Real-Time Apple Detection System Using Embedded Systems With Hardware Accelerators: An Edge AI Application
Open accessThe DOI matches its Crossref record; 56 of the 59 references it cites were found in a registry or an index.
Abstract
Real-time apple detection in orchards is one of the most effective ways of estimating apple yields, which helps in managing apple supplies more effectively. Traditional detection methods used highly computational machine learning algorithms with intensive hardware set up, which are not suitable for infield real-time apple detection due to their weight and power constraints. In this study, a real-time embedded solution inspired from “Edge AI” is proposed for apple detection with the implementation of YOLOv3-tiny algorithm on various embedded platforms such as Raspberry Pi 3 B+ in combination with Intel Movidius Neural Computing Stick (NCS), Nvidia's Jetson Nano and Jetson AGX Xavier. Data set for training were compiled using acquired images during field survey of apple orchard situated in the north region of Italy, and images used for testing were taken from widely used google data set by filtering out the images containing apples in different scenes to ensure the robustness of the algorithm. The proposed study adapts YOLOv3-tiny architecture to detect small objects. It shows the feasibility of deployment of the customized model on cheap and power-efficient embedded hardware without compromising mean average detection accuracy (83.64%) and achieved frame rate up to 30 fps even for the difficult scenarios such as overlapping apples, complex background, less exposure of apple due to leaves and branches. Furthermore, the proposed embedded solution can be deployed on the unmanned ground vehicles to detect, count, and measure the size of the apples in real-time to help the farmers and agronomists in their decision making and management skills.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“The technology is rapidly evolving, with new techniques and frameworks being developed to improve the performance, efficiency, and security of Edge AI solutions [62].”
From Edge AI: A survey (2023) · cites this paper as [62]
Experimental — accuracy not yet measured
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Deep Residual Learning for Image Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.90 · Cited by 235,474 · cited 1×
“Darknet-19 is a light and efficient feature extractor, but can be easily swapped with any other backbone like ResNet [40], DenseNet [41], etc.”
BackgroundVery Deep Convolutional Networks for Large-Scale Image Recognition
Simonyan, K., & Zisserman, A. (2014) · arXiv · doi:10.48550/arxiv.1409.1556 · Cited by 114,148 · cited 2×
“Faster R-CNN with VGG16 net [18] is the state-of-art method in fruit detection [10].”
MethodFaster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
Ren, S., He, K., Girshick, R., & Sun, J. (2017) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2016.2577031 · Cited by 73,472 · cited 5×
“However, Faster R-CNN consists of region proposal networks (RPN) and clas-sification networks that produced excellent results in terms of accuracy, while the detection speed is slow, which can not achieve good results in real-time with high image resolution.”
MethodInfluentialFaster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
Ren, S., He, K., Girshick, R., & Sun, J. (2015) · arXiv · doi:10.48550/arxiv.1506.01497 · Cited by 70,812 · cited 5×
“However, Faster R-CNN consists of region proposal networks (RPN) and clas-sification networks that produced excellent results in terms of accuracy, while the detection speed is slow, which can not achieve good results in real-time with high image resolution.”
MethodInfluentialImageNet: A large-scale hierarchical image database
Deng, J., Dong, W., Socher, R., Li, L.-J., Li, K., & Fei-Fei, L. (2009) · 2009 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2009.5206848 · Cited by 65,333 · cited 1×
“That is not true with larger data set like ImageNet [48], where labels are arranged as a directed graph [49] and not as flat structure.”
BackgroundYou Only Look Once: Unified, Real-Time Object Detection
Redmon, J., Divvala, S., Girshick, R., & Farhadi, A. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.91 · Cited by 53,803 · cited 3×
“The You Only Look Once (YOLO) method [19], [20] deals with the classification and the localization as a regression problem.”
BackgroundMethodDensely Connected Convolutional Networks
Huang, G., Liu, Z., Van Der Maaten, L., & Weinberger, K. Q. (2017) · 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2017.243 · Cited by 48,566
Microsoft COCO: Common Objects in Context
Lin, T.-Y., Maire, M., Belongie, S., Hays, J., Perona, P., Ramanan, D., Dollár, P., & Zitnick, C. L. (2014) · Lecture Notes in Computer Science · doi:10.1007/978-3-319-10602-1_48 · Cited by 45,460
Feature Pyramid Networks for Object Detection
Lin, T.-Y., Dollar, P., Girshick, R., He, K., Hariharan, B., & Belongie, S. (2017) · 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2017.106 · Cited by 30,939 · cited 1×
“That is achieved with 1x1 convolutions arranged in a Pyramid Network [44] structure.”
MethodFocal Loss for Dense Object Detection
Lin, T.-Y., Goyal, P., Girshick, R., He, K., & Dollar, P. (2017) · 2017 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2017.324 · Cited by 30,581 · cited 1×
“It has comparable performance in terms of accuracy with other popular object detection algorithms like RetinaNet [37], Faster-RCNN [38], but it is much faster and compact that makes it an optimal choice for real-time embedded applications.”
Method
Research works printed but not linked: 6
- 2019, Graphics Reinvented NVIDIA GeForce RTX 2080 Ti Graphics Card
- 2019, NVIDIA Jetson—Hardware For Every Situation
- 2019, Intel Neural Compute Stick
- 2019, Intel movidius neural compute stick
- rosebrock, 2017, Deep Learning for Computer Vision with Python Starter Bundle
- silwal, 2016, Apple identification in field environment with over the row machine vision system, Agricultural Engineering International the CIGRE Journal, 16, 66
Cited by
Items 101–200 of 212
FruitVision: Dual-Attention Embedded AI System for Precise Apple Counting Using Edge Computing
Thakur, D., & Kumar, V. (2024) · IEEE Transactions on AgriFood Electronics · doi:10.1109/tafe.2024.3416221 · Cited by 7
“Mazzia et al. [17] proposed a real-time apple detection solution on cost-effective platforms, attaining 83.64% accuracy and 30 fps frame rate.”
BackgroundInfluentialA Scoping Review on Role of Edge Computing in Search and Rescue Operations
Savargaonkar, H. A., Patil, H. Y., Jindal, S. K., Chandra, A., & Gawas, M. A. (2024) · Wireless Personal Communications · doi:10.1007/s11277-024-11543-z · Cited by 11
“E.g. Propellors in drones [97, 98] contemplated solutions to a theoretical stochastic optimization problem with CPU frequency, transmit power, and bandwidth allocation parameters.”
Autonomous Navigation in Rows of Trees and High Crops with Deep Semantic Segmentation
Navone, A., Martini, M., Ostuni, A., Angarano, S., & Chiaberge, M. (2023) · European Conference on Mobile Robots · doi:10.1109/ecmr59166.2023.10256334 · Cited by 7
“The most recent advancements in deep learning also provide competitive advantages for real-world applications, such as model optimization for fast inference on low-power embedded hardware [2, 3] and generalization to unseen data [4, 5, 6].”
BackgroundSelf-citationHung, C., Tsai, C.-Y., & Lee, C. (2024) · IEEE Sensors Journal · doi:10.1109/jsen.2024.3397008 · Cited by 8
“Strategies include partitioning models based on complexity, leveraging AI-accelerated computing, and refining models for edge deployment [8], [9], [10], [11], [12].”
MethodHigh performance inference of gait recognition models on embedded systems
Ruiz-Barroso, P., Castro, F. M., Delgado-Escaño, R., Cózar, J. R., & Guil, N. (2022) · Sustainable Computing Informatics and Systems · doi:10.1016/j.suscom.2022.100814 · Cited by 9
Edge-Oriented Social Distance Monitoring System Based on MTCNN
Valverde, E. C., Angelo, O. P., Caliwag, A. C., Kamali, M. A., Lim, W., & Maier, M. (2022) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2022.3217499 · Cited by 7
“Furthermore, edge computing contributes to building an infrastructure of reliable, secure, and scalable systems, which resolves the issues involving data processing on a cloud computing platform [25].”
BackgroundAn Adaptive Row Crops Path Generator with Deep Learning Synergy
Cerrato, S., Aghi, D., Mazzia, V., Salvetti, F., & Chiaberge, M. (2021) · Asia-Pacific Conference on Intelligent Robot Systems (ACIRS) · doi:10.1109/acirs52449.2021.9519316 · Cited by 7
Self-citationLSTM for State of Charge Estimation of Lithium Polymer Battery on Jetson Nano
Hantono, B. S., Cahyadi, A. I., & Pratama, G. N. P. (2021) · doi:10.1109/icitee53064.2021.9611940 · Cited by 7
Shi, S., Jiang, Q., Jin, X., Wang, W., Liu, K., Chen, H., Liu, P., Zhou, W., & Yao, S. (2023) · Frontiers in Neurorobotics · doi:10.3389/fnbot.2023.1143032 · Cited by 6
“playing a comparative role in the experiment (Mazzia et al., 2020).”
BackgroundReal-time multiple people gait recognition in the edge
Ruiz-Barroso, P., González-Linares, J. M., Castro, F. M., & Guil, N. (2025) · Scientific Reports · doi:10.1038/s41598-025-02351-x · Cited by 8
Revolutionizing machine learning: Harnessing hardware accelerators for enhanced AI efficiency
Zou, C. (2024) · Applied and Computational Engineering · doi:10.54254/2755-2721/47/20241256 · Cited by 6
“These hardware accelerators, including Graphics Processing Units (GPUs), Tensor Processing Units (TPUs), and Application Specific Integrated Circuits (ASICs), have been tailor-made to cater to the unique requirements of machine learning workloads [2, 3].”
MethodA Lightweight Network for Seismic Phase Picking on Embedded Systems
Zhu, Y., Zhao, S., Wei, W., Zhang, F., & Zhao, F. (2024) · IEEE Access · doi:10.1109/access.2024.3416034 · Cited by 6
“Despite limitations in processing speed and parallel computation, embedded CPUs/MPUs have been applied intelligently in diverse domains, such as smart agriculture [9], [10], fisheries [11], and livestock management [12], [13].”
Yamamoto, A., Kusudo, T., Kimura, M., & Matsuno, Y. (2022) · AgriEngineering · doi:10.3390/agriengineering4010002 · Cited by 6
“At present, automation in persimmon harvest is limited, due to difficulties in cutting the fruits from trees with various shapes and the different maturing conditions for each fruit in the field, unlike the developed auto-harvesting machines for apple cultivation [17].”
BackgroundDefauw, N., Malfante, M., Antoni, O., Rakotovao, T., & Lesecq, S. (2023) · Sensors · doi:10.3390/s23031613 · Cited by 7
“These detectors have proven useful for real-time processing as their reduced computational cost allows them to operate in real-time on less computationally performing hardware platforms [26].”
BackgroundPerformance Evaluation of State-of-the-Art Edge Computing Devices for DNN Inference
Rancano, X., Molanes, R. F., González, C. G., Rodriguez-Andina, J. J., & Fariña, J. (2020) · Annual Conference of the IEEE Industrial Electronics Society (IECON) · doi:10.1109/iecon43393.2020.9255055 · Cited by 6
“The use of the GPU is recommended for applications where DNNs are combined with other processing tasks for which GPUs are suitable, such as image and video processing applications, and which do not have power restrictions [20].”
MethodIPT: A Dataset for Identity Preserved Tracking in Closed Domains
Heitzinger, T., & Kampel, M. (2021) · Proceedings - International Conference on Pattern Recognition/Proceedings/International Conference on Pattern Recognition · doi:10.1109/icpr48806.2021.9412979 · Cited by 8
“This variant is of interest for embedded solutions, which are target applications for our dataset, as it has been shown to run at 5 fps on a Raspberry Pi 3 B+ with an Intel NCS2 and at 30 fps on the Nvidia Jetson AGX Xavier [26].”
MethodFruit Identification and Estimation of Calories Using YOLOv3 Algorithm
Migallon, M. F., Santos, T. A. M., & Villaverde, J. F. (2024) · IEEE International Conference on Automatic Control and Intelligent Systems (I2CACIS) · doi:10.1109/i2cacis61270.2024.10649857 · Cited by 5
Pontes, F. A., Schukat, M., & Curry, E. (2024) · IEEE Access · doi:10.1109/access.2024.3520807 · Cited by 7
“Reference [3] studies the viability and cost of running a state-of-the-art DNN CV model in non-simulated GPU-enabled devices for real-time smart-farming applications.”
InfluentialDo, T. (2022) · International Journal of Web Information Systems · doi:10.1108/ijwis-03-2022-0055 · Cited by 8
Recent Challenges on Edge AI with Its Application: A Brief Introduction
Joshi, K. K., Anandaram, H., Khanduja, M., Kumar, R., Saini, V., & Mohialden, Y. M. (2022) · Studies in computational intelligence · doi:10.1007/978-3-031-18292-1_5 · Cited by 8
Underwater Object Detection of AUV based on Sonar Simulator utilizing Noise Addition
Sung, M., Song, Y.-W., & Yu, S. (2022) · IEEE/OES Autonomous Underwater Vehicles · doi:10.1109/auv53081.2022.9965853 · Cited by 5
“Therefore, it has shown fast processing speed even on single board computers [15], [16], and it is suitable for use on the AUV, which has also limited computing power.”
BackgroundAdaptive Information Fusion Using Evidence Theory and Uncertainty Quantification
Arévalo, F., Alison, M. P. C., Ibrahim, M. T., & Schwung, A. (2023) · IEEE Access · doi:10.1109/access.2023.3348270 · Cited by 5
“Fault detection systems are a key component of decision assistance systems because they provide insights into the machine’s condition [1], [2], [3].”
BackgroundAgro-vision IoT-enabled Crop Pest Recognition System based on VGG-16
Udeogu, C. U., Nwakanma, C. I., Ayoade, I. A., Amadi, C. S., & Eze, U. F. (2023) · International Conference on Multidisciplinary Engineering and Applied Science (ICMEAS) · doi:10.1109/icmeas58693.2023.10429830 · Cited by 5
“Furthermore, these methods employed machine learning algorithms that are computationally high with intensive hardware setup, which is ill-suited for infield real-time pest recognition due to their weight and power constraints [8].”
MethodDetection and counting of overlapped apples based on convolutional neural networks
Gao, M., Ma, S., Zhang, Y., & Xue, Y. (2022) · Journal of Intelligent & Fuzzy Systems · doi:10.3233/jifs-213072 · Cited by 5
Emergence of Capsule Network for Automatic Medical Disease Classification
Yadav, S., & Dhage, S. (2022) · Sardar Patel International Conference on Industry 4.0 - Nascent Technologies and Sustainability for 'Make in India' Initiative · doi:10.1109/spicon56577.2022.10180503 · Cited by 6
“Picture categorization [3-5], object recognition [6-8], and instance segmentation [9] are just a few examples of how convolutional neural networks (CNN) have revolutionised artificial visual perception over the last decade.”
BackgroundShahin, A. I., & Almotairi, S. (2021) · PeerJ Computer Science · doi:10.7717/peerj-cs.772 · Cited by 6
“Generally, it was proven that this stage makes the detection task faster (Mazzia et al., 2020; Redmon & Farhadi, 2017).”
BackgroundOliveira, F., Costa, D. G., & Assis, F. (2026) · Internet of Things · doi:10.1016/j.iot.2026.101926 · Cited by 4
Efficient Edge AI Based Annotation and Detection Framework for Logistics Warehouses
Mori, Y., Asai, Y., Higashiura, K., Katayama, S., Urano, K., Yonezawa, T., & Kawaguchi, N. (2025) · IEEE Consumer Communications and Networking Conference · doi:10.1109/ccnc54725.2025.10976142 · Cited by 5
“Object detection and tracking algorithms using edge AI cameras have been studied [6]–[8], primarily targeting objects in everyday environments where open-source datasets are sufficient.”
AI Management Platform with Embedded Edge Cluster
Kum, S., Yu, M., Kim, Y., Moon, J., & Cretti, S. (2021) · International Conference on Consumer Electronics · doi:10.1109/icce50685.2021.9427731 · Cited by 4
“TensorFlow or PyTorch is compatible with ARM architecture, and it can be installed to the embedded devices just the same as we do that on a PC. Mazzia et al [1] has implemented a rea-time detection system based on YOLOv3-tiny on Nvidia Xavier.”
MethodA real-time detection model for smoke in grain bins with edge devices
Yin, H., Chen, M., Lin, Y., Luo, S., Chen, Y., Yang, S., & Gao, L. (2023) · Heliyon · doi:10.1016/j.heliyon.2023.e18606 · Cited by 4
Rincon, J. A., Julián, V., & Carrascosa, C. (2020) · Communications in computer and information science · doi:10.1007/978-3-030-51999-5_3 · Cited by 4
Query-Based Video Summarization System Based on Light Weight Deep Learning Model
(2022) · International journal of intelligent engineering and systems · doi:10.22266/ijies2022.1231.24 · Cited by 8
Chip Contour Detection Based on Real-time Image Sensing and Recognition
Chang, B. R., Tsai, H. F., Hsieh, C.-W., & Chen, M.-L. (2022) · Sensors and Materials · doi:10.18494/sam3378 · Cited by 4
Fast Person Detector with Efficient Multi-level Contextual Block for Supporting Assistive Robot
Putro, M. D., Nguyen, D.-L., Priadana, A., & Jo, K.-H. (2022) · IEEE International Conference on Industrial Cyber-Physical Systems (ICPS) · doi:10.1109/icps51978.2022.9816965 · Cited by 4
Automated neural network generation for industrial datasets, application to laser powder bed fusion
Avval, S. S. P., Abadie, T., & Yaghoubi, V. (2025) · Journal of Intelligent Manufacturing · doi:10.1007/s10845-025-02707-0 · Cited by 3
“However, some edge devices that contain graphics processing units (GPUs) or tensor processing units (TPUs) are quicker with models that have a larger number of parameters but less number of layers (Mazzia et al., 2020; Smolyanskiy et al., 2017).”
Segment and pick any fruit: Text-prompted robotic harvesting
Fan, Z., Su, X., Lu, Y., Fang, Y., Gao, Z., Feng, J., & Hu, Y. (2026) · Pattern Recognition · doi:10.1016/j.patcog.2026.113836 · Cited by 4
Enhancing navigation benchmarking and perception data generation for row-based crops in simulation
Martini, M., Eirale, A., Tuberga, B., Ambrósio, M. C., Ostuni, A., Messina, F., Mazzara, L., & Chiaberge, M. (2023) · Wageningen Academic Publishers eBooks · doi:10.3920/978-90-8686-947-3_56 · Cited by 3
Self-citationToward Real-Time Posture Classification: Reality Check
Zhang, H., Gračanin, D., Zhou, W., Dudash, D., & Rushton, G. T. (2025) · Electronics · doi:10.3390/electronics14091876 · Cited by 3
Plant Leaf Disease Classification using Convolutional Neural Network on FPGA
Shah, P., Rathod, G., Gajjar, R., Gajjar, N., & Patel, M. I. (2023) · International Conference on Device Intelligence, Computing and Communication Technologies (DICCT) · doi:10.1109/dicct56244.2023.10110124 · Cited by 3
“In [14], YOLOv3-Tiny has been deployed on NVIDIA Jetson AGX Xavier.”
MethodMorita, A., & Okuno, H. (2024) · IEICE Transactions on Electronics · doi:10.1587/transele.2024lhp0003 · Cited by 3
“There have also been comparisons of power consumption and frame rate when building embedded image recognition systems for a specific purpose using embedded GPUs and small hardware accelerators [18].”
Real-time Face Mask Detection System on Edge using Deep Learning and Hardware Accelerators
Ruparelia, S., Jethva, M., & Gajjar, R. (2021) · International Conference on Communication, Computing and Industry 4.0 (C2I4) · doi:10.1109/c2i454156.2021.9689421 · Cited by 3
Automotive Embedded Image Classification Systems
Medina, A., Ponce, P., & Ramírez-Mendoza, R. A. (2022) · International Symposium on Electromobility (ISEM) · doi:10.1109/isem55847.2022.9976661 · Cited by 3
“…object classifiers have helped with the speed of object recognition and have proved to have the accuracy needed for these implementations [2], [3], [4], and [5]; how-2022 ever, the processing power required for this solutions is high enough to require a computer aboard the vehicle, and even…”
BackgroundA Real-time Low-cost Artificial Intelligence System for Autonomous Spraying in Palm PlantationsPreprint
Qin, Z., Wang, W., Dammer, K., Guo, L., & Cao, Z. Z. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2103.04132 · Cited by 3
“in [25]) if the backbone network has well enough representational power since more predictors bring more computation complexity.”
BackgroundMethodDesigning a Multiple-User Wearable Edge AI system towards Human Activity Recognition
Silva, M. C., Bianchi, A. G. C., Oliveira, R., & Ribeiro, S. P. (2022) · Brazilian Symposium on Computing System Engineering (SBESC) · doi:10.1109/sbesc56799.2022.9964915 · Cited by 3
“As facing an Edge AI concept is a new task, some authors evaluate aspects of the development of such solutions [9], [10] while others propose novel technologies [11] and solutions [12], [13] based on this convergence.”
BackgroundYOLO on GPU Implementation for a Vision-Based Fruit Select and Place System with a 5-DOF Robotic Arm
Panigrahi, S., Biyani, A., Mithari, A., Sawant, S., Tade, S. L., & Kolte, M. T. (2023) · International Conference on Computing Communication Control and automation · doi:10.1109/iccubea58933.2023.10391950 · Cited by 3
“[ 10]The real-time, on-field application of Apple Detection further concretized the notion that a low-cost, accurate vision system for object detection can be developed.”
BackgroundClassification of soybean pods using the deep learning techniques
Bandeira, P. M. D. C., Villar, F. M. D. M., Bandeira, P. P. D. C., & Dias, I. A. (2023) · Revista Engenharia na Agricultura - REVENG · doi:10.13083/reveng.v30i1.15436 · Cited by 2
Enhancing Traffic Management with Embedded Machine Learning for Vehicle Detection
Talip, M. S. A., Razak, M. Z. A., Mohamad, M., Khairuddin, A. S. M., Izam, T. F. T. M. N., & Azizan, A. (2023) · doi:10.1109/icm60448.2023.10378908 · Cited by 2
Yan, Z., Zhang, H., & Li, L. (2024) · IEEE International Conference on Power, Electronics and Computer Applications (ICPECA) · doi:10.1109/icpeca60615.2024.10471151 · Cited by 2
“MAZZIA V et al [15] utilized the enhanced YOLOv3-Tiny network to attain a high detection speed.”
MethodReviewing Approaches and Techniques for Detecting Suspicious Human Behavior: A Comprehensive Survey
Wong, K. C., & Ahmad, M. S. (2024) · ELEKTRIKA- Journal of Electrical Engineering · doi:10.11113/elektrika.v23n2.538 · Cited by 2
“One of the main challenges in embedded Machin learning is achieving high output accuracy while minimizing power consumption [34].”
BackgroundMethodDeployment of Object Detection Enhanced with Multi-label Multi-classification on Edge Device
Oderhohwo, O., Odetola, T. A., Mohammed, H., & Hasan, S. R. (2020) · IEEE International Midwest Symposium on Circuits and Systems (MWSCAS) · doi:10.1109/mwscas48704.2020.9184601 · Cited by 2
“Apple detection in real time on a RP13 is presented in [9].”
A Fast Real-time Facial Expression Classifier Deep Learning-based for Human-robot Interaction
Putro, M. D., Nguyen, D.-L., & Jo, K.-H. (2021) · International Conference on Control, Automation and Systems (ICCAS) · doi:10.23919/iccas52745.2021.9650034 · Cited by 2
Deep-Learning-Based Visual Odometry Models for Mobile Robotics
de Sousa, F. L. M., Meira, N. F. D. C., Oliveira, R., & Silva, M. C. (2021) · doi:10.5753/sbesc_estendido.2021.18504 · Cited by 2
“(2020) [15] implemented an embedded edge AI solution for real-time apple detection in orchards, with the YOLOv3-tiny algorithm on three embedded platforms.”
MethodMeira, N. F. D. C., Silva, M. C., Vieira, C. B., Souza, A. C. C., & Oliveira, R. (2022) · Lecture notes in business information processing · doi:10.1007/978-3-031-08965-7_8 · Cited by 2
Recent Advances and Challenges of Edge AI and IoT Assisted Covid-19 Alike Detection Systems
Kamruzzaman, M. M., Hossin, M. A., & Alrashdi, I. (2022) · IEEE Globecom Workshops (GC Wkshps) · doi:10.1109/gcwkshps56602.2022.10008591 · Cited by 2
“[5] shows that apple detection which is before the Edge AI system requires more complex hardware and system implementation.”
BackgroundRSSI Fingerprinting Using Machine Learning for Position Estimation
Sagala, A., Lumbantoruan, D. P., & Lubis, R. (2023) · International Conference of Computer Science and Information Technology (ICoSNIKOM) · doi:10.1109/icosnikom60230.2023.10364450 · Cited by 2
“The adaptability and dynamic nature of AI models allow for better interpretation and modeling of the complex patterns within the RSSI datasets [14], [15].”
MethodTowards a Dynamic Edge AI Framework Applied to Autonomous Driving Cars
Muratore, G., Rincon, J. A., Julián, V., Carrascosa, C., Greco, G., & Fortino, G. (2020) · Communications in computer and information science · doi:10.1007/978-3-030-51999-5_34 · Cited by 2
Das, A. (2020) · DalSpace (Dalhousie University) · Cited by 2
Quantification and classification of coffee fruits with computer vision
Bazame, H. C. (2021) · Digital Library of Theses and Dissertations (Universidade de São Paulo) · doi:10.11606/t.11.2021.tde-14122021-160103 · Cited by 2
Edge Computing for Wireless Sensor Network
Nguyen, T. A., Huynh, N. H., & Le, T. N. (2022) · Lecture notes on data engineering and communications technologies · doi:10.1007/978-3-031-15063-0_8 · Cited by 1
Film dance creation practice supported by Cyber Physical System
Lin, Z., & Zhang, Y. (2023) · PLoS ONE · doi:10.1371/journal.pone.0284478 · Cited by 1
“Previously, embedded systems were the core of various digital products for filmmaking [3].”
BackgroundFlood Forecasting Using Edge AI and LoRa Mesh Network
Tham, M., Ng, X. H., Leong, R.-C., & Owada, Y. (2023) · Signals and communication technology · doi:10.1007/978-3-031-43781-6_7 · Cited by 1
Implementaciones de selección visual en frutas: revisión sistemática de literatura
Parraga-Badillo, S. R., & Ygnacio, M. A. C. (2024) · Revista Cientifica de Sistemas e Informatica · doi:10.51252/rcsi.v4i1.591 · Cited by 1
Bishnoi, R., Yaldagard, M. A., Hamdioui, S., Vadivel, K., Sifalakis, M., Rodriguez, N. D., Julian, P. M., Ratschbacher, L., Mallah, M., Patil, Y. R., et al. (2025) · Design, Automation, and Test in Europe · doi:10.23919/date64628.2025.10993206 · Cited by 1
“In this context, embedded systems are specialized computers that are integrated into larger systems, offering control and automation capabilities, and when combined with AI, they enable smarter automation, predictive maintenance, and real-time decision-making [1], [2].”
A Comprehensive Study on Artificial Intelligence of Things Optimization Approaches
Ghamri, M. A., Djamaa, B., Benatia, M. A., & Boumediri, T. (2025) · Information systems engineering and management · doi:10.1007/978-3-031-90758-6_2 · Cited by 1
Toward Inference Latency Optimization for Scalable Collaborative Multi-UAV Analytics
Wang, Y., Yuan, J., Wu, W., Yao, Q., Xu, D., & Shen, Z. (2025) · IEEE Transactions on Green Communications and Networking · doi:10.1109/tgcn.2025.3625726 · Cited by 1
“For the inference engine, we use an e ffi cient YOLOv3-tiny model for each classifier [41].”
From Content to Conversation: Explaining Adoption of On-Device Generative AI Tour Guides
Chen, M., Kuo, F.-K., & Hsiao, K. (2025) · International Journal of Human-Computer Interaction · doi:10.1080/10447318.2025.2604631 · Cited by 4
“The system runs on an NVIDIA Jetson Orin NX 16 GB, whose CUDA-enabled GPU provides on-device acceleration in a compact, lightweight, and low-power form factor suited to mobile deployment (Mazzia et al., 2020).”
Multimodal Remote Sensing Sensors for Edge-Aware Intelligent Perception Systems
Sukhroop, P., Shekhar, Bhardwaj, V. K., Essa, I. M., AL-Shaeli, I. K., & Gharban, H. A. J. (2025) · International Conference on Advanced Computing and Emerging Technologies (ACET) · doi:10.1109/acet67282.2025.11430375 · Cited by 1
“We make architectural choices to prioritize parameter efficiency (low FLOPs), and normalize methods so they can be applied in real-time, on-device conditions [9].”
An enhanced you only look once model for multi-class apple detection in natural orchard environments
Liu, X., Zhang, Z., Yu, J., Hua, W., Li, X., Li, H., Zhang, M., & Saha, C. K. (2026) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2026.114733 · Cited by 1
Persimmon maturity index for chronological analysis using AI object detection algorithm
Yamamoto, A., Shinoda, M., Kusudo, T., Kimura, M., & Matsuno, Y. (2022) · Acta Horticulturae · doi:10.17660/actahortic.2022.1338.22 · Cited by 1
Blockchain and digital enterprise management based on embedded processing system
Li, D., & Liang, S. (2023) · International Journal of Systems Assurance Engineering and Management · doi:10.1007/s13198-023-02166-x · Cited by 2
Weed Recognition Model Based on Embedded Deep Neural Network
曾, 翊. (2023) · Modeling and Simulation · doi:10.12677/mos.2023.126449 · Cited by 1
GPU-based Multivariate IGBT Lifetime Prediction
Moniruzzaman, M., Okilly, A. H., Choi, S., & Baek, J. (2023) · IEEE Energy Conversion Congress and Exposition (ECCE) · doi:10.1109/ecce53617.2023.10362123 · Cited by 1
“However, recent advancements in computational resources, particularly the availability of affordable, high-performance Graphics Processing Units (GPUs), have created new opportunities for multivariate data-driven monitoring systems [12-13].”
BackgroundReal-time AI: Using AI on the tactical edge
Fouad, H. Y., Broadrick, O., Harvey, B., Peeke, C., & Narahari, B. (2024) · Elsevier eBooks · doi:10.1016/b978-0-443-15988-6.00004-2 · Cited by 1
Lavender Autonomous Navigation with Semantic Segmentation at the Edge
Navone, A., Romanelli, F., Ambrosio, M., Martini, M., Angarano, S., & Chiaberge, M. (2024) · Communications in computer and information science · doi:10.1007/978-3-031-74633-8_18 · Cited by 1
Self-citationTyacke, E., Reddy, S. P. J., Feng, N., Edlabadkar, R., Zhou, S., Patel, J., Hu, Q., & Atashzar, S. F. (2022) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2022.02.25.482002
“In the literature, Convolutional Neural networks (CNN) have helped achieve remarkable results in problems including image classification [27]–[29] to object detection [30]–[32].”
BackgroundPortable Citrus Detection System Combining UAV and Edge Equipment
Huang, H., & Kadoch, M. (2022) · Lecture notes in electrical engineering · doi:10.1007/978-981-19-4775-9_49
Deep Learning Based 3D Object Detection with Time-of-Flight
Roy, D. M., Shihabudheen, K. V., Singh, S. K., & Navada, A. (2022) · International Conference on Emerging Techniques in Computational Intelligence (ICETCI) · doi:10.1109/icetci55171.2022.9921375
“…real-time apple detection system which can be extended to any application requiring the detection of tiny objects at the edge is reported in [12], including a comparison between the performance of different hardware platforms (NVIDIA Jetson AGX Xavier and Jetson Nano, Raspberry Pi 3B+ with…”
MethodHabibi, M. A., Nidhom, A. M., Alqodri, F., & Firdaus, A. (2023) · International Conference on Industrial Engineering and Operations Management (IEOM) · doi:10.46254/an12.20221056
“is more efficient in their performance, one alternative is the use of artificial intelligence embedded in the application (Aguirre et al. 2020; Mazzia et al. 2020).”
BackgroundRuaz, F. M. (2023) · Journal of Mechanical Design and Testing · doi:10.22146/jmdt.85645
“However, the network structure of YOLOv3 still can be optimized (Mazzia, 2020), but all this training module size is too big.”
BackgroundNovel Capsule Network Algorithm Based Image Classification with Efficiency
Adhitya Kannan, S. G., & Devi, S. U. (2024) · International Conference on Signal Processing, Computation, Electronics, Power and Telecommunication (IConSCEPT) · doi:10.1109/iconscept61884.2024.10627778
“CNNs are among the most popular and efficient architectures in the deep learning community, particularly for from image classification [3-5] to object detection [6,7] and instance segmentation [8] tasks.”
BackgroundEmbedded Artificial Intelligence in Education: Bibliometric Analysis
Lamrini, M., Benabdelouahab, S., Chkouri, M. Y., & Touhafi, A. (2024) · IEEE International Conference on E-Learning in Industrial Electronics (ICELIE) · doi:10.1109/icelie62250.2024.10814829
Real-time detection of road hazards for autonomous vehicle systems
Szabó, M., Keszler, A., & Tizedes, L. (2025) · IEEE International Conference on Image Processing, Applications and Systems (IPAS) · doi:10.1109/ipas63548.2025.10924564
“However, according to the tests discussed in [19], the NVIDIA Jetson AGX Xavier can perform the same operations in less time than the Raspberry Pi.”
Deep Learning Real Time Class Monitoring System using Artificial Intelligence Vision
Raturi, A., Ahmad, A., Patil, J. B., Joshi, P., Thapliyal, A. G., & Pandey, R. (2024) · Global AI Summit - International Conference on Artificial Intelligence and Emerging Technology (AI Summit) · doi:10.1109/globalaisummit62156.2024.10947982
“Advancements expected in the area, together with efforts to tackle issues, make up a vision [18] that facial recognition becomes an integral and further enhanced element of attendance monitoring systems [19].”
Investigating the Transferability of TOG Adversarial Attacks in YOLO Models in the Maritime Domain
Manasut, P., Ibtasham, M. S., Yaradanakul, Z., Azimi, S., Lafond, S., & Iancu, B. (2025) · IEEE Access · doi:10.1109/access.2025.3589512
“The mean average precision (mAP) is a widely used metric in object detection, as it provides a comprehensive evaluation of detector performance, considering both the accuracy and the precision of detection [34].”
A Learning Package of Computer Vision in AI Topic using Edge-AI Device for STEM Education
Lohakan, M., & Upathamkuekool, C. (2025) · International STEM Education Conference (iSTEM-Ed) · doi:10.1109/istem-ed65612.2025.11129439
“The results of the experiments have revealed great potential for leveraging the tested system to provide real-time locations and numbers of detections while consuming the least amount of power possible [1].”
High Clarity, Low Power: Achieving a 43× Speed Boost in Image Defogging Using FPGA Acceleration
Pham, V. K., & Le, T. (2025) · Engineering Technology & Applied Science Research · doi:10.48084/etasr.11931
“Researchers are now turning to purpose ‑ built accelerators: FPGA implementations [8, 9], Application-Specific Integrated Circuit (ASIC) designs for ultra ‑ low ‑ power dehazing [10], and edge ‑ AI chips tuned for enhancement workloads [11].”
Edge Computing for Computer Vision in IoT: Feasibility and Directions
Savvidis, P., & Papakostas, G. A. (2025) · EAI Endorsed Transactions on Internet of Things · doi:10.4108/eetiot.9404 · Cited by 2
“In telemetry terms, IoT is adapted with Nb-IoT, Sigfox and LoRaWAN, [20].”
Real-Time Smart Home Monitoring and Control Using Raspberry Pi, Sensors, and Edge AI
Bairwa, B., Pathan, F., Noorsaba, & Jadoun, V. K. (2025) · International Conference on Electrical Energy Systems (ICEES) · doi:10.1109/icees67011.2025.11212909
“Mazzia et al. developed an embedded system using AI for fruit detection, demonstrating low-latency performance through hardware acceleration, which can be mapped into home scenarios [16].”
Low-Cost Surveillance Systems for Crops Using IoT
Naresh, K., Goel, K., & Sharma, T. (2026) · Lecture notes in networks and systems · doi:10.1007/978-981-96-9191-3_50
Reduced CNN Model for Rotation-Invariant Classification
Salas, R. R. (2021) · theses.fr (ABES) · doi:10.70675/3a3858bdz5dc9z4523z9a89zd9c4d7238260 · Cited by 1
Cutting-edge technological and management innovations in vegetable seedling nurseries
Fernández, J. A., Gallegos‐Cedillo, V. M., Nájera, C., Gallegos, J., Ochoa, J., del Martínez‐Ballesta, M. C., Gallego, B., Cano, M. D., & Egea-Gilabert, C. (2026) · Acta Horticulturae · doi:10.17660/actahortic.2026.1453.6
Santana, L. S., da Silva, J. M., Ferraz, G. A. E. S., Santana, M. S., dos Santos, L. G., Barros, G. O., Rossi, G., & Bambi, G. (2026) · Journal of Food Measurement & Characterization · doi:10.1007/s11694-026-04320-y
“Since fruit images are a crucial source of high-quality information, comprehensive methods are essential for practical applications [ 24 ].”
TinyML Applications: Airwriting System Applied with TinyML for Vietnamese Handwriting Recognition
Vu, D. N. K., Dang, A. K., & Vu, N. K. A. (2026) · Smart innovation, systems and technologies · doi:10.1007/978-3-032-18974-5_2
SWIRIS: a stereo vision dataset for instance segmentation in inland water reservoirs
Mela, J. L., & García, C. (2026) · Journal of Real-Time Image Processing · doi:10.1007/s11554-026-01888-y
“The success of these models has been pinned by general-purpose benchmarks such as MS COCO, ImageNet, and Pascal VOC [2, 3, 5], as well as domain-specific datasets for autonomous driving (e.g., KITTI and Waymo), medical imaging, and precision agriculture [7, 8, 31, 32].”
Metadata
- Authors
- Mazzia, V., Khaliq, A., Salvetti, F., & Chiaberge, M.
- Affiliations
- Politecnico di Torino (IT)
- Venue
- IEEE Access
- Published
- 2020
- Volume
- 8
- Pages
- 9102-9114
- ISSN
- 2169-3536 (ISSN-L)
- OpenAlex
- W3000086329
- Cited by
- 212
- References
- 59
- Citations by year
- 2026: 21 · 2025: 18 · 2024: 36 · 2023: 42 · 2022: 40 · 2021: 42 · 2020: 13
- FWCI
- 22.13
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Locations
- IEEE Access (publishedVersion, cc-by)
- arXiv (Cornell University) (submittedVersion)
- PORTO@Iris (submittedVersion, cc-by)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- arXiv (Cornell University) (cc-by)
- Funders
- Politecnico di Torino
- Topics
- Life Sciences › Agricultural and Biological Sciences › Plant Science › Smart Agriculture and AI
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
- Physical Sciences › Computer Science › Computer Networks and Communications › IoT and Edge/Fog Computing
- Keywords
- apple detection, real-time object detection, edge AI, embedded systems, YOLOv3-tiny, hardware accelerators, Jetson Nano, NVIDIA Jetson AGX Xavier, Raspberry Pi, unmanned ground vehicles, small object detection, orchard monitoring
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.