Distance Estimation & Driver alert system using ZED Stereo Camera and Raspberry Pi using Socket Communication

In automotive system design, the estimation of the distance is a crucial task, with respect to the autonomous vehicle perception module. The range sensors like LiDAR, RADAR, Ultrasonic sensors, etc. have widely been used for range estimation applications such as adaptive cruise control, lane departure and warning, blind spot estimation etc. The vision- based sensors such as stereo camera, monocular cameras, etc. can also be used for effective estimation of the obstacle distances. As the camera data/frames contains more spatial information compared to a RADAR or ultrasonic sensor, the system could possess a higher understanding of the surroundings. This paper contains an innovative method for distance estimation based semantic information of an obstacle using ZED stereo vision camera. The proposed method is validated using ZED stereo vision camera with Raspberry Pi controller through socket programming.