Lane Detection using Video Processing for Robot Cars

Road accidents are a significant cause of preventable fatalities, with a large number of accidents being attributed to human errors, such as distraction or intoxication. However, computers have the potential to significantly reduce these accidents, as they do not suffer from attention issues while driving. Autonomous systems can control the handling of vehicles effectively, detect and respond to hazardous situations on the road, and thus avoid accidents. Given the increasing need for improved safety, accessibility, and welfare, there is an urgent demand to explore and advance autonomous vehicles. This paper aims to present a Raspberry Pi-based autonomous vehicle prototype that can recognize lanes and plan its movements accordingly. The prototype applies Thresholding smoothed by Canny Edge Detection to detect lane lines, calculate the difference between the frame and lane center, and modify the motor's PWM through an Arduino UNO to change the vehicle's direction. Comprehensive testing has been conducted to verify the prototype's reliability under various lighting and environmental conditions.