BlomURBEX - Measuring accuracy / guidance
BlomURBEX is a powerful tool for rapidly accessing detailed information. As with all tools, there is a need for the user to be aware of the potential for (accidental) misuse in order to maximise confidence in the results derived.
We have identified a few key areas which affect accuracy:
Understanding the "elevation" tool
Perhaps the most significant potential for risk with a tool such as BlomURBEX is user interpretation.
Unlike mapping, aerial imagery whether vertical or oblique is not clearly broken into clear lines and areas and users can not simply snap to a building corner.
User uncertainty is generally related to the following conditions:
- measuring in shadow
- measuring around vegetation
- heights - what is ground level
- heights - what is roof level
The potential for error in such situations can be reduced by taking mulitple measurements:
- of the same point
- of similar points
- from different views
The primary capture of BlomURBEX datasets has been to an average pixel size of 15cm (i.e. each pixel represents 15cm in the real world). Future phases of capture are set to improve this with 10cm capture already progressing.
Pixel size consequently offers an immediate source for potential inaccuracy. Any measurement will be limited to the resolution of the raw data and even a pixel error at either end will equate to an error in the order of 30cm.
All heights require the measurement to be taken vertically, the system will not return a correct value if the line is drawn between two points with differing plan positions.
Understanding the "elevation" tool:
The elevation tool is only intended to return "ground" heights. As such, points taken on buildings or vegetation will represent the interpreted "bald earth" height at that point.
The data is taken from Ordnance Survey Profile and as such is further subject to the accuracies of that dataset (published at +/- 2.5 m metres with 95% RMSE)
