scholarly journals An Edge Computing-Enabled Train Obstacle Detection Method Based on YOLOv3

2021 ◽  
Vol 2021 ◽  
pp. 1-9
Author(s):  
Song Li ◽  
Hongli Zhao ◽  
Jinmin Ma
Keyword(s):  
Cloud Computing ◽  
Deep Learning ◽  
Object Detection ◽  
Real Time ◽  
Detection Method ◽  
Obstacle Detection ◽  
Edge Computing ◽  
Traffic Load ◽  
Learning Tasks ◽  
Onboard Equipment

Rail transit is developing towards intelligence which takes lots of computation resource to perform deep learning tasks. Among these tasks, object detection is the most widely used, like track obstacle detection, catenary wear, and defect detection and looseness detection of train wheel bolts. But the limited computation capability of the train onboard equipment prevents running deep and complex detection networks. The limited computation capability of the train onboard equipment prevents conducting complex deep learning tasks. Cloud computing is widely utilized to make up for the insufficient onboard computation capability. However, the traditional cloud computing architecture will bring in uncertain heavy traffic load and cause high transmission delay, which makes it fail to complete real-time computing intensive tasks. As an extension of cloud computing, edge computing (EC) can reduce the pressure of cloud nodes by offloading workloads to edge nodes. In this paper, we propose an edge computing-based method. The onboard equipment on a fast-moving train is responsible for acquiring real-time images and completing a small part of the inference task. Edge computing is used to help execute the object detection algorithm on the trackside and carry most of the computing power. YOLOv3 is selected as the object detection model, since it can balance between the real-time and accurate performance on object detection compared with two-stage models. To save onboard equipment computation resources and realize the edge-train cooperative interface, we propose a model segmentation method based on the existing YOLOv3 model. We implement the cooperative inference scheme in real experiments and find that the proposed EC-based object detection method can accomplish real-time object detection tasks with little onboard computation resources.

scholarly journals Combining Cloud Computing and Artificial Intelligence Scene Recognition in Real-time Environment Image Planning Walkable Area

2020 ◽  
Vol 5 (1) ◽  
pp. 10-17
Author(s):  
Jia-Shing Sheu ◽  
Chen-Yin Han
Keyword(s):  
Artificial Intelligence ◽  
Cloud Computing ◽  
Object Detection ◽  
Real Time ◽  
Scene Recognition ◽  
Obstacle Detection ◽  
Cleaning Robot ◽  
Artificial Intelligence Technology ◽  
Time Utilization ◽  
Environmental Image

This study developed scene recognition and cloud computing technology for real-time environmental image-based regional planning using artificial intelligence. TensorFlow object detection functions were used for artificial intelligence technology. First, an image from the environment is transmitted to a cloud server for cloud computing, and all objects in the image are marked using a bounding box method. Obstacle detection is performed using object detection, and the associated technique algorithm is used to mark walkable areas and relative coordinates. The results of this study provide a machine vision application combined with cloud computing and artificial intelligence scene recognition that can be used to complete walking space activities planned by a cleaning robot or unmanned vehicle through real-time utilization of images from the environment.

scholarly journals Object Detection Method Based on YOLOv3 using Deep Learning Networks

2019 ◽  
Vol 9 (1) ◽  
pp. 1414-1417
Keyword(s):  
Deep Learning ◽  
Object Detection ◽  
Real Time ◽  
Real World ◽  
Detection Method ◽  
Cross Entropy ◽  
Detection Methods ◽  
Learning Networks ◽  
Learning Techniques ◽  
Localization Errors

—Object Detection is being widely used in the industry right now. It is the method of detection and shaping real-world objects. Even though there exist many detection methods, the accuracy, rapidity, and efficiency of detection are not good enough. So, this paper demonstrates real-time detection using the YOLOv3 algorithm by deep learning techniques. It first makes expectations crosswise over 3 unique scales. The identification layer is utilized to make recognition at highlight maps of three distinct sizes, having strides 32, 16, 8 individually. This implies, with partner contribution of 416 x 416, we will in general form location on scales 13 x 13, 26 x 26 and 52x 52. Meanwhile, it also makes use of strategic relapse to anticipate the jumping box article score, the paired cross-entropy misfortune is utilized to foresee the classes that the bounding box may contain, the certainty is determined and afterward the forecast. It results in perform multi-label classification for objects detected in images, the average preciseness for tiny objects improved, it's higher than quicker RCNN. MAP increased significantly. As MAP increased localization errors decreased.

A Ground-penetrating Radar Object Detection Method Based on Deep Learning

2021 ◽  
Author(s):  
Siyu Chen ◽  
Li Wang ◽  
Zheng Fang ◽  
Zhensheng Shi ◽  
Anxue Zhang
Keyword(s):  
Deep Learning ◽  
Object Detection ◽  
Ground Penetrating Radar ◽  
Detection Method ◽  
Ground Penetrating

scholarly journals Computer vision based obstacle detection and target tracking for autonomous vehicles

2021 ◽  
Vol 336 ◽  
pp. 07004
Author(s):  
Ruoyu Fang ◽  
Cheng Cai
Keyword(s):  
Neural Network ◽  
Computer Vision ◽  
Deep Learning ◽  
Target Tracking ◽  
Real Time ◽  
Autonomous Vehicles ◽  
Autonomous Vehicle ◽  
Obstacle Detection ◽  
Pid Algorithm ◽  
Deep Learning Neural Network

Obstacle detection and target tracking are two major issues for intelligent autonomous vehicles. This paper proposes a new scheme to achieve target tracking and real-time obstacle detection of obstacles based on computer vision. ResNet-18 deep learning neural network is utilized for obstacle detection and Yolo-v3 deep learning neural network is employed for real-time target tracking. These two trained models can be deployed on an autonomous vehicle equipped with an NVIDIA Jetson Nano motherboard. The autonomous vehicle moves to avoid obstacles and follow tracked targets by camera. Adjusting the steering and movement of the autonomous vehicle according to the PID algorithm during the movement, therefore, will help the proposed vehicle achieve stable and precise tracking.

scholarly journals AdaGUM: An Adaptive Graph Updating Model-Based Anomaly Detection Method for Edge Computing Environment

2021 ◽  
Vol 2021 ◽  
pp. 1-12
Author(s):  
Xiang Yu ◽  
Chun Shan ◽  
Jilong Bian ◽  
Xianfei Yang ◽  
Ying Chen ◽  
...  
Keyword(s):  
Internet Of Things ◽  
Anomaly Detection ◽  
Real Time ◽  
Detection Method ◽  
Edge Computing ◽  
Sensor Data ◽  
Computing Environment ◽  
The Real ◽  
Edge Node ◽  
Model Based

With the rapid development of Internet of Things (IoT), massive sensor data are being generated by the sensors deployed everywhere at an unprecedented rate. As the number of Internet of Things devices is estimated to grow to 25 billion by 2021, when facing the explicit or implicit anomalies in the real-time sensor data collected from Internet of Things devices, it is necessary to develop an effective and efficient anomaly detection method for IoT devices. Recent advances in the edge computing have significant impacts on the solution of anomaly detection in IoT. In this study, an adaptive graph updating model is first presented, based on which a novel anomaly detection method for edge computing environment is then proposed. At the cloud center, the unknown patterns are classified by a deep leaning model, based on the classification results, the feature graphs are updated periodically, and the classification results are constantly transmitted to each edge node where a cache is employed to keep the newly emerging anomalies or normal patterns temporarily until the edge node receives a newly updated feature graph. Finally, a series of comparison experiments are conducted to demonstrate the effectiveness of the proposed anomaly detection method for edge computing. And the results show that the proposed method can detect the anomalies in the real-time sensor data efficiently and accurately. More than that, the proposed method performs well when there exist newly emerging patterns, no matter they are anomalous or normal.

scholarly journals Achieving 90% In Data-Centric Industry Deep Learning Task

2021 ◽  
Author(s):  
Tong Guo
Keyword(s):  
Deep Learning ◽  
Object Detection ◽  
Noisy Data ◽  
Learning Task ◽  
Simple Method ◽  
Sequence Generation ◽  
Learning Tasks ◽  
Human Evaluation ◽  
Model Predictions ◽  
Click Through Rate

In industry deep learning application, our manually labeled data has a certain number of noisy data. To solve this problem and achieve more than 90 score in dev dataset, we present a simple method to find the noisy data and re-label the noisy data by human, given the model predictions as references in human labeling. In this paper, we illustrate our idea for a broad set of deep learning tasks, includes classification, sequence tagging, object detection, sequence generation, click-through rate prediction. The experimental results and human evaluation results verify our idea.

Experimental Deep Learning Object Detection in Real-time Colonoscopies

2021 ◽  
Author(s):  
Adrian Ciobanu ◽  
Mihaela Luca ◽  
Tudor Barbu ◽  
Vasile Drug ◽  
Andrei Olteanu ◽  
...  
Keyword(s):  
Deep Learning ◽  
Object Detection ◽  
Real Time ◽  
Learning Object

Prediction of Crime Scene Objects Using Deep Learning

2021 ◽  
Vol 9 (VIII) ◽  
pp. 857-864
Author(s):  
Vibhavari B Rao
Keyword(s):  
Deep Learning ◽  
Object Detection ◽  
Computer Architecture ◽  
Real Time ◽  
Surveillance System ◽  
Crime Scene ◽  
Detection Model ◽  
Detection Algorithms ◽  
Performance Metric ◽  
Smart Surveillance

The crime rates today can inevitably put a civilian's life in danger. While consistent efforts are being made to alleviate crime, there is also a dire need to create a smart and proactive surveillance system. Our project implements a smart surveillance system that would alert the authorities in real-time when a crime is being committed. During armed robberies and hostage situations, most often, the police cannot reach the place on time to prevent it from happening, owing to the lag in communication between the informants of the crime scene and the police. We propose an object detection model that implements deep learning algorithms to detect objects of violence such as pistols, knives, rifles from video surveillance footage, and in turn send real-time alerts to the authorities. There are a number of object detection algorithms being developed, each being evaluated under the performance metric mAP. On implementing Faster R-CNN with ResNet 101 architecture we found the mAP score to be about 91%. However, the downside to this is the excessive training and inferencing time it incurs. On the other hand, YOLOv5 architecture resulted in a model that performed very well in terms of speed. Its training speed was found to be 0.012 s / image during training but naturally, the accuracy was not as high as Faster R-CNN. With good computer architecture, it can run at about 40 fps. Thus, there is a tradeoff between speed and accuracy and it's important to strike a balance. We use transfer learning to improve accuracy by training the model on our custom dataset. This project can be deployed on any generic CCTV camera by setting up a live RTSP (real-time streaming protocol) and streaming the footage on a laptop or desktop where the deep learning model is being run.

Sign in / Sign up

Export Citation Format

Share Document