Skip to content
Camera orientation from monocular images based on edgels
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
EXPERIMENTS
aws
corisco
examples
filtersqp
.gitignore
README.md

README.md

corisco

Camera orientation from monocular images based on edgels

This is a rewrite from the project hosted at https://code.google.com/p/corisco/

How to use

We recommend the use of virtualenv to install this experimental code. Corisco can then be installed as usual for any python package, with:

$ python setup.py develop

We note that the package contains some dependencies such as Cython and PIL.

The corisco command line tool needs to read a json file that defines the intrinsic paramenters from the camere used to capture an image set, and also defines the filenames. An example files and a few xample images are packaged along with corisco. Once corisco is installed you can test it with these images by running:

$ cd examples/
$ corisco set-apasample.json 0 --do_plot

You should see two graphics be plotted, and also an output such as

{"Nedgels": 16249, "ransac_ori_est": [0.9364874945887336, -0.03621275256917897, 0.34785885021235524, -0.02596207540275075], "sqp_itr": 54, "ori_est": [0.9258680293148183, -0.032798202485475124, 0.3760959820129496, -0.015635936816569097], "time": {"ransac": 1.5499999999999998, "total": 2.89, "edgel": 0.16, "sqp": 1.1500000000000001}, "input": {"img_md5": "cf6b240cb4630e6a2aa36138077dec3b", "proc_args": {"ransac_itr": 1000, "int_parm": {"projection_model": "harris", "principal_point": [500, 375.0], "focal_distance": 847.667796003, "distortion_coefficient": 111.897839472}, "err_func": [3.0, 1.0, 0.15], "op_tol": 1e-12, "smooth": null, "grid": {"lim": 20.0, "method": "shigeru", "size": 4}}}}
You can’t perform that action at this time.