相关参考文献
注意:仅列出部分参考文献,下载原文获取全部文献信息。
Article
Multidisciplinary Sciences
Nikita Andriyanov et al.
Summary: Despite the potential of modern neural network architectures in object detection and recognition, additional information about objects is needed for practical tasks. To accurately determine the position of apples for robotic apple picking, the Intel Real Sense depth camera is used to aggregate depth and brightness information. The proposed approach achieves high position estimation accuracy with minimal error.
Article
Optics
N. A. Andriyanov et al.
Summary: The relevance of object detection and recognition tasks in images has been increasing over the years. This paper systematically analyzes the trends in the development of detection methods, the reasons behind these developments, and the metrics used to assess the quality and reliability of object detection. The paper discusses detection techniques based on mathematical models and convolutional neural networks, as well as typical tests used to compare different algorithms.
Article
Agriculture, Multidisciplinary
Longsheng Fu et al.
Summary: Automatic detection of kiwifruit in the orchard is challenging due to varying illumination and color similarity with the complex background. A deep YOLOv3-tiny model (DY3TNet) was developed for efficient and accurate detection, achieving high precision and fast processing speed, especially in images captured with flash.
PRECISION AGRICULTURE
(2021)
Article
Vadim Mizonov et al.
JP Journal of Heat and Mass Transfer
(2021)
Article
Agriculture, Multidisciplinary
Przemyslaw Dolata et al.
Summary: Modern agriculture relies on control and optimization, with monitoring as a key component. Traditional yield monitoring based on spatial mapping of biomass is not effective for uniformity optimization. Developing a method to estimate crop physical dimensions online is crucial for improving the efficiency of precision agriculture, especially for plants like root crops that are difficult to observe directly through monocular vision.
COMPUTERS AND ELECTRONICS IN AGRICULTURE
(2021)
Article
Agronomy
Shunfu Xiao et al.
Summary: The study highlights the importance of selecting and breeding crop varieties with high economic benefits. By utilizing the SFM-MVS method, plant phenotypic traits can be efficiently extracted to estimate the economic benefits of crops. This approach offers a more efficient and accurate method compared to manual measurement, providing a basis for selecting and cultivating crops with high economic benefits.
EUROPEAN JOURNAL OF AGRONOMY
(2021)
Article
Engineering, Mechanical
Qingqi Zhang et al.
Summary: This paper proposes a water meter pointer reading recognition method based on target-key point detection, which utilizes modified networks and algorithms to quickly and accurately identify the pointer and key points on water meter images, achieving automated detection.
FLOW MEASUREMENT AND INSTRUMENTATION
(2021)
Article
Construction & Building Technology
Zhenwei Yu et al.
Summary: The YOLOv4-FPM model improves accuracy and speed by using focal loss and pruning algorithm to optimize the loss function and accelerate detection speed. It expands the detection range and is able to effectively detect cracks in images of different sizes.
AUTOMATION IN CONSTRUCTION
(2021)
Article
Engineering, Electrical & Electronic
Quoc-Khanh Huynh et al.
Summary: This study developed a method to identify and classify cracked chilli fruits caused by the destemming process, utilizing a convolutional neural network model for high accuracy in both static and working conditions.
JOURNAL OF SENSORS
(2021)
Article
Agriculture, Multidisciplinary
Youliang Ni et al.
Summary: A soybean harvester header-height adjustment system was developed in this study, which effectively controls the header height and improves harvesting efficiency through experiments and model analysis. The system has been shown to have good accuracy in controlling the header height, leading to improved harvesting performance.
COMPUTERS AND ELECTRONICS IN AGRICULTURE
(2021)
Article
Chemistry, Multidisciplinary
Hongchao Zhuang et al.
Summary: The study focuses on improving the speed and accuracy of motion blurred gestures recognition through image processing and model optimization, resulting in a model with better performance in real-time interaction of blurred gestures, demonstrating a 30% increase in network training speed, 10% increase in target detection accuracy, and approximately 10% increase in mAP value.
APPLIED SCIENCES-BASEL
(2021)
Article
Agronomy
Gianni Fenu et al.
Summary: The research focuses on the classification and recognition of foliar diseases using machine and deep learning to support agricultural stakeholders. A dataset called DiaMOS Plant is released for diagnosing and monitoring plant symptoms. The study also includes a comparative analysis of existing literature datasets designed for leaf disease classification and recognition, highlighting key features to maximize data value and information content.
Article
Agronomy
Sergey Alekseevich Korchagin et al.
Summary: The article discusses the problem of detecting sick or mechanically damaged potatoes using machine learning methods. An algorithm was proposed and developed to rapidly detect damaged tubers, with the system capable of detecting up to 100 tubers in one second. The system's accuracy reaches 97% under optimal settings, with detection methods varying in outcomes from 80% to 97%.
Article
Agronomy
S. K. Mahmudul Hassan et al.
Summary: Various plant diseases pose threats to agriculture, and automated disease identification is beneficial for timely control. Two methods were proposed to identify plant diseases, with the shallow VGG with Xgboost model showing superior performance in accuracy.
Article
Agronomy
Lifa Fang et al.
Summary: This study utilizes object detection techniques in deep learning to propose a ginger recognition network based on YOLOv4-LITE, which successfully solves the issue of consistent orientation in ginger planting, providing a technical guarantee for automated seeding.
Article
Agronomy
Dasom Seo et al.
Summary: This study aimed to develop a real-time robot monitoring system using deep learning techniques to detect the growth status and maturity of tomatoes in hydroponic greenhouses, with the system's accuracy verified through comparison with expert classification.
Article
Agronomy
Ali Mirzazadeh et al.
Summary: This study proposed a novel technique for classifying damaged crops in rapeseed field images using deep learning algorithms, achieving high accuracy. The results demonstrated the potential of deep neural networks in distinguishing and categorizing damaged crops, with a high classification accuracy.
Article
Agronomy
Faris A. Kateb et al.
Summary: This study introduces a detection mechanism named FruitDet, designed based on YOLO pipeline and implemented using DenseNet architecture, achieving superior fruit detection performance and incorporating new concepts compared to other detection models.
Article
Agronomy
Rolando Miragaia et al.
Summary: The digitization and technological transformation in agriculture has led to better quality harvests and cost savings by using advanced sensors. This study presents a tool based on Deep Learning for analyzing different plum varieties through image analysis. The system's robustness allows for accurate differentiation of plum varieties through uncontrolled photographic acquisition methods.
Article
Agronomy
Ali Shojaeipour et al.
Summary: The study suggests that utilizing biometric identification methods to replace traditional branding or ear tagging can greatly improve livestock welfare and management. By providing a large dataset of cattle face images and developing a two-stage algorithm for automated biometric identification, the research achieved excellent model performance and highlighted the potential for advanced livestock biometric monitoring in future decision support systems for agriculture.
Article
Mathematics
Sergey Korchagin et al.
Summary: The authors modeled the electrophysical properties of composite media and investigated the frequency dependences of the dielectric constant on different inclusions. They also modeled a reflecting screen based on nanostructured composite materials and studied the influence of fractality on optical properties, finding that the proposed structure can increase the operating frequency range and efficiency of the reflecting screen.
Article
Mathematics
Sergey Korchagin et al.
Summary: Composite materials made of dielectric matrix and conductive inclusions are widely used in various systems, with their properties influenced by multiple factors. The anisotropic form of inclusions causes optical anisotropy. This article proposes a model and method for describing the electrical conductivity of layered nanocomposites, removing restrictions related to setting the dielectric constant.
Article
Agronomy
Abozar Nasirahmadi et al.
Summary: This study aimed to determine if a digital two-dimensional imaging system coupled with convolutional neural network (CNN) techniques could be utilized to detect visible mechanical damage in sugar beet during harvesting in a harvester machine. Various detector models based on CNN were developed, and it was found that the YOLO v4 CSPDarknet53 method performed better in detecting sugar beet damage with higher speed compared to other developed CNNs.
Article
Agronomy
Lifa Fang et al.
Summary: An improved YOLOv3 model is proposed in this study, which achieves real-time detection of ginger shoots and seeds by pruning redundant channels and network layers, significantly improving detection speed while maintaining high accuracy.
Article
Environmental Sciences
Bin Yan et al.
Summary: This study proposed an improved YOLOv5s algorithm for apple target recognition in apple picking robots. Experimental results show that the algorithm successfully differentiates between occluded graspable and ungraspable apples, with improved performance and speed.
Article
Agronomy
Chao Qi et al.
Summary: The paper introduces a highly fused and lightweight detection model called F-YOLO for detecting the flowering stage of tea chrysanthemum. By utilizing different networks and fusion modules to enhance gradient flow differences, it achieves superior object detection results compared to state-of-the-art technologies.
Article
Engineering, Ocean
Yan Li et al.
Summary: Zooplankton are important in the global marine carbon cycle and can serve as indicators of aquatic health, providing early warning for natural disasters. With the advancement of observation sensors and platforms, deep neural networks are being pursued for in situ and autonomous zooplankton observation. This paper introduces an improved YOLOV3 model with densely connected structures to enhance feature reusability during transmission, demonstrating superior performance for in situ zooplankton detection compared to other state-of-the-art models.
APPLIED OCEAN RESEARCH
(2021)
Article
Chemistry, Analytical
Liangliang Zheng et al.
Summary: The novel preprocessing algorithm proposed in this paper can simultaneously smooth noise and enhance edges, improving the visual quality of remote sensing images. It outperforms existing methods both visually and quantitatively, with flexible and adjustable processing parameters for different images. This algorithm can play an important role in the remote sensing field to provide more information on interested targets.
Article
Chemistry, Multidisciplinary
Shyang-Jye Chang et al.
Summary: A deep learning algorithm was developed to detect defects in coffee beans, achieving an accuracy rate of 95.2% by using a novel dimensionality reduction method for feature extraction. The proposed model showed high accuracy in detecting coffee bean defects among eight types of coffee beans.
APPLIED SCIENCES-BASEL
(2021)
Article
Agronomy
Jiacheng Rong et al.
Summary: The proposed oyster-mushroom-harvesting robot in this paper can autonomously harvest oyster mushrooms in the greenhouse with high success rates in mushroom recognition and harvesting. Field experiments demonstrate the feasibility and robustness of the system, showcasing its efficient mushroom recognition and harvesting capabilities.
Article
Agronomy
Jannis Machleb et al.
Summary: The study designed a new mechanical weeding tool combined with a bi-spectral camera for intrarow weeding in sugar beets. The system showed good weed control efficacy and can reduce herbicide use effectively. Mechanical weeding, while posing risks, is crucial for addressing herbicide resistance and adaptability.
Article
Agronomy
Adel Ammar et al.
Summary: This paper proposes an original deep learning framework for automated counting and geolocation of palm trees from aerial images using convolutional neural networks. YOLOv4 and EfficientDet-D5 models achieved the best trade-off between accuracy and speed. Geotagged metadata and photogrammetry concepts were utilized for automatic geolocation of detected palm trees.
Article
Agronomy
Maxim Tatarintsev et al.
Summary: It is a state's priority task to analyze the rise in consumer goods prices and regulate pricing, especially for essential commodities like agricultural products. This study focused on analyzing the price changes of sugar during a pandemic and forecasting its impact on production. The forecast model, based on ARIMA time series, successfully predicted the volume of domestic sugar transportation and was validated with actual data.
Article
Agronomy
Andre Silva Aguiar et al.
Summary: This study used deep learning to detect grape bunches in vineyards, training and deploying two state-of-the-art single-shot multibox models on a low-cost, low-power hardware device to achieve satisfactory performance. Experimental results showed that the models performed better in identifying grape bunches at the medium growth stage.
Article
Agronomy
Baohua Yang et al.
Summary: The detection and counting of wheat ears is crucial for crop management. A new CBAM-YOLOv4 model was proposed in this study to accurately detect and count wheat ears in field conditions, providing technical support for the extraction of other crop parameters.
Article
Mathematics
Vladimir Sudakov
Summary: This work focuses on developing new methods to support decision-making in air travel planning using uncertainties in the form of fuzzy numbers, addressing the reduced demand for air travel due to the pandemic and the shift from large to smaller aircraft types. The problem is solved by restoring the fuzzy origin-destination matrix of air traffic statistics and developing algorithms and software for finding potentially promising routes. An example task on choosing new routes between regional airports demonstrates the potential of this approach.
Proceedings Paper
Computer Science, Artificial Intelligence
Nikita Andriyanov et al.
Summary: The paper proposes an algorithm that combines the YOLO v3 convolutional neural network, doubly stochastic filters, and pseudo-gradient procedures for trajectory tracking of a large number of objects on video images in real time. The obtained numerical performance characteristics demonstrate the consistency of such a combination and its potential for application in real video processing systems.
INTELLIGENT DECISION TECHNOLOGIES, KES-IDT 2021
(2021)
Proceedings Paper
Computer Science, Artificial Intelligence
Aleksey Osipov et al.
Summary: This study proposed an algorithm for detecting mechanically damaged potatoes, which combines Viola-Jones method and convolutional networks, achieving a 92.1% accuracy rate for recognizing damaged tubers.
ARTIFICIAL INTELLIGENCE AND SOFT COMPUTING (ICAISC 2021), PT I
(2021)
Proceedings Paper
Computer Science, Artificial Intelligence
Shen Zheng et al.
Summary: Deblur-YOLO is an efficient, YOLO-based, detection-driven approach robust to motion blur photographs. By introducing a generative adversarial network and a new image quality metric, Deblur-YOLO demonstrates superiority on benchmark datasets.
2021 INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS (IJCNN)
(2021)
Article
Economics
Vera Ivanyuk
Summary: The study developed a dynamic model for managing a strategic investment portfolio considering the impact of crisis processes on asset value. It proposed guidelines for framing a long-term investment strategy and elaborated on an efficient method for forecasting financial time series accurately. The methodology for constructing and rebalancing a dynamic strategic investment portfolio based on changing strategies was also developed and empirically estimated.
Article
Computer Science, Artificial Intelligence
Ssu-Han Chen et al.
Summary: This research developed a defect detector based on YOLOv3 for inspecting SMD LED chips, identifying various types of defects. By replacing Darknet-53 with DenseNet and using the Taguchi method for hyper-parameter evaluation, the performance of the defect detector was significantly improved. The testing mAP of YOLOv3-dense was 33.69% higher than CAM of CNN and 14.98% higher than traditional YOLOv3.
ADVANCED ENGINEERING INFORMATICS
(2021)
Article
Biochemical Research Methods
Junfeng Gao et al.
Article
Agronomy
Anna Kuznetsova et al.
Article
Chemistry, Analytical
Noi Quang Truong et al.
Article
Computer Science, Hardware & Architecture
Fanny Spagnolo et al.
INTEGRATION-THE VLSI JOURNAL
(2020)
Article
Engineering, Electrical & Electronic
Hui Yin et al.
Article
Computer Science, Information Systems
Xiukun Wei et al.
Article
Chemistry, Analytical
Zahra Salimi et al.
Article
Agronomy
Pawel Kolodziej et al.
INTERNATIONAL AGROPHYSICS
(2019)
Article
Computer Science, Information Systems
Zechuan Liu et al.
Proceedings Paper
Acoustics
Olivier Barnich et al.
2009 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING, VOLS 1- 8, PROCEEDINGS
(2009)