Efficient Computer Vision ADAS Hardware for Connected and Autonomous Vechicles

适用于互联和自主车辆的高效计算机视觉 ADAS 硬件

• 批准号: 132269
• 金额: $ 18.75万
• 依托单位: MYRTLE SOFTWARE LIMITED
• 依托单位国家: 英国
• 项目类别: Feasibility Studies
• 财政年份: 2016
• 资助国家: 英国
• 起止时间: 2016 至 无数据
• 项目状态: 已结题
• 来源: https://gtr.ukri.org/projects?ref=132269
• 关键词: Efficient; Computer; Vision; ADAS; Hardware

Bringing the next generation of Advanced Driver Assistance Systems (ADAS) hardware to automobiles iscomplex, expensive, iterative and slow. Development and rollout in the marketplace is further slowed by thehigh standards naturally required by the car industry. A major consequence of this situation is that advancedcomputer vision algorithms, which are used in other industries for human safety, are not appearing as quicklyas they should within the increasingly connected cars on the roads today. This project is to explore thefeasibility of developing a new technology in realtime image processing to drastically reduce the iteration timesof producing ADAS hardware. The project will produce hardware versions of key algorithms using our softwareand evaluate the efficiency of our new process. If successful this project would see the UK well-placed to be atthe forefront of owning the IP within all the chips in future car models and leading the way in making our roadssafer.

将下一代高级驾驶辅助系统（ADAS）硬件引入汽车是复杂、昂贵、迭代和缓慢的。由于汽车工业自然要求的高标准，市场上的开发和推出进一步放缓。这种情况的一个主要后果是，先进的计算机视觉算法，这是用于其他行业的人类安全，并没有出现，因为他们应该在越来越多的连接汽车在今天的道路上。本计画旨在探讨开发一种即时影像处理新技术的可行性，以大幅减少生产ADAS硬体的重复次数。该项目将使用我们的软件生产关键算法的硬件版本，并评估我们新工艺的效率。如果成功，这个项目将使英国处于领先地位，在未来的汽车模型中拥有所有芯片的知识产权，并在使我们的道路更安全方面处于领先地位。