Proposed for FilterMate 5.0
Beyond Vectors:
Raster & LiDAR Filtering
A proposed direction for raster zonal statistics and point-cloud filtering, presented here for discussion and planning.
Planning document. FilterMate 4.9.2 filters vector layers. The features, performance targets and schedule below are not part of the current release and may change.
8
EPICs
17
User Stories
5
Sprints
Raster Filtering
Filter Vectors by What's Underneath
The proposed workflow calculates raster statistics inside polygons, then uses those values as filter criteria.
Key feature
Zonal Statistics as Filter
Compute min, max, mean, standard deviation and median of raster values inside each polygon. Then filter: "show only parcels where average elevation > 500 m" or "exclude zones with high NDVI variance". The design calls for background processing to keep the UI responsive.
Raster Value Sampling
Sample raster values at the centroid of every vector feature, with multi-band support, CRS reprojection and NoData handling. Design target: under 500 ms for 10,000 points on the reference benchmark.
Dynamic Raster Highlight
When a zonal filter is active, pixels outside filtered polygons would turn transparent. The proposed visual mask can be toggled without losing the filter; interactive performance is a design target, not a current measurement.
Interactive Histogram
Visualize the distribution of raster values across your zones. Click and drag on the histogram to set filter thresholds. Updates live as you change bands or layers.
Raster Clip & Export
Export the raster clipped to your filtered polygons as GeoTIFF or Cloud-Optimized GeoTIFF (COG). Metadata, CRS and resolution preserved.
Dual Panel UI
Toggle between vector and raster panels with a single click. Auto-detects layer type. Your existing vector workflows remain untouched.
LiDAR & Point Cloud
Filter 3D Point Clouds Natively
A proposed interface for filtering point clouds by classification, elevation, intensity and other attributes.
Key feature
ASPRS Classification Filter
Check and uncheck point classes visually: Ground, Vegetation, Buildings, Water, Noise and more. Standard ASPRS labels (0–18) would use readable names, with only the classes found in the data displayed.
Elevation & Attribute Sliders
Filter by Z range, intensity, return number or any attribute. Sliders auto-calibrate to your data's min/max. Combine freely with classification filters.
Multi-Criteria Combiner
Stack classification + elevation + intensity into a single combined query. Toggle each criterion independently. Save and reload filter presets.
Polygon Clip & LAZ Export
Clip your point cloud to a vector polygon. Export filtered points to LAZ or COPC via PDAL. Requires QGIS 3.32+; gracefully hidden on older versions.
Roadmap
Sprint Timeline
A draft sequence of five estimated sprints, starting with raster foundations and then point-cloud support.
0
Sprint 0 — Foundations (1.5 weeks)
Raster Foundations & Cherry-Pick
Merge the raster domain layer from the development branch, wire the dual-panel UI into the refactored dockwidget and keep the existing test suite passing.
1
Sprint 1 — Raster Sampling (1.5 weeks)
Raster Info + Value Sampling
Display raster metadata, sample values at polygon centroids with automatic reprojection and add multi-band selection. Target benchmark: under 500 ms for 10,000 features.
Alpha Raster
2
Sprint 2 — Zonal Stats (2.5 weeks)
Zonal Statistics + Interactive Histogram
Background zonal statistics, polygon filtering by thresholds and an interactive histogram. Target benchmark: 100 polygons on a 1 GB raster in under 10 seconds.
Beta Raster
3
Sprint 3 — Highlight + PC Foundations (2.5 weeks)
Raster Highlight + Point Cloud Architecture
Dynamic pixel masking for filtered zones. In parallel: hexagonal port/adapter for point clouds, conditional detection (QGIS 3.26+), ASPRS classification filter.
Raster V1 + Alpha Point Cloud
4
Sprint 4 — Point Cloud Complete (2 weeks)
Elevation Sliders + Multi-Criteria + Raster Export
Attribute/Z range filters for point clouds. Multi-criteria combiner. Raster clip-to-polygon export. Full test suite.
FilterMate 5.0
Want to Shape the Future?
FilterMate is open source. Star the repo, report issues, or join the conversation on Discord.