An efficient algorithm for extracting structured information from different vision objects in autonomous vehicle 2019-28-0074
This is the age of technology and innovation. To facilitate and make human life easier, the technologist and scientist move for the innovation of automotive car that can travel towards its given destination without any input from the user. This paper focus on the accurate path tracking performance of a vehicle, and the line detection of an image using the Hough transform algorithm which finds the hidden straight line in larger amount of data. The captured image from the camera is processed through the proposed edge detector algorithm with the least square method and Kalman filter to get an accuracy of nearby edges. The advantage of the proposed algorithm is the reliability of data which paves the way for accurate prediction of the state estimation of vehicle. The simulation test was done with the existing canny edge detection algorithm and the proposed algorithm, which shows the 20% increase in the accuracy of the image using the proposed method. The proposed algorithm has been validated with minimal distance of 1 km in off real time traffic condition with minimum obstacles.
Key words: Hough transform algorithm, least square method, Kalman filter