• Journal of Advanced Navigation Technology
• /
• v.26 no.2
• /
• pp.113-118
• /
• 2022

Motivated by people highlighting important phrases while reading or taking notes we propose a neural network training method by highlighting defective pixel areas to classify effectively defect types of images with complex background textures. To verify our proposed method we apply it to the problem of classifying the defect types of tire band fabric images that are too difficult to classify. In addition we propose a backlight highlighting technique which is tailored to the tire band fabric images. Backlight highlighting images can be generated by using both the GradCAM and simple image processing. In our experiment we demonstrated that the proposed highlighting method outperforms the traditional method in the view points of both classification accuracy and training speed. It achieved up to 13.4% accuracy improvement compared to the conventional method. We also showed that the backlight highlighting technique tailored for highlighting tire band fabric images is superior to a contour highlighting technique in terms of accuracy.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.17 no.4
• /
• pp.577-586
• /
• 2022

Detection and classification of steel surface defects are critical for product quality control in the steel industry. However, due to its low accuracy and slow speed, the traditional approach cannot be effectively used in a production line. The current, widely used algorithm (based on deep learning) has an accuracy problem, and there are still rooms for development. This paper proposes a method of steel surface defect detection combining EfficientNetV2 for image classification and YOLOv5 as an object detector. Shorter training time and high accuracy are advantages of this model. Firstly, the image input into EfficientNetV2 model classifies defect classes and predicts probability of having defects. If the probability of having a defect is less than 0.25, the algorithm directly recognizes that the sample has no defects. Otherwise, the samples are further input into YOLOv5 to accomplish the defect detection process on the metal surface. Experiments show that proposed model has good performance on the NEU dataset with an accuracy of 98.3%. Simultaneously, the average training speed is shorter than other models.

• Smart Structures and Systems
• /
• v.21 no.5
• /
• pp.687-694
• /
• 2018

The problem of wheel tread defects has become a major challenge for the health management of high-speed rail as a wheel defect with small radius deviation may suffice to give rise to severe damage on both the train bogie components and the track structure when a train runs at high speeds. It is thus highly desirable to detect the defects soon after their occurrences and then conduct wheel turning for the defective wheelsets. Online wheel condition monitoring using wheel impact load detector (WILD) can be an effective solution, since it can assess the wheel condition and detect potential defects during train passage. This study aims to develop an FBG-based track-side wheel condition monitoring method for the detection of wheel tread defects. The track-side sensing system uses two FBG strain gauge arrays mounted on the rail foot, measuring the dynamic strains of the paired rails excited by passing wheelsets. Each FBG array has a length of about 3 m, slightly longer than the wheel circumference to ensure a full coverage for the detection of any potential defect on the tread. A defect detection algorithm is developed for using the online-monitored rail responses to identify the potential wheel tread defects. This algorithm consists of three steps: 1) strain data pre-processing by using a data smoothing technique to remove the trends; 2) diagnosis of novel responses by outlier analysis for the normalized data; and 3) local defect identification by a refined analysis on the novel responses extracted in Step 2. To verify the proposed method, a field test was conducted using a test train incorporating defective wheels. The train ran at different speeds on an instrumented track with the purpose of wheel condition monitoring. By using the proposed method to process the monitoring data, all the defects were identified and the results agreed well with those from the static inspection of the wheelsets in the depot. A comparison is also drawn for the detection accuracy under different running speeds of the test train, and the results show that the proposed method can achieve a satisfactory accuracy in wheel defect detection when the train runs at a speed higher than 30 kph. Some minor defects with a depth of 0.05 mm~0.06 mm are also successfully detected.

• Korean Journal of Materials Research
• /
• v.6 no.7
• /
• pp.733-741
• /
• 1996

In this study MeV Si self ion implantations were done to reveal the intrinsic behavior of defect formation by excluding the possibility of chemical interactions between substrate atoms and dopant ones. Self implantations were conducted using Tandem Accelerator with energy ranges from 1 to 3 MeV. Defect formation by high energy ion implantation has a significant characteristics in that the lattice damage is concentrated near Rp and isolated from the surface. In order to investigate the energy dependence on defect formation, implantation energies were varied from 1 to 3 MeV under a constant dose of $1{\times}10^{15}/cm^2$. RBS channe!ed spectra showed that the depth at which as-implanted damaged layer formed increases as energy increases and that near surface region maintains better crystallinity as energy increases. Cross sectional TEM results agree well with RBS ones. In a TEM image as-implanted damaged layer appears as a dark band, where secondary defects are formed upon annealing. In the case of 2 MeV $Si^+$ self implantation a critical dose for the secondary defect formation was found to be between $3{\times}10^{14}/cm^24$ and $5{\times}10^{14}/cm^2$. Upon annealing the upper layer of the dark band was removed while the bottom part of the dark band did not move. The observed defect behavior by TEM was interpreted by Monte Carlo computer simulations using TRIM-code. SIMS analyses indicated that the secondary defect formed after annealing gettered oxygen impurities existed in silicon.

• Journal of Periodontal and Implant Science
• /
• v.32 no.4
• /
• pp.753-767
• /
• 2002

The purpose of this study was to evaluate histologically the effect of LiF-maleic acid added calcium aluminate(LM-CA) bone cement & CA-PMMA composite bone cement on the healing of calvarial defect in Sprague-Dawley rats. The critical size defects were surgically produced in the calvarial bone using the 8mm trephine bur. The rats were divided in three groups : In the control group, nothing was applied into the defect of each rat. LM-CA bone cement was implanted in the experimental group 1 and CA-PMMA composite bone cement was implanted in the experimental group 2. Rats were sacrificed at 2, 8 weeks after surgical procedure. The specimens were examined by histologic analysis, especially about the bone-cement interface and the response of surrounding tissue. The results are as follows; 1. In the control group, inflammatory infiltration was observed at 2 weeks. At 8 weeks, periosteum and duramater were continuously joined together in the defect area. But the center of defect area was filled up with the loose connective tissue. 2. In the experimental group 1, the bonding between implanted bone cement and the existing bone was seen, which more increased in 8 weeks than 2 weeks. Inflammatory infiltration and the dispersion of implanted bone cement particles were seen in both 2 weeks and 8 weeks. 3. In the experimental group 2, implanted bone itself had a dimensional stability and no bonding between implanted bone cement and the existing bone was seen in both 2 weeks and 8 weeks. Implanted bone cement was encapsulated by fibrous connective tissue. In addition, inflammatory infiltration was seen around implanted bone cement. On the basis of these results, when LM-CA bone cement or CA-PMMA composite bone cement was implanted in rat calvarial defect, LM-CA bone cement can be used as a bioactive bone graft material due to ability of bonding to the existing bone and CA-PMMA can be used as a graft material for augmentation of bone-volume due to dimensional stability.

• Journal of the Korean Society of Safety
• /
• v.37 no.4
• /
• pp.20-27
• /
• 2022

This article aims to develop an algorithm that detects fire pump defects by analyzing the current signals of an induction motor, which are triggered by changes in the flow rate and pressure of multistage volute pumps that are used for fire services. The operational status of the pumps was categorized into three: first, normal operation; second, a defect that is caused by a change in the current value; and third, a defect occasioned by a change in current, pressure, and flow rate. When a fire pump was in normal operation, the motor's operating current was measured between 5.06 A and 6.9 A, the flow rate was estimated at 0-0.27 m³/min, and the pressure ranged from 0 to 0.47 MPa. In the event that a defect was caused by an abnormal current value in the motor, it was attributed to the pump's adherence. Furthermore, if there was no source of water, the defect was considered to have been induced by phase-loss operation, no-load operation, or run-stop operation, with the current value of each scenario being measured at > 52.8 A, < 4.13 A, > 45.15 A, and < 3.8 A, respectively, placing its overall range between 0 and 50 A. The sources of defects were detected based on an analysis of the flow rate, pressure, and current, which represent the following causes: air inflow into the casing, inadequate suction of water, and reverse-phase operation, respectively. Each cause entailed the following values: when air seeped into the casing, the pressure was measured at 0.24 MPa irrespective of changes in the flow rate; when there was inadequate suction of water, the pressure was recorded between 0 and 0.05 MPa despite changes in the flow rate; and when the power line's reverse-phase loss was the cause of the defect, the pressure was measured at 0.33 MPa for a flow rate of 0 L/min, and a higher flow rate decreased the pressure to nearly 0 MPa. The results of this study will enable engineers to develop a pump defect detection algorithm that is based on an analysis of current, and this algorithm will facilitate the execution of a program that will control a fire pump defect detection system.

• Korean Journal of Construction Engineering and Management
• /
• v.18 no.5
• /
• pp.59-67
• /
• 2017

At the completion of the construction, various finishing processes are concentrated. This imposes a burden on the on-site manager and imposes on experience based quality control, thereby causing deviations in the quality of construction depending on supervisor or worker's individual competence. In addition, the information related to quality control is frequently scattered in various types of documents such as specifications and drawings, and checkpoints are frequently omitted. It is necessary to provide a tool that can effectively provide the practitioner before or during the inspection work by systematically storing the information related to the defect prevention and linking them in a mutually referential state. This paper proposes an quality inspection support model that can systematically store necessary information on activity or room basis for the quality check of the apartment house finishing work. Establish a defect prevention information base and a information nexus by linking specifications, design standards, checklists, regulations, defect cases, and drawings to the finishing process and the rooms. Based on this, information registration and search interface are presented. It can contribute to securing a certain level of construction quality or more by suggesting a frame that can be utilized by linking various defects prevention information with the focus on closing activity and room.

• Journal of the Korean Ceramic Society
• /
• v.54 no.4
• /
• pp.323-330
• /
• 2017

To understand the defect chemistry of multiferroic $BiFeO_3-based$ systems, we synthesized nonstoichiometric $Bi_{1+x}FeO_{3{\pm}{\delta}}$ ceramics by conventional solid-state reaction method and studied their structural, dielectric and high-temperature charge transport properties. Incorporation of an excess amount of $Bi_2O_3$ lowered the Bi deficiency in $BiFeO_3$. Polarization versus electric field (P-E) hysteresis loop and dielectric properties were found to be improved by the $Bi_2O_3$ addition. To better understand the defect effects on the multiferroic properties, the high temperature equilibrium electrical conductivity was measured under various oxygen partial pressures ($pO_2{^{\prime}}s$). The charge transport behavior was also examined through thermopower measurement. It was found that the oxygen vacancies contribute to high ionic conduction, showing $pO_2$ independency, and the electronic carrier is electron (n-type) in air and Ar gas atmospheres.

• Transactions of the Korean Society for Noise and Vibration Engineering
• /
• v.24 no.8
• /
• pp.636-642
• /
• 2014

The sensor configuration of the magnetostrictive guided wave system can be described as a single continuous transducing element which makes it difficult to separate the individual modes from the reflected signal. In this work we develop the mode decomposition technique employing chirplet transform, which is able to separate the individual modes from dispersive and multimodal waveform measured with the magnetostrictive sensor, and to estimate the time-frequency centers and individual energies of the reflection, which would be used to locate and characterize defects. The reflection coefficients are calculated using the modal energies of the separated mode. Results from experimental results on a carbon steel pipe are presented, which show that the accurate and quantitative defect characterization could become enabled using the proposed technique.

• Journal of Power System Engineering
• /
• v.20 no.4
• /
• pp.26-31
• /
• 2016

The implementation of chirplet transform to locate axially aligned defects in pipes has been investigated. The results are obtained from experiments performed on a carbon steel pipe using magnetostrictive sensors. Chirplet transform is applied to the reflected signal to separate the individual modes from dispersive and multimodal waveform. The separated modes are used to calculate reflection coefficients which would be used to characterize defects. It is found that the reflection from a defect consists of the wave pulses with gradually decaying amplitudes. Also the results show that the reflection coefficient initially increases with the crack length but finally reaches an oscillating regime.