MetricAware: adaptive block optimization for more reliable photogrammetry

Ileron, an enterprise with over ten years of experience in UAV data acquisition and processing, announces MetricAware, a proprietary software technology to reduce reprojection error, down to sub-pixel level, while preserving the structure and connectivity of the photogrammetric block.

MetricAware is delivered exclusively as a service: the client submits the acquired images and receives an optimized photogrammetric block ready for import into COLMAP or Metashape. In this way, the optimization phase is integrated into existing workflows without introducing new software to install and maintain.

At the core of MetricAware is an adaptive refinement process guided by the state of the block. At each step, the image with the highest reprojection error is identified and, within that image, the tie points with the most significant local impact on the error are selected. The selected points are removed and the block is recomputed until the predefined reprojection-error target is reached. This combination of per-image targeting and local-impact-based selection makes it possible to reduce the error while preserving the overall connectivity of the block.

On a test dataset, this procedure reduced RMS reprojection error from 2.14 pixels to 0.99 pixels in a first configuration and to 0.70 pixels in a second, while removing 0.25% and 1.21% of tie points respectively. These results show a significant reduction in error with a minimal loss of redundancy.

Key benefits include greater confidence in metric analyses, thanks to reprojection errors constrained at sub-pixel level, and higher quality of derived products, since a well-conditioned block reduces the likelihood of geometrically weak regions and mitigates the appearance of noise, gaps and local distortions in dense point clouds, meshes and DEMs. MetricAware‘s optimization also supports coherent co-registration of multi-epoch surveys on the same site and helps standardize orientation workflows, replacing the manual sparse-cloud pruning stage.

A joint review of the literature, commercial software and available open-source solutions has not identified tools that implement a technology fully similar to that of MetricAware (adaptive optimization guided by the state of the block). Based on this analysis, MetricAware occupies a specific position within the landscape of automatic pruning technologies applied to photogrammetric blocks.

Further information and updates are available on the official page www.metricaware.com or by writing to [email protected]