https://scholar.dgist.ac.kr/handle/20.500.11750/15727 2026-04-04T10:19:01Z https://scholar.dgist.ac.kr/handle/20.500.11750/59983 Title: YOLO 및 회귀 모델을 활용한 GPR B-scan 영상 기반 공동 크기 및 위치 추정 Author(s): 김세경; 양경택; 송승언; 이종훈 Abstract: Ground Penetrating Radar (GPR) is a non-destructive technique used to detect subsurface voids. In this studypaper, synthetic B-scan data hiddenwithwith spherical voids were generated using gprMax, and a YOLOv8-s model was trained to detect hyperbolic patterns. Since YOLO provides bounding boxes in pixel-level coordinates, regression mapping to ground-truth (GT) labels was applied to enable real-world size estimation. The predicted bounding boxes were mapped to Ground Truth (GT) labels through regression to evaluate void size inference.Since YOLO provides bounding boxes in pixel-level coordinates, a regression-based mapping was developed to convert detection results into real-world metric (m) coordinates, enabling direct estimation of void size and position. Results show that the predicted center coordinates (X, Y) exhibited linearity with GT values (R²>0.99). The bounding box height correlated with GT (R²=0.98). Also, the bounding box width and depth (Y), the estimation of GT width shows (R²≈ 0.98). These results demonstrate that YOLOv8-s based detection can reliably support both void localization and quantitative size estimation in GPR B-scan images.These results demonstrate that the proposed method can reliably support both void localization and real-size estimation in GPR B-scan images. Thus, the proposed approach provides an automated pixel-to-meter translation framework for quantitative GPR interpretation. 2025-11-05T15:00:00Z https://scholar.dgist.ac.kr/handle/20.500.11750/59390 Title: A Multi-template Correlation Time Delay Estimation in GPR System under a Building Collapse Model Author(s): Yang, Gyeongtaeg; Kim, Sekyung; Song, Seungeon; Han, Seonho; Koo, Bontae; Lee, Jonghun Abstract: This paper proposes a multi-template correlationbased time delay estimation method that enhances the accuracy and robustness of ground-penetrating radar (GPR) signal interpretation in noisy environments. Unlike conventional approaches that rely on a single reference waveform, the proposed method utilizes the original signal along with its higher-order derivatives to better match diverse echo patterns. Simulation experiments using a four-layer collapse model generated in gprMax confirm that the proposed method achieves lower mean percentage error (MPE) than traditional techniques under low signal-to-noise ratio (SNR) conditions. By addressing the challenges of echo distortion and environmental noise, the proposed approach demonstrates strong potential for reliable subsurface detection in post-disaster search and rescue scenarios. 2025-10-03T15:00:00Z https://scholar.dgist.ac.kr/handle/20.500.11750/59008 Title: Design and Development of the 24 GHz FMCW Radar Sensor for Blind Spot Detection and Lane Change Assistance Systems Author(s): Ju, Yeonghwan; Kim, Sangdong; Jin, Youngseok; Lee, Jonghun Abstract: In this paper, we designed and implemented the automotive radar based on Frequency Modulated Continuous Wave method for Lane Change Assist and Blind Spot Detection radar system. We also developed digital signal processing module and a 24 GHz FMCW radar RF module composed of a single transmitter, a single transmitting antenna array of five elements and three receivers to measure range, velocity and angle for LCA and BSD. In order to verify the developed radar system, we conducted experiments to measure the detection rate and the range accuracy in the anechoic chamber. The experimental results show that the developed radar is feasible for use in BSD/LCA systems for automotive. 2017-05-23T15:00:00Z https://scholar.dgist.ac.kr/handle/20.500.11750/58415 Title: 2D Spatial-Temporal Simulation of GPR Penetration in the 200-400MHz Band for Detecting Voids and Entrapped Persons in Multi-Layered Collapsed Building Structure Author(s): Song, Seungeon; Yang, Gyeongtaeg; Kim, Bongseok; Kim, Sangdong; Lee, Jonghun Abstract: This paper presents a foundational framework towards designing a portable, site-specific radar system for use in building collapse scenarios, focusing on the detection of voids beneath debris and analyzing GPR signals for locating hidden survivors. To analyze GPR signal characteristics, we employed the Finite Difference Time Domain (FDTD) method, utilizing simulation tools such as gprMax and High Frequency Structure Simulator (HFSS) to model electromagnetic wave propagation within intricate multilayered debris structures. This paper examines the GPR A-scan, B-scan(radargram) and signal propagations across various configurations of collapse site structures, ranging from solely soil compositions to those incorporating thin or thick concrete and wooden layers, each with embedded voids and entrapped persons inside voids.This work aims to enhance the efficiency of survivor search and rescue operations at collapse sites, laying down a theoretical basis for validating experiments designed by real-world collapse scenarios. By scrutinizing the influence of GPR's operating frequency and the structural composition of debris on detection performance, the findings serve as crucial foundational data for tailoring GPR design and understanding detection characteristics pertinent to real-world conditions. 2024-10-23T15:00:00Z