ArchAI: AI in archaeology to de-risk construction and land development.

点击翻译按钮获取中文摘要

Dr Iris Kramer is the founder of ArchAI, a startup based on the world-first technology she developed during her PhD. ArchAI, is de-risking the construction industry by using AI to automatically detect archaeology using earth observation data. Knowing where archaeological sites are located at the earliest planning stages allows accurate estimates of time and cost involved with acquiring planning permission. This means that ArchAI will lower the cost of construction and ensures that vital historical sites are preserved.Construction projects have a big impact on their environment, including archaeology, and it's therefore a legal requirement to assess the potential damages before planning permission is granted. Increasing development and changes to the planning process will likely reduce the accuracy of archaeology assessments and local authorities now fear unrecorded destruction of archaeology and in turn developers fear that archaeology will be found during a construction project.Archaeology is revealed on earth observation data through (1) LiDAR in forested/rough terrain to detect earthwork remains, and (2) satellite imagery in agricultural fields to detect crop stress revealing sub-soil walls and ditches. ArchAI has already developed and tested the LiDAR product across the UK and detected hundreds of previously unknown sites. Our customers include the Forestry Commission and the National Trust.Through initial funding we were able to create a national training database of known archaeological sites and generate national maps with detections of various archaeological site types. The resulting product is highly detailed but requires experts to derive risks proposed by the archaeology. Although this product satisfies some customer needs we have identified that further product development is required to truly address the problem of early stage risk (from archaeology) in the construction industry. In the WIA project we will be addressing this by creating national risk maps which will be based on our national maps but also include risk analysis through (1) predictive modelling based on cultural indicators (e.g. prehistoric settlements are often found near rivers and defended settlements are often found on higher ground) and (2) condition modelling (e.g. archaeological sites in waterlogged condition are more expensive and time consuming to excavate). The resulting map can be interpreted as a traffic light system indicating risk levels and will inform the earliest stages of construction planning such as infrastructure optioneering and scheme design (where archaeology is currently not considered because of the cost).