Raster processing benchmarks

This repository contains a collection of raster processing benchmarks for Python and R packages. The tests cover the most common operations such as loading data, extracting values by points, downsampling, calculating NDVI, writing multilayer, cropping by extent and calculating zonal statistics. The comparison is made from the user's perspective (the simplest functions are used and the code is not optimized), so the results do not represent the best performance.

The detailed results are available at https://kadyb.github.io/raster-benchmark/report.html.

You may also be interested in the vector processing benchmarks.

Software

Python:

• rasterio
• rasterstats
• rioxarray

R:

• stars
• terra
• raster
• exactextractr

Reproduction

1. Download raster data (851 MB) from Google Drive or Earth Explorer (original source, registration required) and then unzip to data/.
2. Run all benchmarks using batch script (run_benchmarks.sh) or single benchmarks files.

Batch script

cd raster-benchmark
./run_benchmarks.sh

Single benchmark

Rscript stars/crop.R

python3 rasterio/crop.py

Dataset

Landsat 8 satellite scene (10 bands, 30 m resolution, 7771 x 7871 pixels) was used for tests.

Scene ID: LC08_L1TP_190024_20200418_20200822_02_T1

Hardware configuration

• CPU: Intel Xeon CPU E5-2620 v2 @ 2.10GHz
• RAM: 64 GB
• OS: Pop!_OS 20.04 LTS

Acknowledgment

Landsat-8 image courtesy of the U.S. Geological Survey, https://earthexplorer.usgs.gov/