Week-to-week construction scanning succeeds when change in the dataset represents change on the site. A shifted route, revised coordinate transform, different crop, or new modeling rule can create apparent movement that has nothing to do with installed work. The cure is a repeatable observation design with enough flexibility to survive real construction.
The team should treat the first scan as a baseline experiment. It establishes the control, route families, checkpoints, naming, processing recipe, comparison zones, and reporting template that later visits inherit. Every repeat then records what stayed constant and what had to change.
Choose comparisons people can act on
“Capture progress” is too broad for a field plan. The owner may want installed quantities, the superintendent may want access conflicts, the VDC lead may want deviations from coordination, and the scheduler may want evidence tied to work packages. Each use needs a different slice of geometry and a different cadence.
| Use | Stable comparison unit | Typical output |
|---|---|---|
| Installed work | Zone and trade package | Colorized status view plus quantity record |
| Clearance review | Named opening or equipment envelope | Section, dimensions, issue location |
| Surface change | Fixed grid or bounded surface | Difference map and exceptions |
| Visual record | Floor, room, or alignment station | Dated viewer and route coverage map |
Fix the comparison unit before the first visit. Changing from floor-based clipping to trade-based clipping halfway through the project makes trend reporting needlessly difficult. If stakeholders ask a new question later, add a versioned output instead of rewriting the historical definition.
Create the baseline kit
The baseline kit travels with every crew as a compact package. It contains the control diagram, checkpoint list, route map, zone boundaries, expected site conditions, capture settings, safety contacts, naming template, and a thumbnail showing the preferred start area. The file should be clear enough that a trained substitute operator can understand the intent.
Permanent reference features are precious on an evolving site. Survey control can provide the coordinate foundation, while visible structural geometry helps the operator repeat the trajectory. Select start and closure locations away from material laydown, temporary partitions, and daily equipment parking. Where a route will disappear behind completed work, plan the successor route before access closes.
Use route families
An identical set of footsteps is rarely possible. Route families preserve the important observation geometry without pretending the site is static. A core route might circle the floor perimeter and cross through two stable corridors. Short branches cover active workfaces. Optional recovery loops handle areas blocked on a particular day.
- Core: repeated every visit and tied to durable geometry.
- Workface: added where installation or earthwork changed.
- Detail: short pass for an opening, embed, rack, or tolerance question.
- Recovery: alternate path used when the preferred route is unsafe or blocked.
The log identifies which family was used and why. That simple note can explain a density change months later.
Hold the coordinate recipe steady
Every visit should use the same project coordinate reference, unit convention, control hierarchy, and transformation method unless a controlled revision is issued. Control used to position the data remains separate from checkpoints used to test it. If a control mark is disturbed, retire it visibly and document the replacement rather than silently substituting a new point under the old name.
Artec currently describes Jet as a SLAM-based system with GNSS-assisted georeferencing options and deployment by hand, backpack, vehicle, robot, or drone. A repeat program can use those options, but changing deployment modes affects trajectory, point distribution, stand-off, and occlusion. The comparison report should flag the change and preserve overlap with the established core route.
Run the visit like a short production
Before entering the route, the crew reviews the daily plan with site supervision. It confirms access, active hazards, trade movements, lighting or weather conditions relevant to complementary imagery, and control availability. Mobile LiDAR can collect geometry in darkness because the laser is active, yet the work area still requires whatever illumination and access controls the safety plan demands.
The operator records device and software versions, configuration, start time, route family, interruptions, and unusual motion in the scene. Swinging cables, moving workers, operating equipment, dust, rain, reflective surfaces, and newly feature-poor open areas can all affect interpretation. The live preview supports a field coverage decision, while independent checkpoints support the later accuracy statement.
The weekly field note should answer two questions: what changed on the site, and what changed in the observation?
Compare only after a repeatable QA gate
Each dataset passes the same pre-comparison gate. Processing reviews trajectory and closures, applies the documented coordinate workflow, exports at the standard resolution, and evaluates checkpoints. The team then checks that the same zone boundaries and masks are in force. A processing upgrade should be tested on prior data before it becomes the project default.
For surface or point-to-point change, set thresholds from the project requirement and the demonstrated uncertainty of the workflow. A threshold pulled from a product page can overstate what the entire process can resolve. Control survey, checkpoint uncertainty, SLAM behavior, surface character, registration, sampling, and comparison method all contribute.
Review spatial patterns
A single root-mean-square value cannot reveal every failure. Plot residuals by checkpoint and compare them across dates. If errors grow with distance from the start, rotate across a floor, or cluster near one transition, the pattern deserves investigation. Also review cross-sections through unchanged structural features. They provide an intuitive stability check for project teams.
Publish a concise weekly record
A good progress package is small enough to use in a meeting. It names the observation date, scan zones, route deviations, control status, checkpoint summary, new exclusions, and processing version. It then presents the actual changes: installed work, movement, quantity, clearance, or documented condition. Link every view to a stable location identifier such as grid, level, room, alignment station, or work-package code.
Archive raw observations, processed master data, comparison inputs, outputs, and the weekly note under versioned names. Never overwrite the prior accepted comparison. When a correction is required, issue a replacement with a reason and preserve the superseded record.
The repeatability scorecard
| Question | Green condition | Action if different |
|---|---|---|
| Same reference? | Control and transform unchanged | Document revision and test overlap |
| Same observation? | Core route and useful stand-off repeated | Flag route family and occlusion impact |
| Same processing? | Approved version and parameters used | Reprocess baseline or isolate change |
| Same comparison? | Zones, masks, thresholds, and outputs fixed | Version the reporting definition |
Comparable scans come from disciplined repetition, not rigid choreography. The site will evolve; the observation record makes that evolution intelligible. Once the system is running, a weekly scan becomes one of the most satisfying records on a project: a geometric timeline that can be checked instead of remembered.