Mach9’s web-based AI CAD application now processes aerial LiDAR and photogrammetry-derived datasets, extending its survey workflow beyond mobile and terrestrial scans. The change brings feature extraction, point-cloud classification and surface modeling into the same application; teams can review and refine results before exporting drawings and surfaces. Mach9’s announcement describes support for data captured by drones, helicopters and aircraft.
Users can upload point clouds and orthomosaic imagery. The software is designed to draft features including curbs, building footprints, pavement edges and vegetation outlines as 3D points and lines. For utility work, Mach9 also describes extracting poles and 3D catenary geometry from suspended wires, which can feed clearance analysis, make-ready workflows and specialist design and pole-loading applications. Surveyors can edit point classifications and linework, then create ground surfaces, breaklines and contours. A deviation heatmap is intended to help identify areas of a surface that may need attention.
The workflow’s value is not simply that AI generates an initial draft: survey teams still need to assess and correct it. Mach9 says its AI can produce an initial draft in hours, and claims centimeter-level accuracy at infrastructure scale, but its announcement provides no test methodology or independent accuracy results. I would not treat that claim as project acceptance: I would accept a surface for export only if elevations at independent surveyed control points meet the project’s stated tolerance and the deviation heatmap shows no unexplained areas. Otherwise, the team should correct the classifications or surface and review it again.
The export handoff is another risk. Mach9 says users can check surfaces against the point cloud and extracted features before LandXML export; Langan’s Russ Hall also reports that the surface can be checked in Mach9 and exported to draw correctly in Civil 3D or OpenRoads. That is useful evidence of a workflow, not a guarantee for every project or software version. I would require the receiving team to inspect the imported breaklines and validate the LandXML surface’s geometry and elevations against project requirements before using it downstream. The announcement does not specify Civil 3D or OpenRoads version requirements.
The listed export formats—DXF, DWG, SHP, GDB and LandXML—offer routes into downstream CAD and GIS processes. The announcement cites use with Autodesk Civil 3D and OpenRoads, but does not detail every supported input configuration. It also lists supported capture systems from RIEGL, Leica, Trimble, GeoCue, Phoenix LIDAR, YellowScan, Wingtra, Teledyne Geospatial and DJI. For teams whose current process separates extraction from surface creation, consolidating those steps in one web-based application could reduce handoffs; that is an inference, not a reported result. The supplied announcement does not identify required browser, hosting or receiving-software versions, so teams should confirm those requirements before adopting the workflow.
Mach9 attributes a roughly 40% delivery-time reduction on multi-thousand-acre Texas projects to customer Langan; Welch Comer reports average savings of four to five hours per project. These are customer-reported results in the vendor’s announcement, not independent benchmarks, and should not be assumed to transfer unchanged to other sites or capture conditions.
The expansion is relevant to survey and civil teams already handling aerial data who want to assess whether classification, drafting and surface review can be consolidated. My decision rule is to withhold delivery if the surface fails the project’s tolerance checks or the receiving software does not preserve the required geometry.
Sources
- Mach9 now works in the sky — Mach9
- Mach9 adds aerial LiDAR and photogrammetry support — AEC Magazine
