vmap

Velocity map generation using the NASA Ames Stereo Pipeline (ASP) image correlator

Overview

The ASP/VW correlator is fast, memory-efficient, multi-threaded, and generalized. This means it can be used for efficient feature tracking operations on two arbitrary input rasters acquired at different times.

Example

Command-line examples (run with no arguments for usage)

Generate a velocity map from two orthorectified images (or any raster in projected coord sys, e.g. high-res DEM shaded relief maps with identical illumination):

• vmap.py img1_YYYYMMDD.tif img2_YYYYMMDD.tif
• Generates ASP disparity maps ('-RD.tif', '-F.tif', '-F_smooth.tif')
• By default, calls disp2v.py on '-F.tif' to generate output velocity products
• Correlator options are currently hardcoded - can edit directly in main(), but eventually will be accepted as command-line arguments

Convert output ASP disparity map to m/yr or m/day:

• disp2v.py img1_YYYYMMDD__img2_YYYYMMDD_vmap/img1_YYYYMMDD__img2_YYYYMMDD_vmap-F.tif
• Output: '_vm.tif' (velocity magnitude in m/yr or m/day), '_vx.tif' (horizontal velocity component), '_vy.tif' (vertical velocity component)

Preview subsampled version of disparity map:

• disp_view.sh img1_YYYYMMDD__img2_YYYYMMDD_vmap/img1_YYYYMMDD__img2_YYYYMMDD_vmap-F.tif
• Creates '-F_b1_5.tif' and '-F_b2_5.tif' and opens linked view with imview.py

Sorry about the long filenames.

Installation

Install the latest release from PyPI:

pip install vmap 

Note: by default, this will deploy executable scripts in /usr/local/bin

Building from source

Clone the repository and install:

git clone https://github.com/dshean/vmap.git
pip install -e vmap/

The -e flag ("editable mode", setuptools "develop mode") will allow you to modify source code and immediately see changes.

Notes on ASP Install

These scripts depend on several command line utilities bundled with the NASA Ames Stereo Pipeline (ASP). These must be in your executable path at runtime. To do this:

1. Download the latest ASP binaries for your operating system (Linux or OS X). For bleeding edge, see the daily builds.
2. Extract: tar -xzvf StereoPipeline-*-x86_64-*.tar.gz (filling in the appropriate filename)
3. Move the extracted folder to a location on your system where you store executables (e.g., bindir="~/sw" or bindir="/usr/local/bin"): mv StereoPipeline-*-x86_64-*/ $bindir
4. Add the ASP bin subdirectory to your path. In bash: export PATH=${PATH}:$bindir/StereoPipeline-*-x86_64-*/bin. To make this permanent, add that line to your ~/.bashrc or ~/.bash_profile.

Core requirements

• NASA Ames Stereo Pipeline (ASP)
• GDAL/OGR
• NumPy
• pygeotools

Optional requirements (needed for some functionality)

• imview
• demcoreg (offset correction over bare ground surfaces)
• matplotlib
• SciPy

License

This project is licensed under the terms of the MIT License.