Crop Disease Detection Research Articles

An Enhanced Deep Learning Model for Effective Crop Pest and Disease Detection

Traditional machine learning methods struggle with plant pest and disease image recognition, particularly when dealing with small sample sizes, indistinct features, and numerous categories. This paper proposes an improved ResNet34 model (ESA-ResNet34) for crop pest and disease detection. The model employs ResNet34 as its backbone and introduces an efficient spatial attention mechanism (effective spatial attention, ESA) to focus on key regions of the images. By replacing the standard convolutions in ResNet34 with depthwise separable convolutions, the model reduces its parameter count by 85.37% and its computational load by 84.51%. Additionally, Dropout is used to mitigate overfitting, and data augmentation techniques such as center cropping and horizontal flipping are employed to enhance the model’s robustness. The experimental results show that the improved algorithm achieves an accuracy, precision, and F1 score of 87.09%, 87.14%, and 86.91%, respectively, outperforming several benchmark models (including AlexNet, VGG16, MobileNet, DenseNet, and various ResNet variants). These findings demonstrate that the proposed ESA-ResNet34 model significantly enhances crop pest and disease detection.

Precision Agriculture Technologies for Early Detection of Crop Pests and Diseases

Crop pests and diseases pose significant challenges to agricultural productivity and food security worldwide. Traditional methods for detecting and managing these threats often rely on manual scouting and blanket pesticide applications, which can be labor-intensive, time-consuming, and environmentally harmful. Precision agriculture technologies offer promising solutions for early detection and targeted management of crop pests and diseases. This review article provides a comprehensive overview of the latest precision agriculture tools and techniques for monitoring crop health, detecting pests and diseases, and guiding site-specific interventions. Key technologies discussed include remote sensing, proximal sensing, machine learning, robotics, and Internet of Things (IoT) sensors. The article highlights the potential of these technologies to improve the timeliness, accuracy, and efficiency of pest and disease detection while reducing reliance on chemical inputs. It also discusses the challenges and opportunities for integrating these technologies into current agricultural practices and extension services. The review concludes with recommendations for future research and development to advance precision agriculture solutions for sustainable crop protection.

Development of a cutting-edge ensemble pipeline for rapid and accurate diagnosis of plant leaf diseases

Selecting techniques is a crucial aspect of disease detection analysis, particularly in the convergence of computer vision and agricultural technology. Maintaining crop disease detection in a timely and accurate manner is essential to maintaining global food security. Deep learning is a viable answer to meet this need. To proceed with this study, we have developed and evaluated a disease detection model using a novel ensemble technique. We propose to introduce DenseNetMini, a smaller version of DenseNet. We propose combining DenseNetMini with a learning resizer in ensemble approach to enhance training accuracy and expedite learning. Another unique proposition involves utilizing Gradient Product (GP) as an optimization technique, effectively reducing the training time and improving the model performance. Examining images at different magnifications reveals noteworthy diagnostic agreement and accuracy improvements. Test accuracy rates of 99.65 %, 98.96 %, and 98.11 % are seen in the Plantvillage, Tomato leaf, and Appleleaf9 datasets, respectively. One of the research's main achievements is the significant decrease in processing time, which suggests that using the GP could improve disease detection in agriculture's accessibility and efficiency. Beyond quantitative successes, the study highlights Explainable Artificial Intelligence (XAI) methods, which are essential to improving the disease detection model's interpretability and transparency. XAI enhances the interpretability of the model by visually identifying critical areas on plant leaves for disease identification, which promotes confidence and understanding of the model's functionality.

A Convolutional Neural Network Model for Crop Disease Detection System

Crop diseases pose a significant challenge to global food security, adversely impacting agricultural output and resulting in considerable economic repercussions. The prompt and precise identification of these diseases is essential for effective intervention and sustainable agricultural practices. This study introduces a model based on Convolutional Neural Networks (CNNs) for the automated detection of crop diseases. The model employs advanced deep learning methodologies to recognize and categorize plant diseases through the analysis of leaf images. Our CNN framework is trained on an extensive dataset comprising both diseased and healthy plant images, employing multiple convolutional layers to extract intricate features, including texture, color variations, and patterns linked to specific diseases. The model demonstrates a high level of accuracy in identifying a variety of diseases across different crop species by learning from both overt symptoms and subtle cues. We evaluate the performance of the system using established metrics such as accuracy and precision, thereby validating its efficacy in practical applications. The proposed system is designed for implementation in low-resource agricultural settings, offering farmers a cost-effective, dependable, and real-time solution for monitoring crop health.

Towards robust crop disease detection for complex real field background images

Most of the work done in image processing-based crop disease detection focuses on images with plain background. This paper presents a technique for crop disease detection for complex real field background images. A segmentation technique is presented to extract leaf patches from the entire image. Transform domain cepstral analysis is proposed for obtaining cepstral coefficients, to attain two level classifications. The first level classifies the crop species while the second level classifies the species into healthy leaf or leaf with specific type of disease. The work is tested on three crops Banana, Soybean and Grape and is checked on plain background laboratory images and on complex real field images. Suggested technique give species level accuracy of 94.33 %, 94.11 % and 98.44 % and disease level average accuracy of 97.75 %, 96.66 % and 97.95 % for Banana, Soybean and Grape, respectively. Comparison with standard features like texture and shape indicate that the presented technique gives the best results for both plain and complex background images suggesting its utilization in crop disease detection to reduce the agricultural and economic losses.

Investigating Transfer Learning based CNN for Multi-Crop Disease Recognition in Horticulture Crops using Wireless Multimedia Sensor Network

Introduction: Achieving sustainable growth in agricultural productivity is crucial for addressing various challenges confronting food systems globally. One of the key aspects is effectively identifying and addressing diseases that can adversely affect crop quality and yield. Method: This research study proposes the use of Wireless Multimedia Sensor Network (WMSN) in conjunction with deep transfer learning model for disease detection in horticulture crops. The proposed system utilizes the image data of horticulture crops obtained using WMSN for performing a comparative assessment of several pre-trained deep learning models, encompassing VGG, ResNet, MobileNet, Inception, Xception, DenseNet, and EfficientNet, to ascertain their effectiveness in identifying diseases across a range of horticultural plants. The models undergo fine-tuning using transfer learning technique and are assessed on different measures, such as accuracy, precision, recall, F1 score, and the number of epochs needed for convergence using the image data obtained from WMSN. Results: Among the compared models, EfficientNetB3 appears as the best performing model, surpassing others in all evaluation metrics while also exhibiting low parameters and a compact size. Notably, EfficientNetB3 achieves a testing accuracy of 93.69% and a training accuracy of 98.85% after only 21 epochs of training. In contrast, DenseNet201, DenseNet169, and EfficientNetB0 demonstrate performance close to the best-performing models, but they have higher parameters or sizes. On the contrary, ResNet50, InceptionV3, and Xception exhibit poorer performance, presumably due to their substantial size in relation to the dataset. Conclusion: This analysis provides valuable insights for optimizing the top-performing models for disease classification across various plant species. Moreover, the fine-tuned EfficientNetB3 model shows potential for integration with mobile devices, facilitating real-time disease classification.

A Summative Review of Advances in Sensor Technology for Precision Agriculture

Sensor technology has become a cornerstone of precision agriculture, offering a wide array of applications that enhance farming efficiency, productivity, and sustainability. By providing real-time data on soil health, crop growth, and environmental conditions, sensors enable farmers to optimize inputs such as water, fertilizers, and pesticides, reducing waste and improving yields. The development of advanced sensors, such as multi-spectral and hyper-spectral imaging, combined with artificial intelligence (AI) and machine learning (ML) algorithms, has revolutionized crop monitoring and disease detection, allowing for timely interventions that prevent significant losses. Additionally, the integration of wireless sensor networks (WSNs) and the Internet of Things (IoT) facilitates the seamless collection and transmission of data, enabling remote and automated farm management. This technology has proven instrumental in addressing the challenges of food security by increasing agricultural productivity and reducing input costs. Sensor technologies also promote sustainable land and water management by improving irrigation efficiency and reducing chemical runoff, thus minimizing environmental impact. Economic benefits are substantial, with farmers experiencing cost savings, increased yields, and higher profitability. Moreover, sensor technology contributes to climate-smart agriculture by improving resource use efficiency and reducing greenhouse gas emissions. However, challenges remain, including the high cost of sensor installation and maintenance, limited accessibility in remote regions, data privacy concerns, and the lack of interoperability between different sensor platforms. The future of sensor technology lies in the development of low-cost, biodegradable sensors, the increased use of autonomous and robotic platforms, and enhanced AI integration for more accurate data interpretation. Expanding the use of sensors in smallholder and developing world agriculture will be crucial for global food security and environmental sustainability, offering a promising path toward a more resilient and efficient agricultural system.

Spectral data driven machine learning classification models for real time leaf spot disease detection in brinjal crops

This study presents the development and evaluation of machine learning models for detecting leaf spot disease in brinjal crops using spectral sensor data. The spectral reflectance of diseased and healthy tissues was recorded across nine wavelength bands (F1: 415 nm, F2: 445 nm, F3: 480 nm, F4: 515 nm, F5: 555 nm, F6: 590 nm, F7: 630 nm, F8: 680 nm, and F9: NIR-750 nm). The data revealed distinct spectral signatures, particularly between F5 (555 nm) and F9 (NIR), where diseased tissues consistently showed lower reflectance compared to healthy tissues. Two machine learning algorithms, Decision Tree (DT) and Support Vector Machine (SVM), were employed to classify the spectral data. The DT model achieved a maximum testing accuracy of 88.2 %, with a Gini index and a depth of 4 as optimal hyperparameters. The confusion matrix indicated that the DT model correctly identified 883 diseased instances and 667 healthy cases, while misclassifying 213 healthy tissues as diseased and 25 diseased tissues as healthy. The SVM model, configured with a cost parameter of 10.0 and a tolerance of 0.01, outperformed the DT model, achieving a testing accuracy of 92.4 %. The SVM model correctly classified 99.3 % of diseased instances and 94.1 % of healthy cases. The results demonstrate the potential of spectral sensor data combined with ML algorithms for precise disease detection, facilitating targeted pesticide application, and reducing input costs. The high accuracy of the SVM model underscores its utility in agricultural disease management, enabling early intervention and enhancing crop health monitoring. Future research may explore integrating multiple sensors and advanced feature extraction methods to further improve the efficiency and accuracy of these systems.

A Review of the Application of Hyperspectral Imaging Technology in Agricultural Crop Economics

China is a large agricultural country, and the crop economy holds an important place in the national economy. The identification of crop diseases and pests, as well as the non-destructive classification of crops, has always been a challenge in agricultural development, hindering the rapid growth of the agricultural economy. Hyperspectral imaging technology combines imaging and spectral techniques, using hyperspectral cameras to acquire raw image data of crops. After correcting and preprocessing the raw image data to obtain the required spectral features, it becomes possible to achieve the rapid non-destructive detection of crop diseases and pests, as well as the non-destructive classification and identification of agricultural products. This paper first provides an overview of the current applications of hyperspectral imaging technology in crops both domestically and internationally. It then summarizes the methods of hyperspectral data acquisition and application scenarios. Subsequently, it organizes the processing of hyperspectral data for crop disease and pest detection and classification, deriving relevant preprocessing and analysis methods for hyperspectral data. Finally, it conducts a detailed analysis of classic cases using hyperspectral imaging technology for detecting crop diseases and pests and non-destructive classification, while also analyzing and summarizing the future development trends of hyperspectral imaging technology in agricultural production. The non-destructive rapid detection and classification technology of hyperspectral imaging can effectively select qualified crops and classify crops of different qualities, ensuring the quality of agricultural products. In conclusion, hyperspectral imaging technology can effectively serve the agricultural economy, making agricultural production more intelligent and holding significant importance for the development of agriculture in China.

A novel approach for detecting Pomegranate leaf pathologies using deep learning

Pomegranate cultivation is essential for ensuring global food security, yet it faces significant threats from diseases. This has resulted in significant losses of yield. Integrated disease management strategies are therefore crucial for addressing this problem and include cultural practices, disease resistant varieties and timely application of fungicides. Artificial intelligence (AI), primarily deep learning, has recently made great strides that can help change the landscape on how agricultural diseases are diagnosed. In this study, therefore, we will develop a highly accurate and efficient system for early detection of diseases in pomegranate crops using deep learning algorithms. Using convolutional neural networks (CNNs) along with transfer learning coupled with data augmentation techniques such as rotation, flipping and scaling automate the identification and classification of disorders related to leaves growing on pomegranate plants. Rigorous experimentation and evaluation using performance metrics like accuracy, precision, recall, and F1 score demonstrate the efficacy of the CNN models. The underlying theory behind this is that AI technologies would be able to successfully detect all BLB diseases associated with pomegranates at an accuracy rate of 96.33 percent using Inception v3 CNN model. These results point out that AI technology changes everything about dealing with crop diseases in agriculture while also emphasizing its importance towards global food security efforts. This study not only presents a novel approach to disease detection but also provides practical solutions for farmers to mitigate crop losses and improve yield quality, promoting more sustainable and resilient farming.

Zero shot plant disease classification with semantic attributes

In the rapidly evolving field of plant disease detection, the number and complexity of crop diseases are increasing, made worse by factors like climate change. Addressing these challenges requires robust and efficient methodologies capable of early and accurate disease identification. This paper explores the integration of advanced deep learning techniques, including pre-trained models, zero-shot learning, and semantic attributes to enhance the effectiveness of plant disease detection systems. High level features extracted from the images by these pretrained models capture crucial patterns, while domain-specific semantic attributes, such as leaf texture and color variations, enhance the understanding. Incorporating zero-shot learning enables adaptation to new and unseen diseases using semantic descriptions. Experimental validation across diverse plant species and disease types underscores the approach’s reliability in real-world agricultural scenarios. Our approach has demonstrated superior performance with plant village dataset, showing a significant improvement in accuracy and generalization. These results underscore the potential of our method to revolutionize plant disease detection and management in agricultural practices.

Image‐based crop disease detection using machine learning

Image‐based crop disease detection using machine learning

Multi-spectral image transformer descriptor classification combined with molecular tools for early detection of tomato grey mould

Accurate early detection of plant diseases is a critical milestone in the development of self-navigating robots for precision agriculture and still remains one of the most significant challenges. Plant pathogens, like Botrytis cinerea causing grey mould disease, pose significant threats to agriculture and food safety. This study focuses on the early detection of B. cinerea in tomato crops using Artificial Intelligence. Specifically, the Deep Learning (DL) YOLOv8 architecture was employed to apply single class segmentation on plant captures to extract the tomato leaves locational information. Each plant capture includes five images from hyperspectral wavelengths (at 460, 540, 640, 775 and 875 nm) and one RGB images. The leaf segmentation achieved 81.7 % Mean Average Precision (mAP) at Intersection over Union (IoU) threshold 0.5 (mAP50). Then the for each leaf segment, and array of descriptors is extracted through various Transformer models. Finally, the descriptors are classified through a KNN or an LSTM solution and the results of all the descriptors for each leaf from each of the Transformer models, for each of the six images of the capture, are ensembled to yield the class of each leaf. The best ensemble results were produced by only accumulating results from some of the Transformer and specifically the MaxViT-L, the ViT-B (P:16 × 16 – C), the VOLO (D:5) and the XCIT-L (L:24 – P:16 × 16), by only using the results provided by the images at the hyperspectral wavelengths at 540, 640 nm and the RGB image, and by applying the LSTM solution. The process achieved a 79.41 % classification accuracy and 71.12 % F1-Score for all classes. Additionally, early, rapid and accurate detection of grey mould at the early stages of the infection or at latent infections when symptoms are not visible was assessed by comparative qPCR-based methods (qRT-PCR) for fungal biomass estimation in plant tissues and the rest of this study's methodology. The findings of this study represent a significant advancement in crop disease detection and management, and highlight the potential for integrating these methods into on-site digital systems (robots, mobile apps, etc.) under real-world settings. The results, demonstrate the effectiveness of combining segmentation and transformer-based classification models for early and accurate detection of grey mould disease.

Assessment of Dataset Scalability for Classification of Black Sigatoka in Banana Crops Using UAV-Based Multispectral Images and Deep Learning Techniques

This paper presents an evaluation of different convolutional neural network (CNN) architectures using false-colour images obtained by multispectral sensors on drones for the detection of Black Sigatoka in banana crops. The objective is to use drones to improve the accuracy and efficiency of Black Sigatoka detection to reduce its impact on banana production and improve the sustainable management of banana crops, one of the most produced, traded, and important fruits for food security consumed worldwide. This study aims to improve the precision and accuracy in analysing the images and detecting the presence of the disease using deep learning algorithms. Moreover, we are using drones, multispectral images, and different CNNs, supported by transfer learning, to enhance and scale up the current approach using RGB images obtained by conventional cameras and even smartphone cameras, available in open datasets. The innovation of this study, compared to existing technologies for disease detection in crops, lies in the advantages offered by using drones for image acquisition of crops, in this case, constructing and testing our own datasets, which allows us to save time and resources in the identification of crop diseases in a highly scalable manner. The CNNs used are a type of artificial neural network widely utilised for machine training; they contain several specialised layers interconnected with each other in which the initial layers can detect lines and curves, and gradually become specialised until reaching deeper layers that recognise complex shapes. We use multispectral sensors to create false-colour images around the red colour spectra to distinguish infected leaves. Relevant results of this study include the construction of a dataset with 505 original drone images. By subdividing and converting them into false-colour images using the UAV’s multispectral sensors, we obtained 2706 objects of diseased leaves, 3102 objects of healthy leaves, and an additional 1192 objects of non-leaves to train classification algorithms. Additionally, 3640 labels of Black Sigatoka were generated by phytopathology experts, ideal for training algorithms to detect this disease in banana crops. In classification, we achieved a performance of 86.5% using false-colour images with red, red edge, and near-infrared composition through MobileNetV2 for three classes (healthy leaves, diseased leaves, and non-leaf extras). We obtained better results in identifying Black Sigatoka disease in banana crops using the classification approach with MobileNetV2 as well as our own datasets.

Application of a Real-Time Field-Programmable Gate Array-Based Image-Processing System for Crop Monitoring in Precision Agriculture

Precision agriculture (PA) technologies combined with remote sensors, GPS, and GIS are transforming the agricultural industry while promoting sustainable farming practices with the ability to optimize resource utilization and minimize environmental impact. However, their implementation faces challenges such as high computational costs, complexity, low image resolution, and limited GPS accuracy. These issues hinder timely delivery of prescription maps and impede farmers’ ability to make effective, on-the-spot decisions regarding farm management, especially in stress-sensitive crops. Therefore, this study proposes field programmable gate array (FPGA)-based hardware solutions and real-time kinematic GPS (RTK-GPS) to develop a real-time crop-monitoring system that can address the limitations of current PA technologies. Our proposed system uses high-accuracy RTK and real-time FPGA-based image-processing (RFIP) devices for data collection, geotagging real-time field data via Python and a camera. The acquired images are processed to extract metadata then visualized as a heat map on Google Maps, indicating green area intensity based on romaine lettuce leafage. The RFIP system showed a strong correlation (R2 = 0.9566) with a reference system and performed well in field tests, providing a Lin’s concordance correlation coefficient (CCC) of 0.8292. This study demonstrates the potential of the developed system to address current PA limitations by providing real-time, accurate data for immediate decision making. In the future, this proposed system will be integrated with autonomous farm equipment to further enhance sustainable farming practices, including real-time crop health monitoring, yield assessment, and crop disease detection.

Deep Learning–Based Enhanced Optimization for Automated Rice Plant Disease Detection and Classification

ABSTRACTEnsuring global food security necessitates innovative solutions for early detection and precise classification of diseases in staple crops like rice. This study introduces an advanced approach for automated rice plant disease detection and classification by integrating deep learning and metaheuristic optimization techniques. Specifically, a deep dense neural network (DNN) is employed for its capacity to capture intricate patterns in images and extreme learning machine (ELM) for classification. To enhance the optimization process, an innovative variant of the Shuffled Shepherd Optimization (SSO) algorithm, known as Enhanced Artificial Shuffled Shepherd Optimization (EASSO), is introduced. EASSO incorporates adaptive strategies and enhanced exploration–exploitation mechanisms, enabling more efficient convergence during the training of the DNN. The proposed system operates by processing high‐resolution images of rice plants through the DNN, extracting nuanced features indicative of various diseases, including blast, bacterial blight, and brown spots. EASSO optimizes the DNN's parameters, maximizing its accuracy in disease classification. The synergy between DNN and EASSO ensures a robust and adaptive model capable of handling diverse and complex disease patterns. This automated approach significantly reduces the reliance on manual inspection, enabling timely intervention and improving overall agricultural productivity. Experimental results demonstrate the superiority of the DNN‐EASSO framework over traditional methods, showcasing higher accuracy rates and faster convergence. The incorporation of Enhanced Artificial Shuffled Shepherd Optimization enhances the precision and reliability of disease classification, making this integrated system a valuable tool for farmers and agricultural practitioners. This research represents a significant stride toward sustainable agriculture, showcasing the potential of advanced technologies in ensuring food security worldwide.

Detection of Rice Leaf SPAD and Blast Disease Using Integrated Aerial and Ground Multiscale Canopy Reflectance Spectroscopy

Rice blast disease is one of the major diseases affecting rice plant, significantly impacting both yield and quality. Current detecting methods for rice blast disease mainly rely on manual surveys in the field and laboratory tests, which are inefficient, inaccurate, and limited in scale. Spectral and imaging technologies in the visible and near-infrared (Vis/NIR) region have been widely investigated for crop disease detection. This work explored the potential of integrating canopy reflectance spectra acquired near the ground and aerial multispectral images captured with an unmanned aerial vehicle (UAV) for estimating Soil-Plant Analysis Development (SPAD) values and detecting rice leaf blast disease in the field. Canopy reflectance spectra were preprocessed, followed by effective band selection. Different vegetation indices (VIs) were calculated from multispectral images and selected for model establishment according to their correlation with SPAD values and disease severity. The full-wavelength canopy spectra (450–850 nm) were first used for establishing SPAD inversion and blast disease classification models, demonstrating the effectiveness of Vis/NIR spectroscopy for SPAD inversion and blast disease detection. Then, selected effective bands from the canopy spectra, UAV VIs, and the fusion of the two data sources were used for establishing corresponding models. The results showed that all SPAD inversion models and disease classification models established with the integrated data performed better than corresponding models established with the single of either of the aerial and ground data sources. For SPAD inversion models, the best model based on a single data source achieved a validation determination coefficient (Rcv2) of 0.5719 and a validation root mean square error (RMSECV) of 2.8794, while after ground and aerial data fusion, these two values improved to 0.6476 and 2.6207, respectively. For blast disease classification models, the best model based on a single data source achieved an overall test accuracy of 89.01% and a Kappa coefficient of 0.86, and after data fusion, the two values improved to 96.37% and 0.95, respectively. These results indicated the significant potential of integrating canopy reflectance spectra and UAV multispectral images for detecting rice diseases in large fields.

The Influence of Parallel Computing on Building Deep Learning Model for the Classification of Bean Diseases

In recent years, the utilization of deep learning techniques for image classification has made significant strides in the field of agriculture. One of the key areas of interest in agriculture is the early detection and classification of diseases in crops, as this can have an insightful impact on crop revenue and quality. This research has investigated the influence of parallel computing on the performance of a deep learning-based classification model for diagnosing bean diseases. Specifically, we have explored the use of parallel computing frameworks to accelerate model training and inference, thereby enhancing the efficiency and effectiveness of disease classification. Our findings demonstrated the potential for parallel computing to accelerate model training. When training a bean disease classification model, we achieved an accuracy of 0.93 using parallel computing, compared to 0.83 with serial computing. Moreover, parallel computing significantly reduced training time, taking only 3 minutes compared to 51 minutes with serial computing.

Near‐infrared spectroscopy and deep neural networks for early common root rot detection in wheat from multi‐season trials

AbstractIn Australia, the soil‐borne disease common root rot (Bipolaris sorokiniana) (CRR) in wheat (Triticum aestivum L.) leads to substantial yield losses, yet has limited visible aboveground symptoms, making detection and identification labor intensive. Near‐infrared (NIR) spectroscopy offers an early potential identification solution for CRR in wheat and has previously been reported with success for crop disease detection. This study investigated the ability of nondestructive NIR spectroscopy in combination with deep neural networks (DNN), logistic regression (LR), and principal component analysis combined with support vector machines (PCA‐SVM) for early‐stage CRR detection in wheat. NIR spectra of five different wheat varieties with varying resistance to CRR were collected in two seasons of glasshouse and three seasons of field trials using a portable spectrometer. Results demonstrated that DNN outperformed LR and PCA‐SVM, achieving 66%–91% average classification accuracy in glasshouse trials and an average accuracy of 73% with up to 87% in field trials, effectively distinguishing inoculated and non‐inoculated wheat plants from seedling to anthesis stages. Validation with a third season of field data achieved an average of 77% accuracy for the most susceptible variety during the stem elongation stage. NIR reflectance within 1600–1700 nm was identified as most important for estimating CRR presence, with initial detection occurring 35 days after sowing (DAS) in the glasshouse and 46 DAS in the field. In conclusion, a NIR spectrometer with a DNN model successfully performed disease classification, with the potential as a portable early disease detection tool to assist farm management decisions.

Lightweight citrus leaf disease detection model based on ARMS and cross-domain dynamic attention

In citrus cultivation, Anthracnose, Scab, and Greasy Spot significantly impact yield and quality. Facing challenges in detecting small targets against complex orchard backgrounds with uneven lighting and obstructions, existing models suffer from low detection accuracy. This study introduces the YOLOv8n-CDDA citrus leaf disease detection model. The Cross-Domain Dynamic Attention (CDDA) mechanism deconstructs the backbone network’s input feature maps into sections, dynamically assigning spatial and channel attention weights to reconstruct critical information and capture the variations and weak semantic features of disease textures. The proposed Adaptive Random Mix-Cut Splicing (ARMS) image augmentation technique blends diseased leaf images with healthy citrus backgrounds, enhancing the diversity and number of background targets. To reduce computational and memory consumption, the network is streamlined through channel pruning; to compensate for the loss in accuracy from pruning, a teacher–assistant–student network format is used for knowledge distillation, where the student network learns from soft knowledge to improve disease recognition accuracy. Finally, Grad-CAM++ technology generates heatmaps of the detections, facilitating the visualization of effective features and deepening understanding of the model’s focus areas. Experimental results demonstrate that the YOLOv8n-CDDA model achieves an average accuracy of 90.89% in disease detection, with an average recall rate of 81.12%, and a mean Average Precision (mAP50) of 88.36%. Compared to the original YOLO v8n and current mainstream detection models such as YOLOv5s, SSD, and Faster-RCNN, the improvements in average accuracy are respectively 2.95%, 4.78%, 14.22%, and 21.01%; in average recall, 2.36%, 3.09%, 15.74%, and 23.27%; and in mAP50, 2.38%, 3.13%, 13.45%, and 20.91%. After pruning and distillation for lightweight adaptation, the YOLOv8n-CDDA model has a parameter size of 0.8M, requires 4.2 GFLOPs, weighs 2.0 MB, and operates at 45 fps. Compared to YOLOv8n, this represents a reduction of 2.2M in parameters, 3.9 GFLOPs, and 4 MB in model weight, with an increase of 7 fps in speed. This model exhibits exceptional performance in the complex environment of citrus leaf disease detection, providing robust technical support for citrus growth monitoring studies, and offering insights for disease detection in other crops as well.