dym/cl-opencv

a common lisp (well, sbcl) wrapper around the opencv vision library

README: cl-opencv

Introduction

These are OpenCV bindings for SBCL. They do not promise to be complete, though eventually I hope to have fairly complete coverage of OpenCV functionality. This package depends on the cffi package and libffi. I use Quicklisp to set up my Lisp environment.

Compatibility

This package is known to work in the following configurations.

• Mac OS 10.6, sbcl 1.0.45 (MacPorts), OpenCV 2.2 (MacPorts)
• Ubuntu 10.10 (64-bit), sbcl 1.0.40, OpenCV 2.1

If you have gotten cl-opencv to run with some other combination of OS and software versions, please let me know.

Installation

Mac OS X

1. Install OpenCV from MacPorts: port install opencv
2. Install libffi from MacPorts: port install libffi
3. In your lisp environment, make sure that cffi is available. In Quicklisp you would just do (ql:quickload "cffi").
4. Run make and sudo make install in the top-level directory to build and install the glue code library.

Ubuntu

1. Install OpenCV from the repos: sudo apt-get install libhighgui-dev.
2. Install libffi from the repos: sudo apt-get install libffi
3. In your lisp environment, make sure that cffi is available. In Quicklisp you would just do (ql:quickload "cffi").
4. Run make and sudo make install in the top-level directory to build and install the glue code library.

Troubleshooting

Ubuntu

If CFFI has problems loading the library, be sure that:

• your lisp and your OpenCV library are both 32 bit or both 64 bit; and
• if you installed OpenCV to a non-standard location, you might try adding that path to cffi:*foreign-library-directories*.

Different camera support different resolutions and framerates. The tests assume a camera capable of about 30 fps at 640x480. Adjust the values at the top of test.lisp to match your camera. A program like VLC can help you determine those.