인문학
사회과학
자연과학
공학
의약학
농수해양학
예술체육학
복합학
지원사업
학술연구/단체지원/교육 등 연구자 활동을 지속하도록 DBpia가 지원하고 있어요.
커뮤니티
연구자들이 자신의 연구와 전문성을 널리 알리고, 새로운 협력의 기회를 만들 수 있는 네트워킹 공간이에요.
초록·키워드
Hotspots on PV (PhotoVoltaic) modules reduce energy yield and risk physical damage, but reliable detection in UAV (Unmanned Aerial Vehicle)–based TIR (Thermal Infrared) imagery is hindered by viewing-angle effects. This study investigates how changes in viewing direction (−90° to +90° in 10° steps) alter radiometric responses even after geometric rectification. Using CLAHE (contrast-limited adaptive histogram equalization)–assisted SuperPoint–SuperGlue matching and RANSAC (Random sample consensus) homography, 19 images were aligned to a 0° reference coordinate system with a mean RMSE (Root-Mean-Square Error) of approximately 0.74 pixels. Hotspots within a fixed ROI (Region of Interest) were segmented and assessed via IoU (Intersection over Union), soft-IoU, DIoU (Distance-IoU), centroid displacement, and area ratio. IoU peaked at the frontal view and declined with increasing |angle|, while Soft-IoU reduced boundary errors but not internal intensity shifts. DIoU often overestimated compact detections, and centroid displacement grew with obliquity, indicating radiometric drift. The area ratio fluctuated unpredictably. These outcomes confirm that geometric correction alone cannot counteract incidence-angle–driven radiometric distortions that obscure true hotspot morphology and intensity. The study recommends near-frontal image acquisition and incidence-aware temperature normalization to enhance reproducibility and operational reliability in UAV–TIR hotspot inspection.
본문·목차
인공지능 문자 인식 모델을 통해 추출된 텍스트로, 일부 오타나 오류가 포함될 수 있으나 지속적으로 개선 중입니다.
오류를 발견하셨다면 해당 부분을 드래그한 후 ' 를 통해 신고해주세요.
오류를 발견하셨다면 해당 부분을 드래그한 후 ' 를 통해 신고해주세요.
최근 본 자료 전체보기
UCI(KEPA) : I410-151-26-02-094499540