노후 구조물의 동적 특성 평가를 위해 대표적인 접촉식 센서인 LVDT와 가속도계를 활용한다. 전통 적인 센서의 데이터 신뢰성은 높지만, 작동 원리로 인하여 대상 구조물의 물리적 접근과 센서의 설치 가 필요하다. 조밀한 센서 설치를 위해선 많은 수의 센서와 데이터 수집장치도 추가적으로 필요하다. 이런 단점을 보완하고자 비접촉식 센서의 개발이 활발히 진행 중이며, 특히 비전센서를 활용한 동적 변위 측정에 관한 연구 및 개발에 많은 진척이 있다. 비전센서를 활용한 동적 변위 측정 시스템은 내 적 파라미터 및 외적 환경조건에 따라 측정 정확도가 크게 변화한다. 주된 내적 파라미터로 영상장비 의 공간분해능은 이미지 센서의 물리적 크리와 촬영거리의 증가 혹은 관심영역이 작아짐에 따라 영향 을 많이 받는다. 외적 환경조건으로 저조도 환경에서 타겟의 밝기차이가 줄어들어 이미지 프로세싱 과 정에서 불리한 조건이며, 이는 동적 변위 측정 정확도 저하로 이어진다. 본 연구에서 저조도 환경에서 비전센서의 운용거리 한계를 초해상화를 적용하여 극복하고자 하며, 인공적 및 자연적 타겟에 대한 동 적 변위 측정 성능을 비교 분석하였다. 동적 변위 측정 실내 실험을 위해 저조도 조건에서 3층 전단 구조물을 9Hz로 가진하였다. 동시에 Sony사의 DSC-100M7 카메라를 활용하여 조화진동으로 인해 발생되는 각층의 변위를 FHD화질 120FPS로 촬영하였고 측정 정확도 비교 분석을 각층의 LVDT 측 정값으로 진행하였다. 촬영거리를 10m를 기준으로 10m씩 증가하면서 최대 40m까지 변위 값을 측정 하였으며, 공간분해능 증가를 위해 GAN기반 초해상화 모델인 RealSR을 적용시켰다. 초해상화를 활용 하여 동적변위를 측정한 결과 저조도 환경에서도 비전센서의 운용 거리가 증가함을 확인할 수 있었으 며, 동시에 변위 측정 정확도도 함께 상승하는 것을 보여줬다.

The second generation of stars in the globular clusters (GCs) of the Milky Way (MW) exhibit unusually high N, Na, or Al, compared to typical Galactic halo stars at similar metallicities. The halo field stars enhanced with such elements are believed to have originated in disrupted GCs or escaped from existing GCs. We identify such stars in the metallicity range −3.0 < [Fe/H] < 0.0 from a sample of ∼36,800 giant stars observed in the Sloan Digital Sky Survey and Large Sky Area Multi-Object Fiber Spectroscopic Telescope survey, and present their dynamical properties. The N-rich population (NRP) and N-normal population (NNP) among our giant sample do not exhibit similarities in either in their metallicity distribution function (MDF) or dynamical properties. We find that, even though the MDF of the NRP looks similar to that of the MW’s GCs in the range of [Fe/H] < −1.0, our analysis of the dynamical properties does not indicate similarities between them in the same metallicity range, implying that the escaped members from existing GCs may account for a small fraction of our N-rich stars, or the orbits of the present GCs have been altered by the dynamical friction of the MW. We also find a significant increase in the fraction of N-rich stars in the halo field in the very metal-poor (VMP; [Fe/H] < −2.0) regime, comprising up to ∼20% of the fraction of the N-rich stars below [Fe/H] = −2.5, hinting that partially or fully destroyed VMP GCs may have in some degree contributed to the Galactic halo. A more detailed dynamical analysis of the NRP reveals that our sample of N-rich stars do not share a single common origin. Although a substantial fraction of the N-rich stars seem to originate from the GCs formed in situ, more than 60% of them are not associated with those of typical Galactic populations, but probably have extragalactic origins associated with Gaia Sausage/Enceladus, Sequoia, and Sagittarius dwarf galaxies, as well as with presently unrecognized progenitors.

We present a method to determine nitrogen abundance ratios with respect to iron ([N/Fe]) from molecular CN-band features observed in low-resolution (R ∼ 2000) stellar spectra obtained by the Sloan Digital Sky Survey (SDSS) and the Large Sky Area Multi-Object Fiber Spectroscopic Telescope (LAMOST). Various tests are carried out to check the systematic and random errors of our technique, and the impact of signal-to-noise (S/N) ratios of stellar spectra on the determined [N/Fe]. We find that the uncertainty of our derived [N/Fe] is less than 0.3 dex for S/N ratios larger than 10 in the ranges Teff = [4000, 6000] K, log g = [0.0, 3.5], [Fe/H] = [−3.0, 0.0], [C/Fe] = [−1.0, +4.5], and [N/Fe] = [−1.0, +4.5], the parameter space that we are interested in to identify N-enhanced stars in the Galactic halo. A star-by-star comparison with a sample of stars with [N/Fe] estimates available from the Apache Point Observatory Galactic Evolution Experiment (APOGEE) also suggests a similar level of uncertainty in our measured [N/Fe], after removing its systematic error. Based on these results, we conclude that our method is able to reproduce [N/Fe] from low-resolution spectroscopic data, with an uncertainty sufficiently small to discover N-rich stars that presumably originated from disrupted Galactic globular clusters.

Deep learning techniques have been studied and developed throughout the medical, agricultural, aviation, and automotive industries. It can be applied to construction fields such as concrete cracks and welding defects. One of the best performing techniques of deep running is CNN technique. CNN means convolutional neural network. In this study, we analyzed crack recognition of sewer with low recognition. Deep learning is generally more accurate with deeper layers, but analysis cost is high. In addition, many variations can occur depending on training options. Therefore, this study performed many parametric studies according to the variations of training options. When analyzed with appropriate training options, the accuracy was over 90% and stable results were obtained

Recent narrow-band Ca photometry discovered two distinct red giant branch (RGB) populations in some massive globular clusters (GCs) including M22, NGC 1851, and NGC 288. In order to investigate the dierences in light/heavy elements abundances between the two subpopulations, we have performed low-resolution spectroscopy for stars on the two RGBs in these GCs. We find a significant difference (more than 4) in calcium abundance from the spectroscopic HK0 index for both M22 and NGC 1851. We also find a more than 8 difference in CN band strength between the Ca-strong and Ca-weak subpopulations. For NGC 288, however, we detect the presence of a large difference only in the CN strength. The calcium abundances of the two subpopulations in this GC are identical within errors. We also find interesting differences in CN-CH relations among these GCs. While CN and CH indices are correlated in M22, they show an anti-correlation in NGC 288. However, NGC 1851 shows no dierence in CH between two groups of stars having different CN strengths. The CN bimodality in these GCs could be explained by pollution from intermediate-mass asymptotic giant branch stars and/or fast-rotating massive stars. For the presence or absence of calcium bimodality and the differences in CN-CH relations, we suggest these would be best explained by how strongly type II supernovae enrichment has contributed to the chemical evolutions of these GCs.

In this study, we adapted very low resolution (240x320 = 76,800 pixels) images by which it is difficult to detect cracks. An automatic crack detection technique has been studied using digital image processing technology for low resolution images of sewage pipeline. Authors have developed two algorithms to detect cracks. The third step covers an algorithm developed to find optimal threshold value, and the sixth step deals with algorithm to determine cracks. As the result, in spite of very low-resolution images, the performance of crack detection turned out to be excellent.

최근 이슈가 되고 있는 도심지 지반함몰로 인하여 주기적인 하수관로 조사의 필요성이 강조되고 있다. 일반적으로 수행되는 조사 방법 중 하나인 하수관로 CCTV조사는 상당한 시간과 노력이 소요된다. 기존 연구들은 주로 하수관로 조사에 소요되는 노력을 줄이기 위한 H/W 및 S/W의 개발에 관한 연구가 주를 이루고 있다. 그러나 기존 CCTV 탐사장치를 이용하여 관리담당자가 보관하고 있는 수많은 조사영상 를 활용하기 위한 연구는 진행되지 않았다. 본 연구는 cross-correlation기법 기반의 이미지프로세싱 방법을 적용하여 CCTV 조사영상의 자막 으로부터 장치의 위치정보를 추출하였다. CCTV 장치의 시간-거리 관계를 분석한 결과 탐사 장치가 정지시간과 하수관로의 손상 사이의 강한 상관관계를 확인하였다. 제안된 CCTV영상의 분석법을 활용하는 경우 CCTV조사 보고서 작성 및 관리에 소요되는 노력을 줄임으로써 하수관 로 유지관리의 효율성과 신뢰도를 높일 수 있을 것으로 기대된다.

This paper proposes a novel face detection method that finds tiny faces located at a long range even with low-resolution input images captured by a mobile robot. The proposed approach can locate extremely small-sized face regions of 12x12 pixels. We solve a tiny face detection problem by organizing a system that consists of multiple detectors including a mean-shift color tracker, short- and long-rage face detectors, and an omega shape detector. The proposed method adopts the long-range face detector that is well trained enough to detect tiny faces at a long range, and limiting its operation to only within a search region that is automatically determined by the mean-shift color tracker and the omega shape detector. By focusing on limiting the face search region as much as possible, the proposed method can accurately detect tiny faces at a long distance even with a low-resolution image, and decrease false positives sharply. According to the experimental results on realistic databases, the performance of the proposed approach is at a sufficiently practical level for various robot applications such as face recognition of non-cooperative users, human-following, and gesture recognition for long-range interaction.