Hyeokjin Hong

Hyeokjin Hong

Auto-generated profile, based on publicly available information

Hyeokjin Hong

Department of Climate-Social Science Convergence, Kyung Hee University, Seoul, Republic of Korea

Research Focus & Metrics

Academic Career Length

1Year (2026-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
30 October 2026 in Journal of Earth Observations and Geospatial Applications
Verified
Journal Article
Closed access
PDF file
HNHyunwoo Nam
Hyunwoo Nam
KRKasim Richards
Kasim Richards
LSLily Sims
Lily Sims
HHHyeokjin Hong
Hyeokjin Hong
SCSeung Hee Choi
Seung Hee Choi
Jeong Chang SeongJeong Chang Seong
Jeong Chang Seong
Satellite-based nighttime light imagery offers a powerful means of monitoring fishing activity across data-sparse ocean regions, yet the influence of image standardization on machine-learning-based pattern classification has received little attention. This study analyzed Visible Infrared Imaging Radiometer Suite Day/Night Band (VIIRS DNB) monthly composite imagery of the Korean Exclusive Economic Zone (EEZ) from 2014 to 2024, applying an identical Convolutional Autoencoder and Gaussian Mixture Model (CAE-GMM) framework under both local and global standardization. Global standardization produced three or four major spatiotemporal clusters that broadly corresponded to seasonal fishing cycles: a high-intensity autumn–winter pattern, an expanded summer–autumn pattern, and a low-intensity winter–spring pattern. In the four-cluster solution, the fourth cluster represented the transition periods among the three seasonal patterns. Particularly, the strong autumn-winter pattern weakened after 2019 and was increasingly replaced by winter-spring clusters. Comparison of the two standardization methods showed that local standardization categorizes images primarily by their internal spatial light distribution, whereas global standardization additionally reflects brightness differences across the time series. As a result, images with similar spatial patterns but lower intensity, such as January through March in recent years, were assigned to a different cluster. These findings demonstrate that standardization is not just a preprocessing step but a decisive methodological choice that depends on the research objective in fisheries monitoring with nighttime lights.