Validation of an AI enabled application for automated road surface marking inspection and auditing

验证用于自动路面标记检查和审核的人工智能应用程序

• 批准号: 10074687
• 金额: $ 6.33万
• 依托单位: ASSETSCAPE LIMITED
• 依托单位国家: 英国
• 项目类别: Grant for R&D
• 财政年份: 2023
• 资助国家: 英国
• 起止时间: 2023 至 无数据
• 项目状态: 已结题
• 来源: https://gtr.ukri.org/projects?ref=10074687
• 关键词: Validation; AI; enabled; application; automated

The applicant, a company called Assetscape, has been contacted by one of their clients, WSP, who is an agent of the UK National Highways Agency. National Highways and their various framework operators monitor and maintain England's strategic road network. One way that they do this is by taking very large quantities of pictures of the road surface using specially modified vehicles that traverse the network over a two year cycle. This very large number of images is then examined and audited by human inspectors to see if areas of the road need repair or preventative maintenance. National Highways has tried to automate this process but has failed to find software systems that can reliably identify worn or damaged sections of the road better than human inspectors.?WSP, on behalf of National Highways, has asked Assetscape if it would be able to develop an Artificial Intelligence solution that is capable of evaluating and auditing the large body of road surface image data to reliably identify road markings (e.g. white lining) that are in need of repair, replacement or removal. This would save a great deal of money, time and provide better maintenance of the roads.Assetscape is applying for grant funding from the Innovate UK Transformative Technologies competition to experiment with the latest generation of AI to enable the team to demonstrate reliable road surface marking evaluation and condition reporting along with data-output auditing functionality. Once this research work is proven, the business will be in a position to invest in the development of a full scale commercial application for UK customers (e.g. National Highways), other highway authorities and road agencies overseas.The project and funding will enable Assetscape to access a large amount of road image data, experiment and develop the AI technology to process these images automatically. Once developed, the team will then test the performance of the AI-based system by comparing the performance of the software with expert human road surface auditors.

申请人是一家名为Assetscape的公司，他们的一个客户WSP联系了他们，WSP是英国国家公路局的代理商。国家公路和他们的各种框架运营商监控和维护英国的战略道路网络。其中一种方法是用经过特殊改装的车辆拍摄大量的路面照片，这些车辆以两年为一个周期在路网上穿行。这些大量的图像随后由人工检查人员进行检查和审计，以确定道路区域是否需要维修或预防性维护。美国国家公路公司（National Highways）曾试图将这一过程自动化，但未能找到能够比人工检查员更可靠地识别磨损或损坏路段的软件系统。科进代表国家高速公路向Assetscape咨询，是否能够开发一种人工智能解决方案，能够评估和审计大量路面图像数据，以可靠地识别需要维修、更换或拆除的道路标记（如白色衬里）。这将节省大量的金钱和时间，并提供更好的道路维护。Assetscape正在申请Innovate UK变革技术竞赛的拨款，以试验最新一代的人工智能，使团队能够展示可靠的路面标记评估和状况报告以及数据输出审计功能。一旦这项研究工作得到证实，该企业将能够为英国客户（例如国家公路）、其他公路当局和海外道路机构投资开发全面的商业应用。该项目和资金将使Assetscape能够访问大量的道路图像数据，实验和开发人工智能技术来自动处理这些图像。一旦开发完成，该团队将通过将软件的性能与专家路面审计员的性能进行比较，来测试基于人工智能的系统的性能。