Where objects without position come from
Building materials plant registers contain objects that exist in the documentation but have no position on the map: old feeders, tanks, chambers, equipment taken out of service. Their names live on in permits and reports, so they cannot simply be deleted.
In addition, the coordinates of some locations are stored in the system of another view. A view without georeferencing cannot be a source of position at the plant: its numbers, without an anchor, rotation and scale, do not say where anything stands.
Recommendation 1: a shared Plant site area without geometry
Register objects without a known position go to the area Plant site, which has no geometry or point. They are invisible on the map until a position is assigned. This is a deliberate decision: the object is in the register and can be selected in the form, but the map does not pretend to know where it stands.
The alternative, placing a pin in the middle of the hall, gives a false sense of knowledge. The person accepting a permit would see a place that nobody has checked.
Example: number of objects without position by quarter
Recommendation 2: do not delete the view with calibration
A hall used as a view may be 1000 x 500 m, with georeferencing (anchor, rotation, scale) and an orthophoto of the plot, and the buildings are areas inside the view. During a test, moving the buildings into separate views deleted the only view with calibration. The editor showed a message about missing georeferencing and the orthophoto disappeared.
Recommendation: before every structure change, make a list of views with calibration and block their deletion. Calibration is the only source for converting local metres to geographic coordinates, so losing it invalidates the whole map.
Recommendation 3: one source of truth for conversions
Converting local metres to geographic coordinates must have one source of truth, and the migration script must be a 1:1 port of this function, with the grid convergence correction for the PUWG 1992 system. Two implementations always end up diverging.
Verification of the transformation on 64 control points gave a maximum error of 0.01 m after converting the outlines back. Every plant should run such a test: convert the points forward and back and measure the largest deviation.
Recommendations 4-6: closing gaps over time
Recommendation 4: assign an owner and a deadline to each gap. A list of objects without position with no owner does not shrink. Recommendation 5: record the position during the first job at the object, when someone is physically there anyway. Recommendation 6: report the number of objects in the Plant site area as an indicator of map maturity, falling over time.
This way the map grows from things that are certain to things that are less certain, not the other way round.
Example of the pace of closing gaps
| Parameter | Value |
|---|---|
| Objects without position at the start | 90 pcs |
| Objects without position after 4 quarters | 10 pcs |
Result: A decrease of 80 objects, about 89%.
Note: Example calculation; the figures are assumptions for your own planning.
Checklist
Video
Frequently asked questions
Is an object in the Plant site area visible on the map?
No. It has no geometry or point, so it does not appear on the map until a position is assigned, but it remains available for selection in the register.
Why not put them all in the middle of the hall?
Because a pin suggests knowledge that does not exist. The person accepting a permit could assess the risk based on a false position.
What works as a test of coordinate conversion?
Converting a set of control points forward and back and measuring the largest deviation. In the test described, it was 0.01 m across 64 points.
Do you have a register with position gaps?
We will show how to move such a register onto the map safely and how to close the gaps without false pins.
Let's talk