Fire Investigation Society of Korea

화재의 환경적 물리적, 화학적, 구조적 배경에 관한 과학적 감식과 조사체계의 구체화 및 인적, 물적 손실의
예방과 안전화의 학문과 기술발전을 도모하고, 산,학,연 정의 상호교류를 통한 화재조사 및 감식의 전문화와
함께 소방관련 정책방향 발전에 공헌하며, 소방 분야 종사자들 간의 정보교환의 장과 사기양양 및 친목도모를
그 목적으로 한다.

학회지검색


pISSN: 2092-531X

한국화재감식학회 학회지 (2018)
pp.19~32

드론을 활용한 감식 데이터의 딥러닝에 관한 연구 : 재난현장 및 실종자 수색을 중심으로

차정훈

(영동군청)

유재석

(충북소방본부)

박창우

(우송대학교 드론아카데미)

In the National Police Agency statistics According to In 2017 alone, children, the disabled and the elderly with dementia are reported missing. There are an average of 100 missing cases per day. In addition, according to data compiled by the National Defense Agency, there are 40,000 to 50,000 fires annually in Korea, Massive loss of life and property will result. Therefore, research on various algorithms for these disaster sites and search methods for missing persons continues, and drone technologies with automated search and analysis features are being developed, particularly using artificial intelligence. As a method of collecting data for searching for missing persons and identifying disaster sites, data such as 360 degree panoramic photos and images can be obtained by attaching aerial video and photography, 3D mapping, and special equipment to drones. Data acquisition aims to extract data analysis elements so that deep learning pattern recognition algorithms can be applied to the latest technology of artificial intelligence. Deep Learning's extraction of elements to search for various missing persons and detect disaster sites is the most important element in unmanned mobile device control technology including AI (Artificial Intelligence) drones. A number of causes of human casualties at the disaster site and the search for missing persons are due to the delay in searching for missing persons and the lack of human error during the inspection and diagnosis of the disaster site.

Download PDF list