The development of real-time road condition systems will better monitor network operation status. However, the weak point all these is their need for comprehensive and reliable data. For traffic data acquisition, two sources are currently available: 1) floating vehicles 2) remote microwave sensors (RTMS). former consists use mobile probe as sensors, latter a set fixed detectors installed in roads. First, structure three-layer BP neural designed to achieve fusion car (FCD) detector (FDD)...

Fuel consumption and air pollution caused by transportation is becoming more serious. Developing an effective meso model to analyze dynamically the temporal spatial distribution of fuel in urban road network has become a research focus. However, parameters used most models traffic field are not detailed enough, generalization ability accuracy these still be improved. For discovering typical driving features relationship with consumption, this paper collected 150 million records within two...

Traditional methods for landslide survey, whether field investigation or human remote sensing-based interpretation approaches, all require considerable labour costs and expert knowledge. Deep learning-based detection have significantly improved the speed of recognition, but their accuracy still has much room improvement. In our work, we propose SA-MFNet to achieve pixelwise based on multisource data fusion analysis. On one hand, feature extraction by utilizing an attention mechanism. other...

Semantic segmentation with convolutional neural networks under a complex background using the encoder-decoder network increases overall performance of online machine vision detection and identification. To maximize accuracy semantic background, it is necessary to consider response values objects components their mutually exclusive relationship. In this study, we attempt improve low component segmentation. The basic encoder selected for segmentation, UPerNet modified based on analysis module....

This study explores a CNN based model to identify abnormal behavior in elevator cabs, which I have named S-LRCN ( S-Long-term Recurrent Convolutional Network). It starts with the detection of key points human skeleton by using Openpose method, then further detects and tracks body through CenterNet DeepSort methods, finally integrates Long Short Term Memory Network (LSTM) Neural (CNN) form deep learning model. In this study, large dataset (500 video clips) collected from real cabs different...

Chassis assembly quality is a necessary step to improve product and yield. In recent years, with the continuous expansion of deep learning method, its application in detection increasingly extensive. The current limitations shortcomings existing methods feasibility improving method are presented discussed this paper. According characteristics numerous parts complex types chassis components, for identification based on component segmentation proposed. proposed first performed using Mask...

Fall detection has always been challenging. There are many current works that have achieved good results in fall detections, but most of the existing only consider single feature dimension (i.e., temporal features or spatial features). This study proposes a method combines LSTM and XGBoost model to detect human body falls through fusing both features. It starts with extracting bond points on each frame video by using AlphaPose. In second step, three vertical height, aspect ratio external...

This paper proposed a polishing path planning method of super accuracy mirror mold with free-form surface by curvature analysis. First, IGES files surfaces are read and the geometry is regenerated as B-spline Automatic Mold Polishing System (AMPS). By using derivative properties surface, normal vector principal curvatures at any point calculated. In addition, effective contact width between tool based on grain size radii also determined. The minimum in 3-D mapped onto (u, v) parameters...

The detachment of railway tunnel lining constitutes a grave danger to train operation safety and drastically curtails the tunnel’s service life. This study endeavors efficiently detect void defects in by creating finite element model structures. Utilizing this model, simulates nonlinear acoustic wave propagation cloud maps for three representative structures: void-free, air void, water void. facilitates thorough examination signal characteristics wavefield, time domain, frequency domain. To...

Based on influencing factors of the diagonal crack widths high-strength reinforced concrete beam, paper presents a model for predict width beam by artificial neural network. The is verified using experimental data; it indicates that network has good effect forecast widths. At same time, provides new way to calculate

Based on the heat-resistant mechanism, orthogonal experimental method and requirements for physical-mechanical performances of concrete, one mixing process project was taken as target analysis mixture ratio design concrete. The conventional raw materials such normal Portland cement, common crushed aggregates, fly ash concrete admixture were adopted to manufacture C30 which gets limit temperature 500°C subsequently. And influencing factors strength also studied in this paper. Experimental...

Owing to the recent development in deep learning, machine vision has been widely used intelligent manufacturing equipment multiple fields, including precision-manufacturing production lines and online product-quality inspection. This study aims at Machine Vision Inspection, focusing on method of semantic segmentation under complex backgrounds. First, fewer-parameters optimization atrous convolution architecture is studied. Atrous spatial pyramid pooling (ASPP) residual network (ResNet) are...

Knowledge representation is one of the key technologies in Chinese character (CC) intelligent formation system. The mechanism human cognition CC term perception and memory cognitive psychology was expounded, formal built, knowledge with semantic network proposed. 6763 CCs National standard GB2312-80 set are taken as experimental subject, through analysis hierarchical structures basic elements each to construct network, platform system developed.

Vehicular fuel consumption and the consequent environmental problems in metropolis attract more attention. However, how to monitoring distribution of dynamically urban road network is still a thorny problem be solved. This paper developed model based on combination multiple meso driving characteristics (e.g. average speed, acceleration). The represented form feature vector this paper. Feature vectors with same correspond pattern. Fuel calculated patterns. experimental results show that...

Understanding the real-time traffic information concerning best routes from commuters' location to their destination is a requirement in most citizens' daily lives. While intersection arterial road key part of network and providing time-dependent delay for intersections seems more desirable. Although efficient algorithmic based on Floating Car Data (FCD) exist, an accurate reliable method estimate OD matrix at missing. A novel estimation proposed overcome drawbacks complex connection links...