• Proceedings of the Korea Information Processing Society Conference
• /
• 2021.11a
• /
• pp.26-27
• /
• 2021

본 논문에서는 딥러닝 영상기술을 활용해 파프리카 잎에서 나타나는 병해를 분류하는 연구를 진행하였다. 비파괴 방법으로 파프리카 잎 뒷면을 촬영하면 잎을 잡는 손이 파프리카 잎을 가리는 영역이 부분적으로 나타나고, 이는 학습을 방해하는 요소가 된다. 이를 해결하기 위해 잎의 영역을 먼저 찾고 그 외의 배경영역을 없애고, 병해를 진단할 수 있도록 모델을 설계하였다. 잎의 영역을 찾아내는 모델은 86.7%의 IoU(Intersection over Union)의 값을 얻었고, 병해를 진단하는 분류 정확도는 86.4%을 얻었다.

• International Journal of Highway Engineering
• /
• v.11 no.2
• /
• pp.17-25
• /
• 2009

The goal of this study is to develop Logistic regression model by accident location(entry section, exit section, inside intersection and pedestrian crossing section). Based on the accident data of Chungbuk Provincial Police Agency(2004$\sim$2005) and the field survey data, the geometric elements, environmental factor and others related to traffic accidents were analyzed. Developed models are all analyzed to be statistically significant(chi-square p=0.000, Nagelkerke $R^2$=0.363$\sim$0.819). The models show that the common factors of accidents are the traffic volume(ADT), distant of crossing and exclusive left turn lane, and the specific factors are the minor traffic volume(inside intersection model) and U-turn of main road(pedestrian crossing model). Hosmer & Loineshow tests are evaluated to be statistically significant(p$\geqq$0.05) except the entry section model. The correct classification rates are also analyzed to be very predictable(more than 73.9% to all models).

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.22 no.1
• /
• pp.81-92
• /
• 2023

Traffic state classification is crucial for time-of-day (TOD) traffic signal control. This paper proposed a traffic state classification technique applying Deep-Embedded Clustering (DEC) method that uses a high dimensional traffic data observed at all signalized intersections in a traffic signal control sub area (SA). So far, signal timing plan has been determined based on the traffic data observed at the critical intersection in SA. The current method has a limitation that it cannot consider a comprehensive traffic situation in SA. The proposed method alleviates the curse of dimensionality and turns out to overcome the shortcomings of the current signal timing plan.

• Journal of the Society of Naval Architects of Korea
• /
• v.58 no.5
• /
• pp.303-313
• /
• 2021

Recently, quality control of the Liquefied Natural Gas Carrier (LNGC) cargo hold and block-erection interference areas using 3D scanners have been performed, focusing on large shipyards and the international association of classification societies. In this study, as a part of the research on LNGC cargo hold quality management advancement, a study on deep-learning-based scaffolding system 3D point cloud object detection and post-processing were conducted using a LNGC cargo hold 3D point cloud. The scaffolding system point cloud object detection is based on the PointNet deep learning architecture that detects objects using point clouds, achieving 70% prediction accuracy. In addition, the possibility of improving the accuracy of object detection through parameter adjustment is confirmed, and the standard of Intersection over Union (IoU), an index for determining whether the object is the same, is achieved. To avoid the manual post-processing work, the object detection architecture allows automatic task performance and can achieve stable prediction accuracy through supplementation and improvement of learning data. In the future, an improved study will be conducted on not only the flat surface of the LNGC cargo hold but also complex systems such as curved surfaces, and the results are expected to be applicable in process progress automation rate monitoring and ship quality control.

• Journal of the Architectural Institute of Korea Planning & Design
• /
• v.34 no.11
• /
• pp.145-156
• /
• 2018

Urban campus, one of the main urban facilities, is the representative place that is struggling with 'gridlock'. Due to privatization of space among different departments and space shortages, gridlock has been occurring as a result. The urban campus trying to solve this problem by changing the quality of space, especially the structure of the shared space, which is expected to be the solution to the grid lock problem. The main purpose of this study is to investigate the structural change in the university's shared space based on paradigm transition. The theoretical consideration is to analyze the spatial characteristics of university shared space that appear at different stages through a new perspective that compares the gridlock phenomenon and the shared paradigm. The framework of the analysis of the shared space, which has recently been restructured, is classified into the spatial characteristics of collaborative space, the creative space, and the common/complex space. In addition, these spatial characteristics are again analyzed through the division of legislative facility classification, management governance subject, area, building location and layout, exposure to the outside as well as the analysis of student and staff entry and exit, sharing structure of site and space, and the classification of program characteristics. The results are as follows: The restructured space is systemized so that the management governance of each space would be connected to each other to share information and space. Furthermore, the spatial boundary between colleges or between campus spaces are not only physically, but categorically clear. The restructured space has semi (or in-between)-spatial characteristics such as the intersection in inside and outside of the pedestrian's circulation and the mixture of programs. This study could serve as principal references in presenting the systematic analysis of directions of the shared spatial structure for the urban campus where new educational space is required due to the changes in the university system.

• Geomechanics and Engineering
• /
• v.38 no.1
• /
• pp.29-44
• /
• 2024

This research investigates the effectiveness of data augmentation techniques in the automated analysis of B-scan images from ground-penetrating radar (GPR) using deep learning. In spite of the growing interest in automating GPR data analysis and advancements in deep learning for image classification and object detection, many deep learning-based GPR data analysis studies have been limited by the availability of large, diverse GPR datasets. Data augmentation techniques are widely used in deep learning to improve model performance. In this study, we applied four data augmentation techniques (geometric transformation, color-space transformation, noise injection, and applying kernel filter) to the GPR datasets obtained from a testbed. A deep learning model for GPR data analysis was developed using three models (Faster R-CNN ResNet, SSD ResNet, and EfficientDet) based on transfer learning. It was found that data augmentation significantly enhances model performance across all cases, with the mAP and AR for the Faster R-CNN ResNet model increasing by approximately 4%, achieving a maximum mAP (Intersection over Union = 0.5:1.0) of 87.5% and maximum AR of 90.5%. These results highlight the importance of data augmentation in improving the robustness and accuracy of deep learning models for GPR B-scan analysis. The enhanced detection capabilities achieved through these techniques contribute to more reliable subsurface investigations in geotechnical engineering.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.34 no.1
• /
• pp.25-33
• /
• 2021

With the development of machine learning techniques, various types of data such as vibration, temperature, and flow rate can be used to detect and diagnose abnormalities in machine conditions. In particular, in the field of the state monitoring of rotating machines, the fault diagnosis of machines using vibration data has long been carried out, and the methods are also very diverse. In this study, an experiment was conducted to collect vibration data from normal and abnormal compressors by installing accelerometers directly on rotary compressors used in household air conditioners. Data segmentation was performed to solve the data shortage problem, and the main features for the fault classification model were extracted through the chi-square test after statistical and physical features were extracted from the vibration data in the time domain. The support vector machine (SVM) model was developed to classify the normal or abnormal conditions of compressors and improve the classification accuracy through the hyperparameter optimization of the SVM.

• Korean Journal of Remote Sensing
• /
• v.39 no.6_3
• /
• pp.1679-1692
• /
• 2023

South Korea is a country that emits a large amount of pollutants as a result of population growth and industrial development and is also severely affected by transboundary air pollution due to its geographical location. As pollutants from both domestic and foreign sources contribute to air pollution in Korea, the location of air pollutant emission sources is crucial for understanding the movement and distribution of pollutants in the atmosphere and establishing national-level air pollution management and response strategies. Based on this background, this study aims to effectively acquire spatial information on domestic and international air pollutant emission sources, which is essential for analyzing air pollution status, by utilizing high-resolution optical satellite images and deep learning-based image segmentation models. In particular, industrial parks and quarries, which have been evaluated as contributing significantly to transboundary air pollution, were selected as the main research subjects, and images of these areas from multi-purpose satellites 3 and 3A were collected, preprocessed, and converted into input and label data for model training. As a result of training the U-Net model using this data, the overall accuracy of 0.8484 and mean Intersection over Union (mIoU) of 0.6490 were achieved, and the predicted maps showed significant results in extracting object boundaries more accurately than the label data created by course annotations.

• Korean Journal of Remote Sensing
• /
• v.40 no.4
• /
• pp.343-350
• /
• 2024

Since the release of Meta's Segment Anything Model (SAM), a large-scale vision transformer generation model with rapid image segmentation capabilities, several studies have been conducted to apply this technology in various fields. In this study, we aimed to investigate the applicability of SAM for water bodies detection and extraction using the QGIS Geo-SAM plugin, which enables the use of SAM with satellite imagery. The experimental data consisted of Compact Advanced Satellite 500 (CAS500)-1 images. The results obtained by applying SAM to these data were compared with manually digitized water objects, Open Street Map (OSM), and water body data from the National Geographic Information Institute (NGII)-based hydrological digital map. The mean Intersection over Union (mIoU) calculated for all features extracted using SAM and these three-comparison data were 0.7490, 0.5905, and 0.4921, respectively. For features commonly appeared or extracted in all datasets, the results were 0.9189, 0.8779, and 0.7715, respectively. Based on analysis of the spatial consistency between SAM results and other comparison data, SAM showed limitations in detecting small-scale or poorly defined streams but provided meaningful segmentation results for water body classification.

• Journal of Korean Society of Steel Construction
• /
• v.9 no.2 s.31
• /
• pp.237-248
• /
• 1997

Three types of fatigue cracks frequently observed in the crane runway girders are verified experimentally using two testing-purpose girders with the size of $6400{\times}600{\times}300$ in millimeters. The fatigue cracks are observed in the vicinity of load-bearing points, at the end of gusset plates and at the fillet welded joints between the lower flange and the web. The load-bearing-point cracks are initiated at the intersection of the fillet welds between the upper flange and the web, where the vertical stiffener is located. The cracks grow up toward the diagonal direction of the web. The cracks observed at the fillet welded joints grow up perpendicularly to the crane runway girder. Compared with the JSSC fatigue design code, the joint class is classified as follows: E for the vicinity of load-bearing points, G or H for the end of gusset plates and D for the lower fillet welded joints. The tests reveal that the class of joint classification at the end of gusset plates and at the lower flange coincides with the fatigue design code.