AI object detection tool
In a sector known for conventional working methods, digital innovation is a challenge. In Q42, BAM found a partner that provides smart AI solutions to work more efficiently and improve existing workflows. Result: the project duration can be shortened by three weeks.
Royal BAM Group designs, builds and maintains buildings and infrastructure for both government and private clients. The renovation of existing buildings presents many challenges. This is often because accurate construction drawings are lacking. Although digital models are made, they are not always available. In addition, recreating a construction drawing from a digital model is a labor-intensive process that can take up to a week. The lack of reliable information creates potential errors in various project phases, including budgeting, planning, scheduling and construction.
Improvement thanks to AI
To speed up this process and make it less error-prone, we have devised a solution in the form of an internal web app together with Renovation Concepts from BAM Wonen. The construction group makes detailed 3D scans with drones for every renovation project, such as making housing blocks more sustainable. Thanks to AI, we have been able to automate and accelerate the conversion of these scans into accurate measurements. By having access to precise information earlier, BAM can more easily plan and carry out work.
How it started
To validate whether our idea would be feasible, we started with a so-called jumpstart: in two weeks we built a prototype and tested it with end users. In the case of BAM, we investigated whether we can automatically extract accurate dimensions from 3D scans. That yielded a positive result, on which we really started building the object detection tool.
A tailor-made solution
To support various stakeholders, we have built internal tooling for exact dimensions, which is in line with the highly regulated working methods of the construction industry. The 3D scans from the drones serve as a basis for generating dimensions of, for example, window frames with machine learning and other code magic. Based on this information, BAM can draw up a budget for the individual activities much more easily and accurately.
Faster turnaround time as a result
Because both clients (BAM) and executors (e.g. painters) in a renovation project now use the same accurate information (instead of different, own measurement data), there is more likely to be agreement about the scope of work. Our new AI tooling leads to a more efficient way of working and an improvement in coordination, planning and construction. This can reduce the lead time of construction projects by three weeks.
Q42 Podcast about AI with engineer Leonard Punt and BAM product owner Christian Alferink
