Autonomous Driving Technology through Image Classfication and Object Recognition based on CNN.

The time has come for humans that the cars themselves think and judge rather than driving by humans, and has already reached a certain level. It hasn't reached the commercialization stage yet, but it's not a story of the future anymore, and even now, a great deal of technological development is taking place every day. However, since this autonomous driving technology does not have a driver when an accident occurs, it can cause problems in terms of responsibility and ethics, so it has to be possible to make the best choice at anytime and anywhere without accident. It requires a high level of technology on the software and the hardware. Most autonomous vehicles currently on the market do not stay or exceed the level of 2 to 2.5. The level of 2~2.5 is a partial automation step, and in a stable environment, autonomous driving is possible in part, but it must be possible for the driver to immediately take over control of the driving of the vehicle. Using CNN-based YOLOv3 implements object recognition and image classification, which are the core technologies of autonomous driving technology and Based on the data, it is implemented through the AI autonomous vehicle kit, so that it is possible to make suggestion or ideas for development of Autonomous driving technology.