Tech Science Press
Not a member yet
3972 research outputs found
Sort by
Numerical Simulation of Derusting Treatment of Steel Parts By Shot Blast
Get PDFIn this paper, we investigated the shot blast treatment for derusting application through finite element (FE) simulations with a large number of random shots. The element deletion technique based on dynamic failure criteria was used to model the removal of rust. The cohesive surface model with damage evolution was used to characterize the decohesion of the rust/substrate interface. The effects of various processing and material parameters on the derusting effectiveness were examined. The results show that the rate of derusting mainly depends on the shot size, velocity and impinging angle, with little relevance to the rust thickness. The spalling of the rust fragments resulted from the decohesion of the rust/substrate interface were observed, especially during the later stage of the treatment. Furthermore, the residual stress and the surface roughness was also investigated and the beneficial effects of shot blast treatment in terms of these aspects were highlighted
A Learning Based Brain Tumor Detection System
Get PDFBrain tumor is one of the most dangerous disease that causes due to uncontrollable and abnormal cell partition. In this paper, we have used MRI brain scan in comparison with CT brain scan as it is less harmful to detect brain tumor. We considered watershed segmentation technique for brain tumor detection. The proposed methodology is divided as follows: pre-processing, computing foreground applying watershed, extract and supply features to machine learning algorithms. Consequently, this study is tested on big data set of images and we achieved acceptable accuracy from K-NN classification algorithm in detection of brain tumor
A Frame Breaking Based Hybrid Algorithm for UHF RFID Anti-Collision
Get PDFMulti-tag collision imposes a vital detrimental effect on reading performance of an RFID system. In order to ameliorate such collision problem and to improve the reading performance, this paper proposes an efficient tag identification algorithm termed as the Enhanced Adaptive Tree Slotted Aloha (EATSA). The key novelty of EATSA is to identify the tags using grouping strategy. Specifically, the whole tag set is divided into groups by a frame of size F. In cases multiple tags fall into a group, the tags of the group are recognized by the improved binary splitting (IBS) method whereas the rest tags are waiting in the pipeline. In addition, an early observation mechanism is introduced to update the frame size to an optimum value fitting the number of tags. Theoretical analysis and simulation results show that the system throughput of our proposed algorithm can reach as much as 0.46, outperforming the prior Aloha-based protocols
Traffic Sign Recognition Method Integrating Multi-Layer Features and Kernel Extreme Learning Machine Classifier
Get PDFTraffic sign recognition (TSR), as a critical task to automated driving and driver assistance systems, is challenging due to the color fading, motion blur, and occlusion. Traditional methods based on convolutional neural network (CNN) only use an end-layer feature as the input to TSR that requires massive data for network training. The computation-intensive network training process results in an inaccurate or delayed classification. Thereby, the current state-of-the-art methods have limited applications. This paper proposes a new TSR method integrating multi-layer feature and kernel extreme learning machine (ELM) classifier. The proposed method applies CNN to extract the multi-layer features of traffic signs, which can present sufficient details and semantically abstract information of multi-layer feature maps. The extraction of multi-scale features of traffic signs is effective against object scale variation by applying a new multi-scale pooling operation. Further, the extracted features are combined into a multi-scale multi-attribute vector, which can enhance the feature presentation ability for TSR. To efficiently handle nonlinear sampling problems in TSR, the kernel ELM classifier is adopted for efficient TSR. The kernel ELM has a more powerful function approximation capability, which can achieve an optimal and generalized solution for multiclass TSR. Experimental results demonstrate that the proposed method can improve the recognition accuracy, efficiency, and adaptivity to complex travel environments in TSR
Super-Resolution Reconstruction of Images Based on Microarray Camera
Get PDFIn the field of images and imaging, super-resolution (SR) reconstruction of images is a technique that converts one or more low-resolution (LR) images into a highresolution (HR) image. The classical two types of SR methods are mainly based on applying a single image or multiple images captured by a single camera. Microarray camera has the characteristics of small size, multi views, and the possibility of applying to portable devices. It has become a research hotspot in image processing. In this paper, we propose a SR reconstruction of images based on a microarray camera for sharpening and registration processing of array images. The array images are interpolated to obtain a HR image initially followed by a convolution neural network (CNN) procedure for enhancement. The convolution layers of our convolution neural network are 3×3 or 1×1 layers, of which the 1×1 layers are used to improve the network performance particularly. A bottleneck structure is applied to reduce the parameter numbers of the nonlinear mapping and to improve the nonlinear capability of the whole network. Finally, we use a 3×3 deconvolution layer to significantly reduce the number of parameters compared to the deconvolution layer of FSRCNN-s. The experiments show that the proposed method can not only ameliorate effectively the texture quality of the target image based on the array images information, but also further enhance the quality of the initial high resolution image by the improved CNN
A Scalable Approach for Fraud Detection in Online E-Commerce Transactions with Big Data Analytics
Get PDFWith the rapid development of mobile Internet and finance technology, online e-commerce transactions have been increasing and expanding very fast, which globally brings a lot of convenience and availability to our life, but meanwhile, chances of committing frauds also come in all shapes and sizes. Moreover, fraud detection in online e-commerce transactions is not totally the same to that in the existing areas due to the massive amounts of data generated in e-commerce, which makes the fraudulent transactions more covertly scattered with genuine transactions than before. In this article, a novel scalable and comprehensive approach for fraud detection in online e-commerce transactions is proposed with majorly four logical modules, which uses big data analytics and machine learning algorithms to parallelize the processing of the data from a Chinese e-commerce company. Groups of experimental results show that the approach is more accurate and efficient to detect frauds in online e-commerce transactions and scalable for big data processing to obtain real-time property
Valorization of a Good Bioceramic from Moroccan Waste Fish Bone by a Heat Treatment Method
No full textA tricalcium phosphate (TCP) material was produced from sardine and mackerel waste using a heat treatment method after a Soxhlet extraction to obtain the non-soluble portion of fish waste. The bones were annealed at temperatures between 400°C and 1200°C. The thermal analysis (TG-DTA) was carried out to investigate the thermal stability of TCP and to confirm the removal of organic matter from the raw fish. The calcined bones were characterized by Fourier transform infrared spectroscopy (FT-IR), X-ray diffraction (XRD), and field emission scanning electron microscopy (FE-SEM). The calcium to phosphorous weight ratio was determined by ICP- AES. FT-IR and XRD confirmed the similarities to synthetic β-TCP (JCPDS-09-169), FE-SEM results revealed the formation of nanostructured TCP. These results suggest that varying the isolation temperature between 600-1200°C has tremendous impact on the production of TCP from fish bone waste with the required properties
Toddy Palm (Borassus Flabellifer) Fruit Fibre Bundles as Reinforcement in Polylactide (PLA) Composites: An Overview About Fibre and Composite Characteristics
No full textToddy palm fruit have an apparent density below 0.8 g/cm³ and offer an interesting lightweight construction potential in polylactide (PLA) composites reinforced with 37 mass-% fibres. Single fibre bundles show similar mechanical properties compared with coir: tensile strength of 240 MPa, Young´s modulus of 3.8 GPa and an elongation at break of 31%. However, density and diameter (~ 50 μm) of fruit fibre bundles are significantly lower. The compression moulded composites have a density of 0.9 g/cm³ and achieved an unnotched Charpy impact strength of 12 kJ/m², a tensile strength of 25 MPa, Young’s modulus of 1.9 GPa and an elongation at break of 9%. Due to the high porosity of the composites and the different stress-strain behaviour of fibre and matrix the fibre-reinforcement potential could not be fully used. Maximum stress of the composite was reached at the elongation at break of the PLA-matrix (~2%) while the fibre achieved its maximum stress at an elongation of ~31%. After reaching the maximum stress of the composite, the fibres were pulled out from the matrix with low energy absorption, resulting in a decrease in stress and a limited reinforcement potential. Additionally, the study investigates whether an insect attack by the Asian fruit fly on the mesocarp has a significant influence on the mechanical fibre characteristics. The results have shown that only the rough surface of the fibre bundles is smoothed by insect infestation. The mechanical properties were not significantly affected. For this reason insect-infested fruits of the toddy palm, which are no longer suitable for food production, can be used for the production of sustainable composite materials
ia-PNCC: Noise Processing Method for Underwater Target Recognition Convolutional Neural Network
Get PDFUnderwater target recognition is a key technology for underwater acoustic countermeasure. How to classify and recognize underwater targets according to the noise information of underwater targets has been a hot topic in the field of underwater acoustic signals. In this paper, the deep learning model is applied to underwater target recognition. Improved anti-noise Power-Normalized Cepstral Coefficients (ia-PNCC) is proposed, based on PNCC applied to underwater noises. Multitaper and normalized Gammatone filter banks are applied to improve the anti-noise capacity. The method is combined with a convolutional neural network in order to recognize the underwater target. Experiment results show that the acoustic feature presented by ia-PNCC has lower noise and are well-suited to underwater target recognition using a convolutional neural network. Compared with the combination of convolutional neural network with single acoustic feature, such as MFCC (Mel-scale Frequency Cepstral Coefficients) or LPCC (Linear Prediction Cepstral Coefficients), the combination of the ia-PNCC with a convolutional neural network offers better accuracy for underwater target recognition