An autonomous vehicle needs to be familiar with its surroundings. The safety of thetransportation system is greatly enhanced by advanced driving assistance systems (ADASs). Road detectionis one of the steps that a driving car must do. Is it possible for a computer to recognize a road in a singlephotograph for this purpose? This question is addressed using the lane detecting techniques. Roads andlanes are tough for machine learning to differentiate because of training a machine to recognize a road.Over the past few decades, a number of lane identification technologies have been created and integratedinto various autonomous cars. It is still very difficult to create lane recognition technology that caneffectively identify a road lane in a range of road conditions. This research provides a composite approachfor road detection from image processing using convolutional neural networks by testing 150 photographsthat include a road, jungle, muddy road, and barriers. It will decide if an image contains a road or not. Inthis essay, we first establish whether a road exists. The second step is to find a lane on the finished road.The benefit of the proposed technology is that if there is a road, the automobile can continue to moveforward; otherwise, it will stop.