Ship Target Detection Based on Improved YOLO Network

Ship target detection is an important guarantee for the safe passage of ships on the river. However, the ship image in the river is difficult to recognize due to the factors such as clouds, buildings on the bank, and small volume. In order to improve the accuracy of ship target detection and the robustness of the system, we improve YOLOv3 network and present a new method, called Ship-YOLOv3. Firstly, we preprocess the inputting image through guided filtering and gray enhancement. Secondly, we use k-means++ clustering on the dimensions of bounding boxes to get good priors for our model. Then, we change the YOLOv3 network structure by reducing part of convolution operation and adding the jump join mechanism to decrease feature redundancy. Finally, we load the weight of PASCAL VOC dataset into the model and train it on the ship dataset. The experiment shows that the proposed method can accelerate the convergence speed of the network, compared with the existing YOLO algorithm. On the premise of ensuring real-time performance, the precision of ship identification is improved by 12.5%, and the recall rate is increased by 11.5%.

Download Full-text

Linear Static, Real-Time Finite Element Computations Based on Element Masks

ASME 2011 World Conference on Innovative Virtual Reality ◽

10.1115/winvr2011-5501 ◽

2011 ◽

Cited By ~ 2

Author(s):

Holger Graf ◽

Andre´ Stork

Keyword(s):

Finite Element ◽

Stress Field ◽

Boundary Conditions ◽

Real Time ◽

Stiffness Matrix ◽

New Method ◽

Stress Distributions ◽

Post Processing ◽

Time Performance ◽

Computational Overhead

This paper presents a new method for the manipulation of a given CAE domain in view of VR based explorations that enables engineers to interactively inspect and analyze a linear static domain. The interactions can ideally be performed in real-time in order to provide an intuitive impression of the changes to the underlying volumetric domain. We take the approach of element masking, i.e. the blending out of computations resulting from computational overhead for inner nodes, based on the inversion of the stiffness matrix. This allows us to optimize the re-simulation loop and to achieve real-time performance for strain and stress distributions with immediate visualization feedback caused by interactively changing boundary conditions. The novelty of the presented approach is a direct coupling of view dependent simulations and its close linkage to post-processing tasks. This allows engineers to also inspect the changes of the stress field inside of the volume during, e.g. cross sectioning.

Download Full-text

A New Method for High Speed CAN Protocol Conversion Circuit

Applied Mechanics and Materials ◽

10.4028/www.scientific.net/amm.596.873 ◽

2014 ◽

Vol 596 ◽

pp. 873-876

Author(s):

Qi Li ◽

Yan Fei Liu ◽

Da Cheng Luo ◽

Jing Jing Yang

Keyword(s):

Real Time ◽

High Speed ◽

Large Scale ◽

Can Bus ◽

New Method ◽

Time Performance ◽

Test Instrument ◽

Protocol Conversion

Due to high correspondence speed, great real-time performance and good expansibility, CAN bus has been used widely in aerospace, large-scale equipments and other fields these years .This paper introduces a kind of CAN Bus Test Instrument based on PXI bus and FPGA, which is used to test and monitor the CAN bus equipment. The result of test shows that this kind of test instrument has great advantages in reliability, stability and extensibility.

Download Full-text

YOLO-SD: A Real-Time Crew Safety Detection and Early Warning Approach

Journal of Advanced Transportation ◽

10.1155/2021/7534739 ◽

2021 ◽

Vol 2021 ◽

pp. 1-12

Author(s):

Xinwei Lin ◽

Shengzheng Wang ◽

Zhen Sun ◽

Min Zhang

Keyword(s):

Real Time ◽

Data Augmentation ◽

Effective Means ◽

Recall Rate ◽

Training Effect ◽

Site Conditions ◽

Monitoring Methods ◽

Time Performance ◽

Work Area ◽

Stage System

Wearing safety rope while working at the loft and over the side of a ship is an effective means to protect seafarers from accidents. However, there are no active and effective monitoring methods on ships to control this issue. In this article, a one-stage system is proposed to automatically monitor whether the crew is wearing safety ropes. When the system detects that a crew enters the work area without a safety rope, it will warn the supervisor. In this regard, a safety rope wearing detection dataset is established. Then a data augmentation algorithm and a boundary loss function are designed to improve the training effect and the convergence speed. Furthermore, features from different scales are extracted to get the final detection results. The obtained results demonstrate that the proposed approach YOLO-SD is effective at different on-site conditions and can achieve high precision (97.4%), recall rate (91.4%), and mAP (91.5%) while ensuring real-time performance (38.31 FPS on average).

Download Full-text

Cloth Simulation with a Modified Implicit Method Based on a Simplified Mass-Spring Model

Applied Mechanics and Materials ◽

10.4028/www.scientific.net/amm.373-375.1920 ◽

2013 ◽

Vol 373-375 ◽

pp. 1920-1926 ◽

Cited By ~ 4

Author(s):

Jing Hu ◽

Wen Qing Huang ◽

Ke Qiang Yu ◽

Mu Huang ◽

Jun Bai Li

Keyword(s):

Real Time ◽

Time Effect ◽

Implicit Method ◽

Spring Model ◽

Cloth Simulation ◽

Time Performance ◽

Bounding Volume ◽

Mass Spring Model ◽

Bounding Boxes ◽

Mass Spring

Firstly, we introduced the development of cloth simulation in recent years. Based on physical model of cloth simulation, we established the simulation system with a simplified mass-spring model. The computational efficiency is increased with this model. A modified implicit method was proposed in this paper. This method produces plausible animation, and it is easy to be realized with a stable and good real-time performance. The paper adopted AABB (Axis-Aligned Bounding Boxes) bounding volume approach for the detection of cloth collision, it obtains an excellent real-time effect of cloth simulation.

Download Full-text

Research on Function Real-Time of Substation Automation System Based on IEC61850

Advanced Materials Research ◽

10.4028/www.scientific.net/amr.765-767.2021 ◽

2013 ◽

Vol 765-767 ◽

pp. 2021-2025

Author(s):

Xian Pei Wang ◽

Qi Lin Zhang ◽

Yong Biao Zhao

Keyword(s):

Real Time ◽

Network Structure ◽

Processing Speed ◽

Simulation Models ◽

Automation System ◽

Time Performance ◽

Performance Of Systems ◽

Communication Events ◽

Distribution Mode ◽

Substation Automation

The function real-time of substation automation system based on IEC61850 depends on the distribution mode of logical nodes, the processing speed of logical nodes and the communication delay between logical nodes.The realization principles of functions free distribution are introduced based on IEC61850, take bus protection function and distributed interlock function examples, the functions are deposed into the connection diagram of logical nodes and their communication events are introduced in great detail. The simulation models of D2 substation which has unified network or two-tier network are established with OPNET. The communication events of bus bar protection function and distributed interlock function are simulated and analyzed based on this mode. The rationality of distribution mode of systems functions is quantitative analyzed which provide guidance to improve the real-time performance of systems functions and network structure.

Download Full-text

Pose recognition in sports scenes based on deep learning skeleton sequence model

Journal of Intelligent & Fuzzy Systems ◽

10.3233/jifs-189834 ◽

2021 ◽

pp. 1-10

Author(s):

Chen Li-quan ◽

Li You ◽

Fengjun Shen ◽

Zhaoqimeng Shan ◽

Jiaxuan Chen

Keyword(s):

Deep Learning ◽

Real Time ◽

Target Detection ◽

Gesture Recognition ◽

Video Frame ◽

Skeleton Extraction ◽

Recognition Method ◽

Time Performance ◽

Basic Network ◽

Human Joints

Human skeleton extraction is a basic problem in the field of computer vision. With the rapid progress of science and technology, it has become a hot issue in the field of target detection such as pedestrian recognition, behavior monitoring, and pedestrian gesture recognition. In recent years, due to the development of deep neural networks, modeling of human joints in acquired images has made progress in skeleton extraction. However, most models have low modeling accuracy, poor real-time performance, and poor model availability. problem. Aiming at the above-mentioned human target detection problem, this paper uses the deep learning skeleton sequence model gesture recognition method in sports scenes to study, aiming to provide a gesture recognition method with strong noise resistance, good real-time performance and accurate model. This article uses motion video frame images to train the VGG16 network. Using the network to extract skeleton information can strengthen the posture feature expression, and use HOG for feature extraction, and use the Adam algorithm to optimize the network to extract more posture features, thereby improving the posture of the network Recognition accuracy. Then adjust the hyperparameters and network structure of the basic network according to the training results, and obtain the key poses in the sports scene through the final classifier.

Download Full-text