AI in construction: a site that watches itself
An execution error becomes expensive when it is seen late. Miras Vision turns site cameras, daily photos and execution drawings into an always-on inspector that compares every stage with what it should be.
From excavation to finishing
Visual site monitoring
Cameras and daily photos are analysed by machine vision: progress per zone, crews present, equipment and safety — with nobody sitting behind a monitor.
Execution control against drawings and BIM
What was built is compared with the execution drawings and the building information model; deviations in rebar, services or levels are flagged before the pour and before they are covered.
Safety and site discipline
PPE detection, entry into hazard zones, work at height; instant alerts to the site manager.
Automatic progress reporting
Physical progress is extracted from images and daily reports and placed next to the schedule and cost in the Miras PM portal.
Miras PM portal →Miras Vision: 83% fewer execution errors
Costly execution errors on construction projects.
Miras machine vision; automatic scan and benchmark of every stage.
83% fewer execution errors · foundation to finishing.
Miras built this system first for its own projects — the group’s villa communities and buildings — and then rolled it out to other projects. Every stage is scanned and benchmarked against the reference; the error is seen while it is still cheap.
A builder that builds AI
Miras itself builds and designs, so our systems are made for the real site: weak internet, the site manager’s Android phone, the supervising engineer’s iPhone photos, and reports that must reach the project manager the same night.
- Works on Iranian mobile data without a VPN.
- Images and data stay in Iran, on site or on Miras servers in Tehran.
- Integrated with the Miras PM portal: image → progress → schedule → cost.
- For client, contractor and consultant, each with their own access level.
Miras group projects
Miras villa communities in Mazandaran, the projects of the Maryam Amjadi architecture office and the Miras Amlak real-estate portal — where every product is tested before release.
Frequently asked questions
What exactly does AI do on a construction site?
From images and data it complements the work of the inspector and project manager: it measures physical progress, shows deviations from drawings, monitors safety and automates the daily report. The decision remains with the engineer; AI shows early what used to be seen late.
What equipment is needed for visual monitoring?
Ordinary site cameras (IP/NVR) are enough; where there are none, the supervising team’s daily phone photos are also an input. The processing server can be on site or on Miras infrastructure.
Does it work with our BIM model?
Yes. The building information model (IFC/Revit) is the comparison reference; without BIM, 2D execution drawings are the baseline.
Is it worth it for small projects?
Any project where one mistake in rebar, services or levels is expensive benefits from visual monitoring. The system scales with the number of cameras and project stages.
Bring intelligence to your plant or project
We start with a data audit: where your data is, where cost leaks, and which model pays back first. Tehran · Isfahan · Noshahr · Dubai.