车载驾驶辅助系统及其深度学习与视觉技术

Advanced Driver Assistance System(ADAS) and Its Deep Learning & Embedded Vision Technology

本文简要介绍自动驾驶技术的基础,包括ADAS的原理、架构和功能。深度学习嵌入式视觉技术在实现车载驾驶辅助系统(ADAS)的功能架构中起着关键作用。从深度神经网络机器学习的原理和特点出发,分析了驾驶辅助系统中深度学习和嵌入式视觉技术的原理、功能与架构,简述ADAS中视觉技术的块拼贴、场景分割等技术手段的相关特性,以及简述卷积神经网络CNN在ADAS中应用的现状与发展展望。

The paper introduces the basics of automatic driving technology-the principles,architecture and functionof ADAS.Deep learning and embedded vision technology play a key role in realizing in implementing advanceddriving assistant system(ADAS).From the depth of principles and characteristics of the neural network of machinelearning,this paper expounds the driving assistant system in depth study and the architecture of embedded visiontechnology,ADAS patch-based scene segmentation techniques such as related feature.The present situation and developmentprospect of the application of convolution neural network CNN are also introduced.