Oblique photography review: a GenAnywhere checklist for teams
How to run a review that survives questions about dates, coverage and the difference between a photo and a measurement.
When a team reviews oblique photography, the work goes better if it separates three things: what the images show, what the model was built from, and what the decision actually needs. Most disputes in these meetings come from mixing those three. A short checklist keeps them apart and makes the scene easier to defend.

1 Prepare the pack
1.1 What should the pack contain?
The capture itself, with its flight date in the file name. The processing report, even a short one, naming the software and the settings that produced the mesh.
The area drawing, so the room can see what was flown and what was not. A boundary that stops at the site edge explains a blank neighbour before anyone asks.
The questions the meeting needs to answer. A review without a stated question drifts into admiration of the texture, and texture is not a decision. Put the question on the opening slide, where it can discipline everything that follows.
1.2 Who reads which part?
The designer reads the surfaces: conditions, materials, the parts that will be altered. The planner reads the arrangement: footprints, heights, the relationship between blocks.
The engineer reads the geometry and wants to know how it was controlled. The client reads the view and wants to recognise the place.
Send the pack before the meeting and label each part for its reader. A scene that arrives without a note gets read by everyone in the same way, which suits nobody. Ask each reader to bring one question; a pack that answers four questions is doing its job. The OGC CityGML standard is a useful reference when a team needs to agree what level of detail a model is meant to carry.
2 Run the review
2.1 What do you show at the start?
Show the extent before the detail. A plan view with the coverage boundary marked tells the room what is evidence and what is context.
Then show the known failures. Water, glass, vegetation and moving objects should be pointed out by the presenter rather than discovered by the client.
Then show the question. A review that opens with the decision to be made keeps the texture in its proper place as supporting material rather than as the subject. If the room starts with the pictures, let it, then bring the decision back before the discussion hardens.
2.2 How do you answer a challenge?
Answer with the source, not with the picture. If someone asks whether a wall is load-bearing, the capture shows the wall and the drawing or the surveyor answers the question.
Separate what the scene shows from what it implies. "The façade is rendered and intact on the flight date" is a statement about the record; "the building is sound" is an inference that needs another source.
Where the answer is unknown, say so and note what would settle it. A review that records its open questions is more useful than one that closes them with guesses. An open question with an owner is a plan; an open question without one is a worry.

3 Record the outcome
3.1 What goes in the minutes?
The date of the capture, the version of the scene and the software used to view it. A screenshot without that context becomes an orphan document within a month.
The decisions taken and the evidence behind each. If the team accepts a massing because the capture shows the neighbouring roof line, write that down.
The open items, with an owner. A question about a boundary or a surface is easier to chase when someone's name sits beside it. Give each item a date by which it will be answered, even if that date is provisional.
3.2 What gets reissued?
Reissue the scene when the capture is replaced or when a known failure is corrected. A new flight should carry a new date, not a quiet overwrite of the old file.
Reissue the note when the decision changes. A pack that still argues for a scheme the team abandoned will be quoted back at the wrong moment.
Reissue the drawing when the coverage changes. People plan site visits around the extent they were given. Keep a one-line change log at the top of the note so a reader can see what moved since the last issue.
4 Place the map base
4.1 What does GenAnywhere add?
GenAnywhere builds a site model from a drawn map area and presents buildings, terrain, water and vegetation on separate layers. A team can replace a building, move a block or change a height without editing a mesh.
In a review, that makes the base the place to test alternatives. The capture stays as the record of what exists; the base carries the proposal that sits beside it.
GenAnywhere draws its model from a map area and keeps buildings, terrain, water and vegetation on separate layers. A review can hide the trees, replace a block or change a height in the scene without reworking the capture. The GLB export then carries the arrangement into a design or GIS tool, where the proposal is drawn properly. The area can also be redrawn if the review expands, which keeps the pack current without another flight.
The export is a general GLB for 3D, BIM, GIS and engine workflows. It is a working scene rather than a survey deliverable, and the note should say so. That sentence prevents most of the misunderstandings that follow a base into a technical meeting.
4.2 Where does the base stop?
It stops at surface truth. It shows a plausible mass rather than the state of a façade, and it should not be used to argue about condition.
It stops at currency. Map data lags, and a base built last quarter may not show last month's demolition.
It stops at the boundary of what was drawn. A team that needs the neighbour's yard has to include it in the area, or accept that the review will have a hole at the edge. A hole is acceptable when the note names it and misleading when the scene pretends it is a wall.

5 Avoid the usual failures
5.1 What do teams miss?
The date. A capture shown without a date invites the room to assume it is current.
The coverage. A scene that ends at the site boundary makes the surroundings look empty, and empty surroundings flatter a scheme.
The proposal. A new volume added to a captured scene can read as existing unless it is styled and labelled as proposed. Style it in a flat colour and give it a name that begins with the word proposal.
5.2 How do you prevent a rebuild?
Agree the coordinate frame and the unit before anything is merged. Two sources that look aligned and sit a metre apart will cost more to fix than to check.
Keep the capture and the model as separate files. A merged scene hides provenance, and provenance is what a review is for.
Test one known dimension on import. The receiving tool may apply its own units, and a quick check saves a rebuild later. Run the test on the machine that will present, not on the one that exported.
6 Set the bar
6.1 What makes a review defensible?
A stated question, a dated source, a coverage drawing and a note that names who checked what.
A clear separation between what was observed and what was inferred. The room can disagree with an inference; it cannot argue with a record.
A record of the limits. A review that says where the capture fails is harder to unsettle than one that presents a smooth surface as the whole story. Defensibility is mostly a matter of having written down what was known at the time.
6.2 When is the checklist too heavy?
When the decision is small and reversible. A quick look at a single façade does not need a pack, a note and a minute.
When the same team reviews the same site weekly. Keep the standing note and update the date rather than rebuilding the process.
When the audience is internal and already knows the limits. The checklist is for handovers and formal reviews, not for every conversation.
6.3 When do you hand over?
Hand over when the decision is made and the next team needs the record rather than the discussion.
Hand over with the capture, the base, the note and the open items. A scene without its note is a picture, and a picture invites assumptions.
Hand over before the files become slow or the software changes. A format that opens today may need conversion tomorrow, and the note should record what was tested. Include the version numbers of the software used to view the scene, as well as the one used to make it.
Draw the area in the GenAnywhere app and keep the review pack beside the capture it describes.
7 FAQ
FAQ-1 Does GenAnywhere review the capture for me?
No. The platform does not inspect photogrammetry or judge its quality. It builds an editable site model from map data, and the review of a capture remains a human task supported by the note and the coverage drawing. The scene supplies the arrangement; the judgement about the capture still belongs to the team that flew it.
FAQ-2 How many images should a review pack include?
Enough to show the extent, the known failures and the features the decision depends on. A handful of labelled views usually beats a long sequence, because each view in a pack should answer a question the room has already asked. Drop any view that answers nothing and keep the ones a reviewer will want to revisit.
FAQ-3 Can the model be edited after the review?
Yes. Buildings, terrain and vegetation sit on separate layers, so a replacement or a height change can be made without touching the capture. Keep both files so the record of existing conditions stays intact. Record the edits in the note, so a later reader can see which volumes changed after the area was drawn.
FAQ-4 What if the client brings their own model?
Compare the two on one known dimension and agree which governs the decision. Then label each source in the scene. A review with two unlabelled models invites the room to treat both as equally reliable, which is rarely true.
See how a map area becomes an editable site model in the GenAnywhere workflow.
Content and technical notes
This article was prepared by the GenAnywhere team with AI-assisted drafting and human review.
Product features, formats, and availability change as the platform evolves; refer to the official GenAnywhere website for current details.
For technical questions or corrections, contact support@genanywhere.com.
Last reviewed: 2026-11-06.