REAL-TIME LANE DETECTION USING RASPBERRY PI FOR AN AUTONOMOUS VEHICLE

Abstract

The latest developments include the smart world, smart cars, and other technologies. The development of intelligent vehicles must be able to detect and identify traffic signs to ensure traffic safety. To control the speed of an autonomous vehicle, environmental perception is essential. The traffic regulations listed on traffic signs must be fed as input to autonomous vehicles. However, traffic regulation is one of the essential factors in autonomous vehicles, but many more factors need to be taken into consideration. In this paper, machine learning techniques for stop sign detection, traffic sign detection, and object detection with Obstacle avoidance and distance calculation play a crucial role in regulating the longitudinal velocity of an autonomous vehicle. The stop sign disappears from the camera’s field of view as the car approaches it, making it challenging to stop the car at the desired distance from the sign. To know exactly where to stop the vehicle, knowledge of the location of the stop line is crucial. Obstacle avoidance and detection of the object are other challenging factors that analyse the potential. The Haar cascade classifier method is the optimizer approach utilised here. Features of both the Hue Saturation Value grey scaling space have faster detection capability in speed and low lighting suffering. The proposed technique is evaluated using an Indian Traffic Sign that has set a benchmark. The proposed method provides an accuracy of almost 80%.