Papers by Keyword: Image Processing

Abstract: The process of welding is prone to many defects and these defects can cause the formation of many defective regions. It is necessary to identify the regions of defects as these may cause problems and breakages. In this work, we have proposed a method to detect and identify the defects that are commonly seen in seam welds. Manually identifying the detects is not only error prone and time consuming, most of the defects are not visible to the human eyes. In recent days, X-ray images of weld seam are used for this purpose. In this paper we have applied computer vision techniques and proposed an image processing pipeline to generate a binary segmentation of the image to identify the regions of slag and porosity defect seen in weld seams. From the experimental results on the publicly available dataset, GDX-ray images, it could be observed that, there is a significant improvement in detecting various defects with the proposed approach.

Abstract: Aiming at the problems of long detection time and large detection error when agricultural machinery extracts corn row navigation lines, a method of corn row navigation line extraction based on the adaptive edge detection algorithm is proposed. First, the improved super-green feature algorithm and the maximum inter-class variance method are used to automatically obtain the green feature binary image, and the morphological processing is used to improve the image quality, determining dynamic regions of interest by constraining pixel thresholds, extraction of corn edge contour using adaptive edge detection algorithm, finally, the feature points are fitted by the Theil-Sen estimation method. Experimental results show: the super-green feature algorithm reflects the green content in the image more realistically, using the adaptive edge detection algorithm to extract corn row features, the accuracy rate is 94%, and the processing time of a single frame image is 104ms. Compared with the Hough algorithm extraction and the vertical projection algorithm, the navigation line extraction accuracy is increased by 15% and 8% respectively, and the time-consuming is reduced by 258ms and 150ms respectively. In addition, the stability of the algorithm is analyzed in different environments, all with good timeliness.

Abstract: Meeting stringent emission regulations, the demand for environmentally friendly fuels is increasing by the day. Alternative fuel must be burned alongside conventional fuel to increase the availability of such clean energy sources. The current experimental study investigates the characteristics of the premixed LPG flames with CO₂ dilution in tube swirling and non-swirling burners. The study including testing the effects of equivalence ratios, φ, (0.8, 1, 1.2, & 1.4), CO₂ dilution ratios (0%, 5%, 7.5%, & 10%), and aspect ratio of the non-swirling burner (2, 4, 6, 8, & 10). Two swirling burners with swirl number was tested, namely 0.78 & 0.48. The dilution of CO₂ has been observed lengthens the flame, particularly at higher equivalence ratios and/or flow rates since there is more than one influence, they all agree on a similar influence on flame height. The flame shortens clearly when using a swirling burner. Besides, when increasing the swirl number, the flame height increases slightly. Also, the swirling burner divided the flame's inner core into segments equal to the number of swirl vanes, and a flower-shaped flame was generated at low flow rates. The burner’s aspect ratio affects flame height insignificantly. Flame stability limits increase for a higher equivalence ratio and it enhances due to CO₂ addition. The LPG-CO₂/air mixture has an improved reply to beat flame flashback. The addition of CO₂ expands the flow rate of stable flame by about 40% and 25% for φ = 1 and 1.2 respectively. Utilizing a swirling burner improves flame stability greatly. The limit between flashback and blowout increased by about three times as a result of using a swirling burner.

Abstract: Abstract. 3D printing is a rapidly developing technology in the medical world that has been used for pre-operative planning, prosthetic manufacturing, and training for medical education. This 3D printing is needed for medical education to make it easier for students to study anatomical structures. The advantages of 3D printing provide more detail and tactile representation of anatomical aspects of organs to address the problems of online learning and cadaveric limitations. This research aimed to develop the manufacture of 3D printed models of the human heart organ to improve understanding in learning for medical students. Making a 3D printed model of a heart organ is divisible into six parts: the aorta, right ventricle, left atrium, left ventricle, right atrium, and pulmonary artery. The 3D printing model creation procedure consisted of several steps: image acquisition, image post-processing, and 3D printing. This research used Computed Tomography Scanning (CT-Scan) images of the normal heart in Digital Imaging in Medicine (DICOM) format from Saiful Anwar Hospital, Malang. The segmentation uses the grow from seed technique with 3D Slicer software and is saved in STL format. The accuracy of the 3D printing was carried out by measuring dimensions and volume. Measurements are required to ensure the accuracy of 3D printing so that the resulting organs match the initial image data and can be used as learning media in anatomical structures by medical students.

Abstract: Agriculture improvement is a global economic issue and ongoing challenge in this covid-19 pandemic that is highly dependent on effectiveness. The recognition of the diseases in plant leaf performs a major role in the agriculture industry and city-side greenhouse farms approximate analysis of the leaf disease this article intends to integrate image processing techniques with the “convolutional neural network”, which is one of the deep learning approaches, to classify and detect plant leaf disease and publicly available plant the late data to help treat the leaf as early as possible, which controls the economic loss. This paper has a set that was used which consists of 10 classes of disease and three classes of a plant leaf, this research offers an effective method for detecting different diseases in plant leaf variations. The model was created to detect and recognize healthy plant kinds, such as tomato and potato, and pepper these three leaves will perform under the algorithm called a convolutional neural network. By modifying the parameters and changing the pulling combination, models that have been used to train and test these types of leaf sample images can be created. leaf disease recognition was based on these 10 different types of classes in three different species tomato, potato, and pepper the classification of sample images has reached diseases identification accuracy.

Abstract: Gas metal arc welding process has the capability of producing high quality, all position welding, and is easily adaptable for automated welding applications. Repair welding of random cracks on existing assembly/structure through automatic welding would need real time crack/gap identification and weld path generation. In this work, an image processing-based system is presented for identifying the crack geometry. Graphical user interfaces are also developed to take necessary user inputs required at different stages for crack identification, predicting weld bead dimension, and weld path generation. Based on the identified crack geometry and predicted weld bead feature, linear and curved weld path planning methodology is proposed. The proposed modules are validated for a case study by successfully generating the desired weld paths. Different natures of velocity profiles are considered to appraise the role on motion behaviour and a suitable profile is selected for reducing the jerks at sharp corners/via points on the weld path and maintaining uniform bead geometry.

Abstract: Non-destructive evaluation of structures is a key procedure in operation of aircraft structures, necessary to maintain their quality and integrity. Numerous non-destructive testing (NDT) techniques have been adapted to inspect aircraft structures and are currently used according to appropriate protocols. However, many of them provide only qualitative results, such as the D-Sight optical NDT technique used for inspections in aviation. In this study, a concept of improvement of the D-Sight technique is proposed by means of appropriate experimental program and processing procedures applied to the resulting images from inspection. It was demonstrated that appropriately selected processing methods may allow assessing damage quantitatively and improve the overall sensitivity and applicability of a given NDT technique.

Abstract: Two-phase flow has been used in so many industrial processes, such as boilers, reactors, heat exchangers, geothermal and others. Some parameters which need to be studied include flow patterns, void fractions, and pressure changes. Research on void fractions aims to determine the composition of the gas and liquid phases that will affect the nature and value of the flow property. The purpose of this study is to find out the characteristics of the void fraction of various patterns that occurs and to determine the characteristics of the velocity, length, and frequency of bubbly and plug. Data acquisition was used to convert the data from analog to digital so that it can be recorded, stored, processed, and analyzed. High-speed camera Nikon type J4 was used to record the flow. The condition of the study was adiabatic with variation of superficial gas velocity (J_G), superficial fluid velocity (J_L), and also working fluid. To determine the void fraction by using the digital image processing method. The results of the study found that the flow patterns which occurred in this study were bubbly, plug, annular, slug-annular and churn flows. It also showed that the void fraction value is determined by the superficial velocity of the liquid and air. The higher the superficial velocity of the air, the lower the void fraction value.

Abstract: Welded metal porosity significantly influences the mechanical properties of dissimilar metal joints. In this study, the comparison of porosity evaluation methods was held using the sample of welded-brazed zinc coated steel and Al-Mg-alloy plates joint. Relative porosity was measured through cross-sections’ images area analysis, as well as it was evaluated through 3D-fitting of spotted on these images pores. Area and size of pores was measured, volume and distribution were evaluated. It was found that relative porosity values estimated by 2D and 3D methods are equal.

Abstract: Architectured copper clad aluminium composites processed by a restacking drawing method at room temperature are reported in this work. Wires were drawn to severe plastic strain without any intermediate annealing. Three different diameters were studied in order to examine the influence of a different plastic deformation level on the structure of the different wires. Thanks to image processing it has been shown that independently of the plastic deformation, inserted fibers remain continuous and are homogeneous in size and shape. Furthermore, XRD and TEM characterizations confirm that there is no significant intermetallic growth during the deformation. Thus, the improvement and/or degradation of the functional properties of the wires can be well controlled by performing an appropriate post-processing annealing treatment. Keywords: Cu/Al composite, architectured wire, drawing, microscopy, image processing