stereopipeline-quickstart

A guide to the NASA Ames Stereo Pipeline (ASP), runnable in a pre-configured GitHub Codespace with ASP and asp_plot installed.

Open this repo in a GitHub Codespace

This site is the static companion to the GitHub repo. The repo contains the notebooks; this site explains why each step exists.

Pick your path

New to stereo photogrammetry?

What ASP does, what a “stereo pipeline” is, and where the moving parts fit. Read this first if you’re not sure why bundle adjustment exists.

What is the Ames Stereo Pipeline?

Ready to run something?

Launch the Codespace, open a notebook, hit Run All.

Run the tutorials in a Codespace

Want a deeper concept primer?

The end-to-end mental model: from raw imagery to aligned DEM. With diagrams.

Pipeline overview

Want to run the tutorials?

Two walkthroughs: ASTER (medium-res) and WorldView-3 (high-res). Both fully open.

Tutorials

Need to look something up?

Glossary, ASP output-file naming, parameter cheat sheet, links to the canonical ASP docs.

Glossary

Want pretty plots from your run?

How asp_plot produces diagnostic figures, PDF reports, and ICESat-2 comparisons.

Visualization with asp_plot

Curious about alternative open-source stereo pipelines?

The same WorldView-3 pair run through three open-source stereo pipelines, side by side.

Alternative Open-Source Stereo Pipelines

What this is not

A replacement for the official ASP documentation. ASP has hundreds of options for dozens of sensors. This guide covers the path of least resistance: two openly-available datasets, a small set of parameters, and explanations of why each step exists. The official docs are the authoritative reference.

Contents

Get started

• What is the Ames Stereo Pipeline?
  • A short history
  • A toolchain of modular executables
  • Why “stereo”?
  • Where asp_plot fits
  • What’s next
• Run the tutorials in a Codespace
  • Launch
  • Machine size
  • Run a tutorial
  • Persistence and storage
  • Costs

Concepts

• Pipeline overview
  • The five stages
  • Stage 1: Sensor preparation
  • Stage 2: Bundle adjustment
  • Stage 3: Orthorectification
  • Stage 4: Stereo + DEM generation
  • Stage 5: Alignment
  • What asp_plot does at every stage
• Stereo photogrammetry
  • What ASP actually computes
  • Knobs that matter
  • Geometry that matters
  • Where matching fails
  • Where to read more
• Bundle adjustment
  • What the optimizer does
  • Why this matters for stereo
  • Outputs you can visualize
  • Common knobs
  • Where to read more
• Orthorectification
  • The intuition
  • When you can skip it
  • What reference DEM to use
  • The two-pass trick
  • Where to read more
• Alignment
  • The reference dataset
  • What pc_align does
  • When pc_align helps and when it doesn’t
  • Reading the alignment report
  • The --pc_align flag in asp_plot
  • Where to read more
• Visualization with asp_plot
  • Two ways to use it
  • Where to read more

Tutorials

• Tutorials
  • Suggested order
  • Beyond the tutorials
• Tutorial 1 — ASTER Mt. Rainier
  • What we’ll do
  • 1. Download the scene
  • 2. Sensor prep — aster2asp
  • 3. Stereo (no orthorectification) — first pass
  • 4. Inspect the rough DEM with asp_plot
  • 5. Orthorectified pass
  • 6. Generate the full PDF report
  • What’s next
• Tutorial 2 — WorldView-3 UCSD
  • Pair details
  • What we’ll do
  • 1. Download the imagery from SpaceNet
  • 2. Stereo geometry
  • 3. Reference DEM from AWS Copernicus DEM
  • 4. Bundle adjustment
  • 5. Mapproject
  • 6. Stereo + DEM
  • 7. Inspect with asp_plot
  • 8. Full report with ICESat-2 alignment
  • What’s next

Comparisons

• Alternative Open-Source Stereo Pipelines
  • Key Differences

Reference

• Glossary
• ASP output files
  • After bundle_adjust -o ba/run
  • After mapproject
  • After parallel_stereo -o stereo/run
  • After point2dem
  • After pc_align
• Local install
  • Requirements
  • Step 1 — Install ASP
  • Step 2 — Install asp_plot
  • Step 3 — Clone this repo and run a tutorial
  • Conda-only install (no binary download)
  • Verifying everything works
• Further reading
  • ASP itself
  • Visualization and diagnostics
  • Reference data
  • Companion repositories