resource-intensive. A data augmentation method is proposed to overcome this problem. It was applied to a small dataset of healthy and diseased chili leaf by utilizing geometric transformation method. Eventually, two deep learning models of CNN and ResNet-18 were evaluated using augmented and original datasets. From a series of experiment, it can be concluded that the trained deep learning models using original and augmented datasets perform better with an average accuracy performance of 97%.

show abstract

Image-Based Hot Pepper Disease and Pest Diagnosis Using Transfer Learning and Fine-Tuning

Cited by 23 publications

References 34 publications

An Industrial-Grade Solution for Crop Disease Image Detection Tasks

An Industrial-Grade Solution for Crop Disease Image Detection Tasks

Editorial: Machine learning and artificial intelligence for smart agriculture, volume II

An Improved Deep Learning Model of Chili Disease Recognition with Small Dataset

Contact Info

Product

Resources

About