repository
생체측정 플랫폼 기반 실시간 차량용 운전자 모니터링 시스템
- Author(s)
- 박화범
- Advisor
- 김영길
- Department
- 일반대학원 의용공학과정
- Publisher
- The Graduate School, Ajou University
- Publication Year
- 2020-02
- Language
- eng
- Alternative Abstract
- This study aims to implement a biometric based real-time driver monitoring and emotion recognition platform for vehicle. In recent years, inadvertent traffic accidents in various driving situations accounted for about 41%[1-2], according to the 2016 Korea Road Traffic Authority statistics and according to the 2025 Roadmap of European New Car Assessment Program Euro NCAP[3], the Driver Monitoring System will be adopted as a safety assessment standard in 2020. In addition, in order to develop autonomous driving, driver intervention is required for certain situations while driving, and research on driver monitoring system (DMS) that can identify drivers beyond detecting simple driver carelessness in order to switch driving control, is ongoing[4]. In this paper, we implemented the Driver Monitoring System that determines and alerts the driver's attention and drowsy driving by analyzing the results of the extracted shape of a driver's face, eyes, nose, mouth, etc[5]. inputted through camera-based image algorithm, and also analyzes driver's emotional state through facial expressions of drivers and categorizes them into joy, sadness, anger, surprise and etc[6-12]. Detailed implementation items are as follows. First, camera-based Driver Monitoring System Hardware platform that satisfies the Automotive Spec is implemented. The DMS hardware platform consists of a DMS unit that handles face recognition, direction, eye opening, and emotion recognition software algorithms and IR Camera[13-15]. Second, Proposed and implemented camera-based driver careless and drowsy driving detection algorithm[16]. To detect the driver's face, Haar-Like and Hough Circle Transform algorithm is used to detect the driver's face, eyes, and mouth area, and then create a Mesh by connecting the detected eye and mouth areas, and detect the driver carelessness according to the change of driver's angle and determine driver’s drowsiness by analyzing the degree of eye closure[19-32]. Vehicle performance recognition rate and image-based frame detection recognition were detected through real vehicle test of driver's carelessness and drowsy driving. Third, a driver emotion recognition algorithm is proposed and implemented by detecting facial expression biometric information[34-37]. AAM (Active Appearance Model) algorithm detects facial expression changes in real time using Landmark of driver's face image and analyzes driver's emotion through 4 facial expressions of driver's joy, sadness, anger, surprise. and then propose service to change music genre of vehicle. Also, measured vehicle performance recognition rate through actual vehicle test[38-45]. Fourth, the vehicle biometric based real-time driver monitoring and emotion recognition platform implemented in this paper consisted of experiment environment by evaluator in actual vehicle situation and evaluated based on actual environment recognition rate. As a result, we could confirm the high performance that can be commercialized, but it is necessary to continuously update and research the image DB learning exposed to various environments and the software algorithm robust design accordingly.
- Fulltext
- Files in This Item:
- There are no files associated with this item.
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.
-
- License
- STATISTICS
- Total Visit :3,297,219
- Total Download :1,737
- Today View :5,036
