The automation industry is rapidly growing and coming up with new and
improved techniques for reducing time and efforts. One such example is the
autonomous cars which are said to be the future of the automobile industry since they
would be driver less, very efficient and relieve the stress of daily commuting [1].
Advances in technology using the AI and deep learning techniques help in improving
the safety of the passengers and also in minimizing the efforts of the driver. For the
study of autonomous vehicles, a lot of data needs to be collected, some of which
include warning signals, speed limits, obstacles, collision avoidance, etc. This paper
shows how IoT devices i.e. cameras and LiDAR sensors help in data collection, how
deep learning is a solution, and how image recognition methods that use deep learning
can help in object or any obstacle detection. An image processing algorithm based on
deep learning is proposed in which the image perception can be made by an optical
camera communication technique that can be used for collecting the data. Hence it will
highlight how deep learning is used in the field of image processing or image
recognition.
Keywords: Autonomous driving, Camera, CCD, LiDAR, Convolution neural network(CNN), Deep learning, Image processing, Lane detection, LED area detection, Machine learning, Object detection, Vehicles, YOLO.
