정오표

강유진,조동진,한대현,임정호,임중빈,오금희,권언혜,Kang, Yoojin,Cho, Dongjin,Han, Daehyeon,Im, Jungho,Lim, Joongbin,Oh, Kum-hui,Kwon, Eonhye 대한원격탐사학회 2022 大韓遠隔探査學會誌 Vol.38 No.2

Role of Linguistic and Extralinguistic Factors in the Acquisition of /w/ Retainment

강유진(Yoojin Kang) 한국사회언어학회 2023 사회언어학 Vol.31 No.3

This study investigates how native Korean speakers, who moved from Kyungsang province to Seoul, acquire /w/ retainment in Seoul Korean (SK). The specific objective is to examine how mobile speakers adopt the SK-like /w/ retainment and evaluate how linguistic and non-linguistic factors influence this acquisition. Furthermore, the study aims to establish whether having explicit awareness of the SK-like /w/ retainment is a prerequisite for acquiring it. The overall results suggest that most speakers have successfully adopted the SK-like /w/ retention, but there is noticeable diversity among speakers in terms of the extent to which they retain /w/. This variability can be linked to linguistic and non-linguistic factors, such as where /w/ appears in speech and individual attitudes towards SK. This research sheds light on the acquisition of /w/ retainment among native Korean speakers who have relocated from Kyungsang province to Seoul, highlighting the impact of linguistic and non-linguistic factors. It underscores the role of explicit awareness and reveals significant interspeaker variation in the adoption of SK-like /w/ retainment, contributing to our understanding of dialect acquisition dynamics.

Phonetic Convergence and Divergence in the Seoul dialect

강유진(Yoojin Kang) 한국사회언어학회 2020 사회언어학 Vol.28 No.3

This study examines whether the lexical pitch accent of the Kyungsang dialect and /wɑ/ monophthongization are imitated by Seoul dialect speakers in an auditory naming task. The goal of the study relates to what is imitated in the phonetic signal. To answer this question, the study examines how relative salience between two linguistic features affects phonetic accommodation. Fifty words as produced by a speaker of the Kyungsang dialect served as stimuli for a shadowing task. The first and second formants of disyllabic words containing the diphthong /wɑ/ and the F0 of disyllabic words with the HL tone were acoustically analyzed for phonetic accommodation. Overall, the results suggest that in terms of the diphthong /wɑ/, participants were more likely to converge toward the Kyungsang model talker, producing their vowel as more monothongized. With respect to the lexical pitch accent, participants were less likely to converge to the model talker, producing the target words with the same tone.

기계학습을 적용한 정지궤도 위성 기반 산불 탐지와 모니터링

강유진(Yoojin Kang),장은나(Eunna Jang),임정호(Jungho Im),심성문(Seongmun Sim) 대한공간정보학회 2018 한국지형공간정보학회 학술대회 Vol.2018 No.11

모의영상을 이용한 농림위성 대기보정의 주요 파라미터 민감도 분석 및 타위성 산출물 활용 가능성 제시

강유진 ( Yoojin Kang ),조동진 ( Dongjin Cho ),한대현 ( Daehyeon Han ),임정호 ( Jungho Im ),임중빈 ( Joongbin Lim ),오금희 ( Kum-hui Oh ),권언혜 ( Eonhye Kwon ) 대한원격탐사학회 2021 大韓遠隔探査學會誌 Vol.37 No.5

차세대 중형위성 사업의 일환으로 농지 및 산림에서의 원격 탐사를 위하여 농림위성 (차세대 중형위성 4호)이 발사 예정에 있다. 위성 영상에서 식생의 정량적인 정보를 얻기 위해서는 대기보정을 통한 지표 반사도 취득이 선행되어야 하므로 농림위성을 위한 대기보정 기술 개발은 불가피할 것으로 생각된다. 특히 대기에서의 흡수와 산란 특성은 파장에 따라 다르게 나타나므로 농림위성 파장 영역을 고려한 대기보정 파라미터 민감도 분석이 필요하다. 또한, 농림위성은 5개 채널(Blue, Green, Red, Red edge, Near-infrared)을 보유하고있어 대기보정 주요 파라미터인 AOD (Aerosol optical depth)와WV (Water vapor)를 직접 산출하기 어려우므로 이를 외부에서 제공할 수 있는 방안을 마련할 필요가 있다. 따라서, 본 연구에서는 농림위성과 유사한 사양을 가진 Sentinel-2 위성 영상을 이용하여 주요 파라미터인 AOD, WV, O₃ 민감도 분석을 수행하고, 파라미터 제공을 위해 천리안 2A (GK2A; GEO-KOMPSAT-2A) 정지궤도 복합위성의 산출물을 이용하여 대기보정 파라미터로서의 활용 가능성을 살펴보았다. 민감도 분석 결과는 AOD가 가장 중요한 파라미터임을 보여주었으며, 근적외선 채널보다는 가시광 채널에서 더 큰 민감도를 가지는 것으로 나타났다. 특히 Blue 채널에서 AOD의 20%의 변화는 지표 반사도에서 약 100%의 오차율을 야기하므로 정확한 지표 반사도 취득을 위해서는 높은 신뢰성을 가진 AOD가 필요할 것으로 생각된다. GK2A AOD 산출물을 이용한 대기보정 결과는 토지피복별 분류 가능성을 이용하여 Sentienl-2 L2A 자료와 비교한 결과, 두 모델별 분류 가능성은 유사하였으나, 파장대가 짧은 영역일 수록 GK2A AOD 산출물을 적용한 대기보정 결과가 Sentinel-2 L2A보다 높게 나타났다. 이를 통해 GK2A에서 제공되는 산출물이 향후 농림위성 대기보정 파라미터로서 충분히 활용될 수 있을 것으로 판단된다. 본 연구의 결과는 추후 농림위성 발사 후 대기보정에 참고 자료로서 활용될 수 있을 것으로 기대된다. As part of the next-generation Compact Advanced Satellite 500 (CAS500) project, CAS500- 4 is scheduled to be launched in 2025 focusing on the remote sensing of agriculture and forestry. To obtain quantitative information on vegetation from satellite images, it is necessary to acquire surface reflectance through atmospheric correction. Thus, it is essential to develop an atmospheric correction method suitable for CAS500-4. Since the absorption and scattering characteristics in the atmosphere vary depending on the wavelength, it is needed to analyze the sensitivity of atmospheric correction parameters such as aerosol optical depth (AOD) and water vapor (WV) considering the wavelengths of CAS500-4. In addition, as CAS500-4 has only five channels (blue, green, red, red edge, and near-infrared), making it difficult to directly calculate key parameters for atmospheric correction, external parameter data should be used. Therefore, this study performed a sensitivity analysis of the key parameters (AOD, WV, and O₃) using the simulated images based on Sentinel-2 satellite data, which has similar wavelength specifications to CAS500-4, and examined the possibility of using the products of GEO-KOMPSAT-2A (GK2A) as atmospheric parameters. The sensitivity analysis showed that AOD was the most important parameter with greater sensitivity in visible channels than in the near-infrared region. In particular, since AOD change of 20% causes about a 100% error rate in the blue channel surface reflectance in forests, a highly reliable AOD is needed to obtain accurate surface reflectance. The atmospherically corrected surface reflectance based on the GK2A AOD and WV was compared with the Sentinel-2 L2A reflectance data through the separability index of the known land cover pixels. The result showed that two corrected surface reflectance had similar Seperability index (SI) values, the atmospheric corrected surface reflectance based on the GK2A AOD showed higher SI than the Sentinel-2 L2A reflectance data in short-wavelength channels. Thus, it is judged that the parameters provided by GK2A can be fully utilized for atmospheric correction of the CAS500-4. The research findings will provide a basis for atmospheric correction of the CAS500-4 in the future.

농림위성을 위한 기계학습을 활용한 복사전달모델기반 대기보정 모사 알고리즘 개발 및 검증: 식생 지역을 위주로

강유진 ( Yoojin Kang ),김예진 ( Yejin Kim ),임정호 ( Jungho Im ),임중빈 ( Joongbin Lim ) 대한원격탐사학회 2023 大韓遠隔探査學會誌 Vol.39 No.5

식생에서의 고해상도 정보를 제공하기 위한 목적으로 농림위성이 발사 예정에 있다. 농림위성 자료를 식생 분야에 활용하기 위해서는 정확한 지표면 반사도 추출이 필수적이므로, 이를 위한 대기보정 기술의 개발과 식생에 대한 검증이 선행되어야 한다. 따라서, 본 연구에서는 Sentinel-2를 대체 위성 영상으로 활용하여 복사전달모델 대기보정을 모사하는 기계학습 기반 대기보정 알고리즘을 개발하고, 산림 지역을 위주로 검증을 수행하였다. Sentinel-2 및 GEOKOMPSAT-2A (GK-2A)에서 추출된 대기보정 파라미터를 기반으로 Random Forest와 Light Gradient Boosting Machine (LGBM)을 통하여 대기보정 알고리즘을 개발하고, 산림에 위치한 총 4개 지역에서의 1년 간의 현장 관측 자료를 이용하여 검증하였다. 두 가지 기계학습 기법 중 효율성을 고려했을 때 LGBM이 탁월할 것으로 판단되었으며, 한 관측소를 제외하면 0.91 이상의 상관계수를 보이며 Normalized Difference Vegetation Index를 활용한 연간 식생 활력도의 시계열 변동을 잘 반영할 수 있음을 확인하였다. 대기보정 필수 파라미터인 Aerosol Optical Depth (AOD) 및 water vapor 수급을 위하여 GK-2A 산출물을 활용할 수 있으나, AOD의 지속적인 결측 문제가 지표면 반사도 산출에 치명적일 수 있기 때문에 이를 보완할 필요가 있다고 사료된다. 본 연구는 농림위성을 위한 대기보정 모사 알고리즘을 개발하고, 한계점과 동시에 보완되어야 할 방향에 대해 제시함으로써 추후 농림위성 자료의 정밀 대기보정을 위한 토대로 기여할 수 있을 것으로 기대된다. Compact Advanced Satellite 500-4 (CAS500-4) is scheduled to be launched to collect high spatial resolution data focusing on vegetation applications. To achieve this goal, accurate surface reflectance retrieval through atmospheric correction is crucial. Therefore, a machine learning-based atmospheric correction algorithm was developed to simulate atmospheric correction from a radiative transfer model using Sentinel-2 data that have similar spectral characteristics as CAS500-4. The algorithm was then evaluated mainly for forest areas. Utilizing the atmospheric correction parameters extracted from Sentinel-2 and GEOKOMPSAT-2A (GK-2A), the atmospheric correction algorithm was developed based on Random Forest and Light Gradient Boosting Machine (LGBM). Between the two machine learning techniques, LGBM performed better when considering both accuracy and efficiency. Except for one station, the results had a correlation coefficient of more than 0.91 and well-reflected temporal variations of the Normalized Difference Vegetation Index (i.e., vegetation phenology). GK-2A provides Aerosol Optical Depth (AOD) and water vapor, which are essential parameters for atmospheric correction, but additional processing should be required in the future to mitigate the problem caused by their many missing values. This study provided the basis for the atmospheric correction of CAS500-4 by developing a machine learning-based atmospheric correction simulation algorithm.

Dialect Contact and Linguistic Accommodation

Yoojin Kang(강유진) 언어과학회 2016 언어과학연구 Vol.0 No.79

This study investigates standard Seoul Korean speakers’ linguistic accommodation in dialect contact situations. In particular, the study examines how ‘gender’, ‘age’, and ‘length of residency in Gyeongsang’ can affect participants’ accommodation in dialect contact. The study also examines speakers’ motivations for accommodation in dialect contact. In addition, this study investigates correlations of participants’ accommodative acts and attitudes towards both the Gyeongsang dialect and standard Seoul Korean as well as attitudes towards Gyeonsang Province and Seoul. To find this correlation, the present study collected two kinds of data : phonetic data and identity assessment data. Participants carried out three tasks: TEST1, designed for investigating phonological variation, TEST2, designed for exploring lexical pitch accents, and an identity assessment questionnaire.

Himawari-8 정지궤도 위성 영상을 활용한 딥러닝 기반 산불 탐지의 효율적 방안 제시

이시현,강유진,성태준,임정호,Sihyun Lee,Yoojin Kang,Taejun Sung,Jungho Im 대한원격탐사학회 2023 大韓遠隔探査學會誌 Vol.39 No.5

As wildfires are difficult to predict, real-time monitoring is crucial for a timely response. Geostationary satellite images are very useful for active fire detection because they can monitor a vast area with high temporal resolution (e.g., 2 min). Existing satellite-based active fire detection algorithms detect thermal outliers using threshold values based on the statistical analysis of brightness temperature. However, the difficulty in establishing suitable thresholds for such threshold-based methods hinders their ability to detect fires with low intensity and achieve generalized performance. In light of these challenges, machine learning has emerged as a potential-solution. Until now, relatively simple techniques such as random forest, Vanilla convolutional neural network (CNN), and U-net have been applied for active fire detection. Therefore, this study proposed an active fire detection algorithm using state-of-the-art (SOTA) deep learning techniques using data from the Advanced Himawari Imager and evaluated it over East Asia and Australia. The SOTA model was developed by applying EfficientNet and lion optimizer, and the results were compared with the model using the Vanilla CNN structure. EfficientNet outperformed CNN with F1-scores of 0.88 and 0.83 in East Asia and Australia, respectively. The performance was better after using weighted loss, equal sampling, and image augmentation techniques to fix data imbalance issues compared to before the techniques were used, resulting in F1-scores of 0.92 in East Asia and 0.84 in Australia. It is anticipated that timely responses facilitated by the SOTA deep learning-based approach for active fire detection will effectively mitigate the damage caused by wildfires.

GK-2A 위성 영상과 기계 학습 기법을 이용한 실시간 산불 탐지

성태준(Taejun Sung),강유진(Yoojin Kang),임정호(Jung-Ho Im) 한국통신학회 2022 한국통신학회 학술대회논문집 Vol.2022 No.6

압전재료 기반 지능형 조종익의 구조 및 서보공력탄성학 응답 해석

박철우(Chul-Woo Park),강유진(YooJin Kang),빈영빈(Young-Bin Bin),신상준(SangJoon Shin),윤범수(Bum Soo Yoon),박지원(JiWon Park),윤광준(Kwang Joon Yoon),최현영(HyunYoung Choi) 한국항공우주학회 2016 한국항공우주학회 학술발표회 논문집 Vol.2016 No.11