Autonomous Vehicle Driver Assistance System for Road Signs and Lane Detection Based on Deep Neural Networks Framework

The aim is to recognise the different traffic sign boards and identify lanes along with neighbourhood vehicles. Environmental perception research is critical to the realisation of automated driving, and lane detection algorithms are at the forefront of this field. To run autonomous vehicles, humans need to be able to multitask and finish multiple jobs fast. Advanced driving assistance systems and driverless cars both depend on lane detection and vehicle detection as a key component. Traditional methods only identify the lane from a single frame, and they offer performance with difficulties when handling extreme scenarios like lane line erosion, big shadows, significant vehicle occlusion, noisy image inputs, etc. In practical terms, lanes are intended to be continuous line constructions on the road. As a result, information from earlier frames can be used to extrapolate a lane that is not clearly visible in the current frame. The deep neural network architecture is used to analyse highways to provide driving assistance for vehicles. Two more subtasks are included in road analysis. The deep neural network framework is introduced with the goal of using the least amount of hardware and software setup possible to help this driving function, if not completely automate it. The proposed framework's ability to be used in real-time due to its low complexity. The accuracy of the pipeline's classification of road signs, its capacity to collect data about the road, and its time bench-marking performance are the first three metrics used to quantify the pipeline's performance.