Diego Ripley
5.9 KiB
title, weight
| title | weight |
|---|---|
| Orthoimagery | 3 |
I briefly worked on collecting orthoimagery a couple of years ago. It started from this. Current data processing pipeline is being defined, but you can preview some of the datasets from the process.
Development Environment
While a key goal is to utilize open source as much as possible, We utilize the proprietary software MapTiler Engine Pro as it is superior to current open source solutions. A limitation is that we are limited to 8 CPU cores for rendering of the map tiles.
Specifications of Tile Packages
The specifications for the tile packages are defined in this code.
#!/bin/bash
PROJECT_DIR="~/Documents/Personal/Projects/dataforcanada-orthoimagery"
DATASET_ID="ca-mb_winnipeg-2024A00054611040_orthoimagery_2024_075mm"
DATA_DIR="${PROJECT_DIR}/data"
DATA_INPUT_DIR="${DATA_DIR}/input/${DATASET_ID}"
DATA_OUTPUT_DIR="${DATA_DIR}/output/${DATASET_ID}"
MBTILES_OUTPUT_FILE="${DATA_OUTPUT_DIR}/${DATASET_ID}.mbtiles"
PMTILES_OUTPUT_FILE="${DATA_OUTPUT_DIR}/${DATASET_ID}.pmtiles"
# Define arguments in an array
ARGS=(
-progress
-name "City of Winnipeg Orthoimagery for 2024 / Ortho-imagerie de la Ville de Winnipeg de 2024"
-description "Orthoimagery 7.5cm resolution. / Ortho-imagerie à résolution de 7,5 cm."
-attribution "Source: data.winnipeg.ca / Source: data.winnipeg.ca"
-srs_epsg
-mbtiles_compatible
-wo "NUM_THREADS=ALL_CPUS"
-wo "USE_OPENCL=TRUE"
-sparse
-scale 2.000000
-work_dir ~/tmp/maptiler_engine
-f webp32
-webp_quality 85
-webp_lossy
-webp_preset photo
-resampling cubic
-overviews_resampling average
-o "${MBTILES_OUTPUT_FILE}"
$DATA_INPUT_DIR/*.ecw
)
# Run the command with the array
maptiler-engine "${ARGS[@]}"
pmtiles convert --tmpdir=~/tmp/pmtiles ${MBTILES_OUTPUT_FILE} ${PMTILES_OUTPUT_FILE}
Download and Preview
Here is a table of some of the datasets created from the current process.
The Plan
The plan is to include the original dataset, for example, the orthoimagery files from Vancouver are provided in MrSID and ECW file formats, which are proprietary and require special drivers being used. I am looking into using formats such as Cloud Optimized GeoTIFFs, but I have to make sure that there is no degradation of visual quality in the process. I hope to experiment with multispectral data in the future.