Fast and accurate detection of road to control the vehicle movement using Haar like features and Yolo classifier [articol]

Abstract

Autonomous driving is the most focused research problem in real world to ensure the transportation services to the people. In this research work, autonomous driving is focused by introducing the techniques which can ensure the uninterrupted vehicle movement in the road even with presence of obstructs. This is ensured by introducing the technique Fast and Accurate Road Detection System (FARDS). In the proposed research method, initially preprocessing is done on the traffic surveillance video to remove the noises by applying the Hybrid Median Filter method. And then background extraction is done by using markov random field process to extract the required objects. After extracting the required objects from the videos, Haar like Feature extraction is done to detect the objects present in the videos. Based on those feature road and objects detection is performed dynamically by applying Yolo classifier. This information is utilized to adjust the vehicle speed, acceleration and steering values to ensure the uninterrupted vehicle movements without any accidents. The overall evaluation of the research method is done in the matlab simulation environment from which it is proved that the proposed method FARDS leads to ensure the better outcome than the existing research techniques.