CADStoriesGet a quote
EN ⌄

Leica BLK360 G2 accuracy: a façade survey checked against GNSS

Aleksei Pozmogov · Field notes ·

After reviewing the Leica BLK360 G2 and comparing it with the G1, I wanted to look at accuracy on a real façade survey. I used a survey target with a mounting point for a GNSS receiver to compare the registered scan network with external measurements.

The setup: scanner, target and 14 points

The equipment was a Leica BLK360 G2 and a Meridian M20L receiver, working in RTK mode from an EFT base station. The target and receiver were mounted on a tripod. I measured 14 points in a challenging urban environment, keeping points independent for checking. The aim was to compare processing with and without control data.

BLK360 G2 and the survey target on site. Nearby buildings and parked vehicles formed part of the working environment.
BLK360 G2 and the survey target on site. Nearby buildings and parked vehicles formed part of the working environment.
Meridian M20L mounted above the scan target. This arrangement connects GNSS observations to a target visible in the scans; antenna height and mounting geometry need to be accounted for during processing.
Meridian M20L mounted above the scan target. This arrangement connects GNSS observations to a target visible in the scans; antenna height and mounting geometry need to be accounted for during processing.

What the registration report shows

The final bundle contains 42 setups and 91 links. Cyclone Register 360 reports an overall error of 0.002 m and the same cloud-to-cloud error, with 79% overlap. That was a good registration result for this façade dataset. The 2 mm value describes registration consistency; it does not establish 2 mm absolute accuracy for every façade point.

Russian report labels: 42 setups, 91 links; Overall error and Cloud-to-cloud both read 0.002 m. Overlap is 79%, Strength 55%; no target-error value is displayed.
Russian report labels: 42 setups, 91 links; Overall error and Cloud-to-cloud both read 0.002 m. Overlap is 79%, Strength 55%; no target-error value is displayed.

The control-links table reports a mean error of 0.011 m. Checked rows range from 0.004 to 0.019 m. Two further rows, at 0.024 and 0.030 m, are unchecked. The original post's shorthand — from 4 mm, more often around 1 cm — therefore does not describe the full table range. The post does not document why those two rows were unchecked.

Control-links report: mean error 11 mm; checked rows 4–19 mm. Pt9 at 24 mm and Pt11 at 30 mm are unchecked. NEZ columns are error-vector components, not separate scanner specifications.
Control-links report: mean error 11 mm; checked rows 4–19 mm. Pt9 at 24 mm and Pt11 at 30 mm are unchecked. NEZ columns are error-vector components, not separate scanner specifications.

The practical lesson for façade surveys

The BLK360 G2 produced a well-aligned façade cloud in this test. External observations added a useful check beyond the registration figure. GNSS observations in built-up streets carry their own uncertainty, so the residual cannot be attributed entirely to the scanner. Equally, a tight registration alone does not demonstrate centimetre-level georeferencing.

The building point cloud and setup-link network in Cyclone Register 360. For façade work, the connections around the building deserve inspection alongside the summary statistics.
The building point cloud and setup-link network in Cyclone Register 360. For façade work, the connections around the building deserve inspection alongside the summary statistics.

This was an encouraging field result for the compact scanner in my façade-survey workflow. It remains a project-specific check, rather than a metrological certification of the instrument. PPK was not part of this experiment; I left it for a separate comparison.

More field experience with the BLK360 G2

First impressions and kit — part 1
BLK360 G2 versus G1 — part 2

For façade geometry and existing-building survey data in Montenegro, see building surveys for renovation.