메뉴 건너뛰기
.. 내서재 .. 알림
소속 기관/학교 인증
인증하면 논문, 학술자료 등을  무료로 열람할 수 있어요.
한국대학교, 누리자동차, 시립도서관 등 나의 기관을 확인해보세요
(국내 대학 90% 이상 구독 중)
로그인 회원가입 고객센터 ENG
주제분류

추천
검색

논문 기본 정보

자료유형
학위논문
저자정보

강남우 (고려대학교, 고려대학교 대학원)

지도교수
조훈희
발행연도
2022
저작권
고려대학교 논문은 저작권에 의해 보호받습니다.

이용수11

표지
AI에게 요청하기
추천
검색

이 논문의 연구 히스토리 (2)

초록· 키워드

오류제보하기
국내 노후 건축물 비중이 증가함에 따라 건축물의 유지관리를 위해 건축물의
생애 이력 정보체계 구축의 필요성이 부각되고 있다. 특히, 건축물 관리법 제 7조
에 따르면 유지관리를 위해 건축물의 생애 이력 정보체계 구축을 요구하고 있으
나, 유지관리가 필요한 노후 건축물은 도면이 없거나 소실된 경우가 대부분이다.
또한, 리모델링 시장이 성장함에 따라 기존 건축물의 As-built 도면 획득 방안에
관한 관심이 증대되고 있다.
이에 도면이 소실되거나, BIM을 구축하지 않은 건축물을 대상으로 실내
As-Built BIM 구축을 위해 스캐너를 통해 실내를 역설계하는 방안이 주목받고
있으며 이와 관련하여 많은 연구가 진행되고 있다. 이러한 As-Built BIM을 구
축하기 위해 건축 부재의 형상뿐만 아니라 부재의 종류를 파악해야 하며, 이에
건축 부재 객체 인식이 필요하다.
하지만 실내는 가구와 사람과 같은 다양한 객체의 간섭으로 인해 벽과 같은 건
축 부재의 포인트가 온전하게 획득되지 않은 폐색영역이 발생하며 정확한 객체
인식을 어렵게 한다. 특히 열린 문과 창문과 같은 부재는 포인트가 습득되지 않
은 부재이기 때문에 객체 인식에 어려움이 있다.
이에 본 연구에서는 객체로 인해 간섭된 실내 3차원 포인트 클라우드 데이터
를 대상을 건축 부재가 서로 직교한다는 가정인 맨해튼 가정과 픽셀 기반의 인식
을 통해 벽, 천장, 바닥, 문을 인식하는 건축 부재 인식 알고리즘을 제안하였다.
본 연구에서는 픽셀 기반의 건축 부재 객체 인식 알고리즘을 제안하였다. 천장
의 픽셀을 기반으로 벽면에 해당하는 픽셀을 도출한 뒤, 벽면 픽셀의 해상도를
낮추어서 벽면을 분류하였다. 분류된 벽면과 인접한 복셀의 포인트를 벽면으로
정사영 하여 폐색영역을 제거한 뒤 벽면의 픽셀이 없는 개구부의 치수를 기반으
로 문을 인식하였다.
본 연구는 문 인식 성능의 정밀도 및 재현율 검사를 통해 알고리즘의 성능을
검증하였으며, F1 score 91.23%로 준수한 인식 성능을 보였다. 본 알고리즘은
포인트 클라우드 데이터의 3차원 좌푯값만을 활용하기 때문에 실내 역설계를 위
한 스캐너 종류의 폭을 넓히는데 기여할 수 있다고 판단되며, 건축 부재와 문과
같은 개구부를 자동으로 인식하고 분류할 수 있기 때문에 As-built BIM 구축에
필요한 기초 자료를 제공하는 데에 기여할 수 있을 것으로 기대된다.

목차

제1장. 서 론 ················································································································ 1
1.1 연구의 배경 및 목적 ······················································································· 1
1.2 연구의 범위 및 절차 ······················································································· 3
제2장. 예비적 고찰 ······································································································ 4
2.1 포인트 클라우드 데이터의 개요 및 특성 ··················································· 4
2.1.1 스캐너의 종류 및 특성 ·········································································· 4
2.1.2 스캔 방식의 종류 및 특성 ···································································· 6
2.1.3 포인트 클라우드 데이터 전처리 과정 ················································ 8
2.2 산업 현황 ········································································································· 12
2.3 건축 부재 인식 알고리즘 관련 연구 ························································· 17
2.3.1 RGB-D 데이터 기반 ·········································································· 17
2.3.2 3차원 포인트 클라우드 데이터 기반 ··············································· 20
2.4 소결 ··················································································································· 26
제3장. 복셀화 기반 건축 부재 인식 알고리즘 ··················································· 27
3.1 알고리즘 개요 ································································································· 27
3.2 벽면 부재 인식 ······························································································· 29
3.2.1 복셀화 및 아웃라이어 제거 ······························································· 29
3.2.2 천장 복셀 탐색 ····················································································· 30
3.2.3 벽면 복셀 탐색 ····················································································· 30
3.2.4 벽면 복셀 분류 ····················································································· 30
3.3 벽면 포인트 보간 및 개구부 탐색 ····························································· 32
3.3.1 객체 간섭 판별 및 포인트 보간 ······················································· 32
3.3.2 개구부 탐색 ··························································································· 33
제4장. 연구 결과 및 분석 ······················································································· 34
4.1 연구개요 ··········································································································· 34
4.1.1 데이터셋 ································································································· 34
4.1.2 정밀도/재현율 검사 ·············································································· 36
4.1.3 사례연구 데이터 ··················································································· 37
4.2 사례연구 결과 ································································································· 39
4.2.1 폐색영역 제거 ······················································································· 39
4.2.2 정밀도/재현율 검사 결과 ···································································· 40
4.3 결과 분석 ········································································································· 42
제5장. 결론 ················································································································· 44
참고 문헌 ····················································································································· 46

최근 본 자료

전체보기

댓글(0)

0