Method of Indoor Positioning Intel-ligence Using Image Analysis on Hallway Objects
영상내 복도 연관 객체의 인식을 통한 지능형 실내위치인식 방법
- 주제(키워드) 실내위치인식 , 컴퓨터비젼
- 발행기관 아주대학교
- 지도교수 노병희
- 발행년도 2021
- 학위수여년월 2021. 2
- 학위명 석사
- 학과 및 전공 일반대학원 AI융합네트워크학과
- 실제URI http://www.dcollection.net/handler/ajou/000000030724
- 본문언어 영어
- 저작권 아주대학교 논문은 저작권에 의해 보호받습니다.
초록/요약
Conventional indoor location positioning technology generally uses WLAN, GPS, and Bluetooth technologies. However, this technique may cause an error due to the strength of the received signal in an indoor environment. An indoor location positioning method that does not use a sensor is a method of referencing a 3D building model. This paper proposes a method of indoor location positioning using reference to a 3D building model, which is one of the indoor location recognition methods. The proposed method uses an image to recognize and use a permanent static object in a building, such as a door in the image, to predict and calculate the real-world distance ratio between the static object and the static object. After estimating real-world distance ratio, we compare estimated distance ratio with the real distance ratio to estimate the camera position. Through this method, it is expected that the indoor location can be recognized inexpensively with a single image without special infrastructure or sensors.
more목차
1. Introduction 1
2. Related Works 2
2.1. Indoor Positioning Technology 2
2.2. Indoor Positioning Using Image 3
3. System and Methodology 4
3.1. System Overview 4
3.2. Static Object Detection 5
3.3. 2D factor Detection 7
3.4. Real-World Distance Ratio Estimation 20
3.5. Position Estimation 21
4. Experiments 28
4.1. Static Object Detection Performance 28
4.2. 2D factor Detection Performance 30
4.3. Positioning Result 33
4.4. Time Performance 33
4.5. Limitation 34
5. Conclusion 35
Reference 36