International Journal of Engineering and Manufacturing @ijem
Статьи журнала - International Journal of Engineering and Manufacturing
Все статьи: 532
Development of machine vision system for automatic inspection of vehicle identification number
Статья научная
The vision system is developed to reduce the human effort and improve productivity in the Vehicle Quality Assurance (VQA) shop for inspection of a car bearing a Vehicle Identification Number (VIN), assigned to it at the assembly shop. This project work is carried out in association with M/s Renault Nissan Automotive India Pvt Ltd, Chennai. The vision system consists of a camera fixed on a pan-tilt camera frame and an Optical Character Recognition (OCR) software. The camera frame is mounted on a belt conveyor with remote control of forward, backward and tilting motion. The image of VIN present at the car door is captured through a digital camera placed adjacent to the car. The characters in the VIN image thus obtained are extracted using MATLAB, with configurable OCR software. Template matching method is followed in the OCR process. The MATLAB code can overcome trivial issues in VIN image inspection at the quality shop. Development of a graphic user interface to the software is also described.
Бесплатно
Diabetic Retinopathy Severity Grading Using Transfer Learning Techniques
Статья научная
Diabetic Retinopathy is a severe eye condition originating as a result of long term diabetes mellitus. Timely detection is essential to prevent it from progressing to more advanced stages. Manual detection of DR is labor-intensive and time-consuming, requiring expertise and extensive image analysis. Our research aims to develop a robust and automated deep learning model to assist healthcare professionals by streamlining the detection process and improving diagnostic accuracy. This research proposes a multi-classification framework using Transfer Learning for diabetic retinopathy grading among diabetic patients. An image based dataset, APTOS 2019 Blindness Detection, is utilized for our model training and testing. Our methodology involves three key preprocessing steps: 1) Cropping to remove extraneous background regions, 2) Contrast enhancement using CLAHE (Contrast Limited Adaptive Histogram Equalization) and 3) Resizing to a consistent dimension of 224x224x3. To address class imbalance, we applied SMOTE (Synthetic Minority Over-sampling Technique) for balancing the dataset. Data augmentation techniques such as rotation, zooming, shifting, and brightness adjustment are used to further enhance the model's generalization. The dataset is split to a 70:10:20 ratios for training, validation and testing. For classification, EfficientNetB3 and Xception, two transfer learning models, are used after fine-tuning which includes addition of dense, dropout and fully connected layers. Hyper parameters such as batch size, no. of epochs, optimizer etc were adjusted prioir model training. The performance of our model is evaluated using various performance metrics including accuracy, specificity, sensitivity and others. Results reveal the highest test accuracy of 95.16% on the APTOS dataset for grading diabetic retinopathy into five classes using the EfficientNetB3 model followed by a test accuracy of 92.66% using Xception model. Our top-performing model, EfficientNetB3, was compared against various state-of-the-art approaches, including DenseNet-169, hybrid models, and ResNet-50, where our model outperformed all these methodologies.
Бесплатно
Diffraction Tomography: It's Application in Ultrasound
Статья научная
Ultrasound Diffraction Tomography (UDT) is an important alternative to conventional B-mode imaging. Generally, in diffraction tomography, the most universal available computational strategies for reconstructing the object from its projections are interpolation in the frequency domain and interpolation in the space domain. They are analogous to the direct Fourier inversion and backprojection algorithms of straight ray tomography. In this paper two B-spline interpolation functions are introduced. Due to the computational expenses in the space domain interpolation, we apply the interpolation in the frequency domain to implement our new interpolation functions. We also compare our results with filtered backprojection algorithm result. The validity and feasibility of our method was tested using an agar phantom to mimic the human tissue, olive to mimic the cancer, and water to mimic the cyst. The experimental results show that this method has a promising impact in clinical applications.
Бесплатно
Статья научная
Malaria remains a pervasive global health challenge, affecting millions of lives daily. Traditional diagnostic methods, involving manual blood smear examination, are time-consuming and prone to errors, especially in large-scale testing. Although promising, automated detection techniques often fail to capture the intricate spatial features of malaria parasites leading to inconsistent performance. In order to close these gaps, this work suggest an improved technique that combines a Self-Attention Mechanism and a Dilated Convolutional Neural Network (D-CNN) to allow the model to effectively and precisely classify malaria parasites as infected or uninfected. Both local and global spatial information are captured by dilated convolutions, and crucial features are given priority by the attention mechanism for accurate detection in complex images. We also examine batch size variation and find that it plays a crucial role in maximizing generalization, accuracy, and resource efficiency. A batch size of 64 produced superior results after testing six different sizes, yielding an AUC of 99.12%, F1-Score of 96, precision of 97.63%, recall of 93.99%, and accuracy of 96.08%. This batch size balances efficient gradient updates and stabilization, reducing overfitting and improving generalization, especially on complex medical datasets. Our approach was benchmarked against existing competitors using the same publicly available malaria dataset, demonstrating a 2-3% improvement in AUC and precision over state-of-the-art models, such as traditional CNNs and machine learning methods. This highlights its superior ability to minimize false positives and negatives, particularly in complex diagnostic cases. These advancements enhance the reliability of large-scale diagnostic systems, improve clinical decision-making, and address key challenges in automated malaria detection.
Бесплатно
Статья научная
This paper presents a distribution network reconfiguration based on bacterial foraging optimization algorithm (BFOA) along with backward-forward sweep (BFS) load flow method and geographical information system (GIS). Distribution network reconfiguration (DNR) is a complex, non-linear, combinatorial, and non-differentiable constrained optimization process aimed at finding the radial structure that minimized network power loss while satisfying all operating constraints. BFOA is used to obtain the optimal switching configuration which results in a minimum loss, BFS is used to optimize the deviation in node voltages, and GIS is used for planning and easy analysis purposes. Simulation is performed on the 33-bus system and results are compared with the other approaches. The obtained results show that the proposed approach is better in terms of efficiency and having good convergence criteria.
Бесплатно
Driver Drowsiness Detection System
Статья научная
When you are driving a car and you are being responsible for your co-passenger and other innocent being on the road, you should be extra responsible. Many fatal and minor accidents happen on the road due to the drowsiness of drivers only. Hence, there is a need to detect drowsiness while driving a car. It has become an important requirement for everyone’s safety. The main objective of this study is to create a highly accurate drowsiness detection system using methods that are both affordable and easy for any car manufacturer to include in their cars. The ultimate objective is to increase road user’s protection by raising the level of safety for both drivers and their cars. This study's main contribution is the implementation of a bimodule method for drowsiness detection. The first module effectively detects signs of drowsiness by analyzing a constant stream of images of the driver in real time using a reinforcement learning model. Simultaneously, the car's second module, which is built into the steering wheel grip, keeps track of the driver's hand pressure when performing turns and emergency scenarios. The findings of the study highlight how well the proposed system works to reduce the risks associated with drowsy driving. It further highlights the value of cutting-edge technology in protecting other drivers and improving driving safety, which has the potential to save lives and avoid accidents.
Бесплатно
Dynamic Modeling and H∞ Control of Offshore Wind Turbines
Статья научная
In this paper, vibration control problem in tension leg platform offshore wind turbines is investigated. First a non-linear model of the wind turbine is obtained. Six degrees of freedom are considered in the model including surge, heave and pitch of the platform, tower fore-aft vibrations, rotor rotation and drivetrain torsional vibration. Moreover all external loads acting on the offshore wind turbine such as aerodynamic loads, hydrodynamic loads and mooring line forces are taken into account. To achieve an accurate model of the wind turbine, tower and drivetrain are modelled as flexible components. The model output is compared with FAST simulator; a popular open source software for modeling wind turbines. Then, a robust controller is designed to regulate rotor speed and output power, increase wind turbine efficiency and attenuate tower fore-aft vibration. The controller is implemented on the non-linear dynamic model to investigate the closed loop performance.
Бесплатно
Dynamic Modelling and Control of Flapping Wing Micro Air Vehicle for Flap-Glide Flight Mode
Статья научная
Flapping wing micro-air-vehicles (FWMAV) are micro-air-vehicles that use biomimetic actuation (oscillatory flapping wing) for aerodynamic force generation. The realization of such bionic flight, which offers small size, low speed, and flexible maneuverability has significant military and civilian values. Thus the design of FWMAV (ESB-I) will be very important for security related sectors since they have all the right stuffs for surveillance and reconnaissance. Since everything about a bird is made for flight the kinematic and dynamic modeling as well as control algorithm of bird like FWMAVs is more complex than that of serial robots. Thus balancing the main requirements for the design of FWMAVs which includes excellent aerodynamic performance, high efficiency, and satisfactory maneuverability is very important. With the aim to improve the performance of a FWMAV this work incorporates an intermittent flapping and gliding flight mode. Flap-gliding flight mode, which is often used by large bird species, effectively combines the aerodynamic advantages of fixed and flapping wings. Inspired from it, a kind of flexible flap-gliding Micro Air Vehicle, named Ethio-Smart Bird-I (ESB-I), was successfully designed. An expression describing the mechanical energy cost of travelling of this flight mode in terms of work per range for one flap-glide flight cycle was presented. It is shown that there is an energy saving of flap-gliding flight compared to continuous flapping flight. However due to a system dynamic variation in this flight mode, it possesses difficulty in control surface design. To implement this specific flight mode, this thesis proposes a closed-loop active disturbance rejection control, ADRC, strategy to stabilize the attitude during the processes of flapping flight, transition and gliding flight. To verify the control effect, the unsteady aerodynamic estimation method of the flapping wing based on modified strip theory approach and the dynamics of the FWMAV in Lagrangian form were modelled in the MATLAB/SIMULINK platform and applied in the simulation. Using this model longitudinal stability of ESB-I was analyzed. Simulation results show that even if the FWMAV is in different flight modes, ADRC controller can track the target pitch signal effectively with tracking error less than 0.05rad. To further explain the effects of ADRC in this specific flight mode, the control effects of a PID controller is presented. As per the simulation result ESB-I with PID controller has a target pitch angle tracking error greater than 1rad. This shows that, in flap-gliding flight mode ADRC can track the target pitch signal better than PID controller.
Бесплатно
Early Detection of Dementia using Deep Learning and Image Processing
Статья научная
Dementia is the world's most deadly disease. A degenerative disorder that affects the thinking, memory, and communication abilities of the human brain. According to World Health Organization, more than 40 million people worldwide suffer from this illness. One of the most common methods for analyzing the human brain, including detecting dementia, is using MRI (Magnetic resonance imaging) data, which provides insight into the inner working of the human body. Using MRI images a deep Convolution neural network was designed to detect dementia, we are utilizing image processing to help doctors detect diseases and make decisions on observation, in an earlier stage of the disease. In this paper, we are going to get to the bottom of the DenseNet-169 model, to detect Dementia. There are approximately 6000 brain MRI images in the database for which the DenseNet-169 model has been used for classification purposes. It is a Convolution Neural Network (CNN) model that classifies Non-Dementia, Mild Dementia, Severe Dementia, and Moderate Dementia. The denseNet-169 model helps us determine Dementia disease. And also present the 97% accuracy for clarification of disease is present in the patient body. we are conducted this survey for providing effective disease prediction model for physicians to conclude that the disease stage is accurate and provide proper treatment for that.
Бесплатно
Edge Detection of Image Using Image Divergence and Downsampling Method
Статья научная
Classically, the points where digital image brightness transforms rapidly are ordered into a group of curved line segments termed as edges. Edge detection is an important feature and tool in digital image processing to analyze the significant changes in gray level image intensity. In this paper, an edge detection method is proposed. In the proposed method divergent operation is applied to the image to compute the Laplacian of the image. After then the sample rate of Laplacian of image is decreased by downsampling. A threshold value is yielded by computing the mean on the down sample value. Laplacian of image and threshold value is compared and pixel values are set according to the threshold value. Then the morphological operation is performed on the processed image to produce the final edge detection image. The significance and value of this research are reducing image noise by downsampling and searching vital edge information through divergence operation. The present study introduces a new method of edge detection. The finding of this research work is to detect the edges of objects. The proposed method is compared with other existing edge detection methods i.e., Canny, Sobel, Robert, Zero cross, and Frei-Chen. Quantitative evaluation is performed through various metrics i.e., Entropy, Edge-based contrast measure (EBCM), F-Measure, and Performance ratio. Experimental results obtained from MATLAB 2018a show that the proposed method performs better than other well-known edge detection methods.
Бесплатно
EduCloud: A Dynamic Three Stage Authentication Framework to Enhance Security in Public Cloud
Статья научная
Now-a-days, one of the most exciting technology is cloud computing. Accessing dynamically virtualized resources through internet is called as cloud computing. Security and confidentiality are the major concerns in public cloud. Though EduCloud (Educational Cloud) uses public cloud, moving data from one location to another location may lead to risk. Information related to staff, student and management or admin that can be shared in EduCloud, are to be secured in public educational cloud environment. In this scenario, data security is the most critical issue in cloud. But present authentication system available does not provide enough security in public EduCloud. Hence, we propose new authentication framework to enhance security in public educational cloud. The features of various authentication techniques are discussed in this paper and a novel framework is proposed for pubic EduCloud, which provides not only security but also increases the response time. The developed software tool is best suited and provably a secured solution to the public educational cloud environment.
Бесплатно
Effect of Thar Coal Fly Ash on Compressive and Tensile Strength of Concrete
Статья научная
This study's subject is the effectiveness of substituting Thar Coal Fly Ash (TCFA) for ordinary Portland cement, also known as OPC. Tharparkar, Pakistan, possesses the world’s third largest coal reserves, with deposited coal fuel of 175 billion tons and capable of providing energy for over 200 years. Thar Coal is a lignite type that produces 7-10% of by-products in ashes; among them, Fly Ash is a significant waste. Reusing this waste as a partial cement replacement offers an environmentally friendly solution. This study prepared concrete specimens with varying proportions of TCFA (0%, 10%, 20%, and 30% by mass) as cement substitutes. Compressive strength tests were conducted on 36 cubes (100mm x 100mm x 100mm) with different fly ash percentages at a proportion to water to cement of 0.52. Ages 7, 14, and 28 days for curing were considered. The findings demonstrate that a higher TCFA component enhances the workability of the concrete. At all curing ages, the strength in compression at a 20% TCFA replacement level was greater than that of standard concrete. However, as the cement replacement was increased to 30%, there was a slight decrease in the comparative compressive strength compared to regular concrete. The tensile strength of the splitting test, performed after twenty-eight days of curing age, reveals that it surpassed conventional concrete for all replacement levels. Considering the favorable outcomes in workability, constrictive strength, durability strength, and substantial economic and environmental benefits, there is much potential for using TCFA as a cement substitute in the construction sector.
Бесплатно
Статья научная
The main objective of the study was to investigate the effect of various filler materials on inter laminar shear properties of glass/epoxy composite materials. In this study three different filler materials were selected (SiO2, TiO2, Glass powder) for high strength and stiffness applications. The physical properties are also studied with the composites. The fiber volume fraction of material plays vital role in preparation of composite laminates. The fabrication method used was conventional hand lay-up process depending upon the resin system hot temperature cure and room temperature curing processes were selected. The short beam shear test was carried out to test inter laminar shear strength (ILSS). The short beam shear test results inter laminar shear strength and to find the significant influence of filler material on characterization of glass fiber reinforced plastic (GFRP) composite. The test results shown that high shear strength were observed with silicon dioxide(SiO2) filler material and with 10% of silicon dioxide filler material shows maximum ILSS value.
Бесплатно
Статья научная
The mechanical properties of a friction material primarily depend on the interfacial adhesion between all the ingredients of a friction material. In this work, a new friction material is developed by combination carbon fiber (CF), polymer matrix and other ingredients. The surface of CF is chemically inert and hydrophobic in nature and does not possess good bonding property with resin. Therefore, an attempt is made to improve the bonding strength between all the ingredients of a friction material. CF surface is modified by three different surface treatment techniques to increase hydroxyl or carboxyl groups on the surface. First, surface oxidation treatment, Second nitric acid treatment and third grafting multi walled carbon nano tubes functionalized (MWCNT-F) on CF surface. CF content after surface modifications is varied in wt % and mixed with remaining ingredients. Friction composite sheets are fabricated by using hand layup method. The resulting materials are characterized by SEM, TGA and FTIR analysis. MWCNTs-F on CF surface is observed. Twelve composite sheets with varying content of CF and surface treatment method is fabricated. Sample specimens are cut from the friction composite sheets to evaluate tensile and flexural properties of friction material. The best surface treatment method and optimum ingredients are selected for the improvement of tensile and flexural properties of friction material.
Бесплатно
Статья научная
To Monascus spp., light is not the required factor for its growth. But Monascus spp. has the capacity to sense and respond to light. This paper investigated the effects of blue light on growth and the changes of citrinin yield in Monascus 15. Our results demonstrated blue light was a stimulating signal for citrinin formation. Under the blue light illumination, the biomass of Monascus 15 was inhibited, but the citrinin yield increased when comparing with no light culture condition. Spores statistical results revealed that the blue light also influences the development of mycelium and spore formation.
Бесплатно
Efficient Design of Compact 8-bit Wallace Tree Multiplier Using Reversible Logic
Статья научная
Reversible logic is now employed in low-power CMOS circuits, optical data processing, DNA calculations, biological studies, quantum circuits, and nanotechnology. When building quantum computers, for example, the use of reversible logic is unavoidable. The structure of a reversible logic circuit is far more complex than that of an irreversible logic circuit. The multiplication operation is regarded as one of the most crucial in the ALU unit. In this study, the Wallace tree method is utilized to minimize the depth of circuits in 8x 8 reversible unsigned multiplier circuits. The proposed design is an attempt to enhance design factors including the number of gates, garbage outputs, constant inputs, and quantum cost for an 8-bit Wallace Tree multiplier using reversible logic. The Proposed design offers 27% less quantum cost compared to the existing 8-bit Wallace tree multiplier design.
Бесплатно
Efficient Parallel Design for Edit distance algorithm in DNA Sequence Alignment
Статья научная
The focus of Bioinformatics research is usually on two aspects—genomics and proteomics, specifically, it’s starting from nucleic acid and protein sequences, analyzing the structural and functional biological information expressed in the sequences. Biological sequence alignment is one of the common problems, the Needleman-Wunsch algorithm based on dynamic programming is the most basic algorithm, and Edit Distance(Levenshtein Distance) algorithm is also widely used in DNA sequence alignment. Nowadays, there are large amount of improvements on the Needleman-Wunsch algorithm, while few on Edit Distance algorithm, so this paper focuses on revealing the effects of parallel design on optimizing the Edit Distance algorithm, and it also compares the two algorithms’ different significances in DNA sequence alignment objectively.
Бесплатно
Electromagnet separator of different particles
Статья научная
In this paper we simulate and realize an electromagnetic separator excited with an alternating current. And, we will to see the factors that influence the separation. This realization is used for the separation of the ferromagnetic and non ferromagnetic particles. The aim of this application is to recover the particles of non ferrous and ferrous from mixture by attractive magnetic force with big rate of separation. Experiments’ results and simulation of this separator, finally the parameters of separation are optimized with prediction method is presented in this paper.
Бесплатно
Статья научная
In this paper, we have introduced the structure of resonance linear motion actuator which applied to coal mine. And we have built two-dimensional analyzing model by using finite element method. On the basis of the analysis, we have calculated the motor’s air-gap field and magnetism, and we also have studied the affection of motor which caused by permanent magnet’s size. We have studied motor’s winding current when it operates at constant amplitude, and the corresponding experiment results have verified the correctness of finite element model. We have studied the numerical value calculation of resonance linear motion actuator and calculation of its magnetism characteristic. What we have analyzed shows that the air-gap field of motor changes continuously and the magnetism is proportional to winding current. And the results can help technologists to design this kind of motor.
Бесплатно
Статья научная
Cloud computing is a fastest growing technology in the research and industry field.It provides the on demand resources to the customers on the rent basis. These resources are provided through the virtual machines. Resources required by the virtual machines can change dynamically. So load balancing in the cloud is more challenging task as compared to the traditional computing, where the resource requirements are not changed with time. Overall performance of the cloud system can be increased by the efficient load balancing approach. Three steps are involved in the load balancing method i.e., physical machine selection, virtual machine selection and destination physical machine selection. In the past few years a number of load balancing approaches have been proposed to increase the resource utilization and minimize the energy consumption. This paper has proposed a load balancing approach which uses the lower and upper threshold to select the physical machine (PM) for migrating the virtual machine (VM). Then place the selected VM to the PM which consumes minimum power to minimize the energy consumption. To create the cloud environment, CloudSim simulator is used which provides the interface to deal with the physical and virtual machines. To evaluate the performance, the proposed method is compared with already present load balancing approaches. Simulation result shows that proposed approach minimize the energy consumption, migrations and total simulation time.
Бесплатно
