sba_python repository

Repository Summary

Checkout URI https://github.com/safijari/sba_python.git
VCS Type git
VCS Version python-devel
Last Updated 2019-10-16
Dev Status DEVELOPED
CI status No Continuous Integration
Released UNRELEASED
Package Tags No category tags.
Contributing Help Wanted (0)
Good First Issues (0)
Pull Requests to Review (0)

Packages

Name Version
sba_python 0.4.2

README

* Sparse Bundle Adjustment Library
  Originally developed at Willow Garage as part of the vslam stack, 
  this library is currently used by =open_karto=. The python wrapper
  currently only supports 2D mode and is being used as a backend for
  [[https://github.com/safijari/mp-slam][=mp-slam=]].

* Simple usecase example
  #+begin_src python
from sba_cpp import SPA2d, Node2d
import numpy as np

s = SPA2d()

"""
Test prep

start at x = 0, y = 0, yaw = 0 (0, 0, 0, 0)
start at x = 1, y = 1, yaw = 0 (1, 1.1, 1.1, 0)
start at x = 0, y = 1, yaw = 0 (2, 0.1, 1.1, 0)
"""

# x, y, yaw, node_id
s.add_node(0, 0, 0, 0)
s.add_node(1.1, 1.1, 0, 1)
s.add_node(0.1, 1.1, 0, 2)

# inverse of covariance
precision_matrix = np.identity(3)

# from_node_id, to_node_id, xdiff, ydiff, yawdiff, precision matrix
s.add_constraint(0, 1, 1.1, 1.1, 0, precision_matrix.tolist())
s.add_constraint(1, 2, -1.1, 0.1, 0, precision_matrix.tolist())
s.add_constraint(2, 0, -0.1, -1.1, 0, precision_matrix.tolist())

# the parameters below are
# max iterations, diagonal augmentation for LM
# use CSparse (set to True for the really fast version)
# initial tolerance for CG
# max iterations for some step in LM
s.compute(100, 1.0e-4, True, 1.0e-8, 50)

# The above values are defaults in the original c++ code and they work
# well there

for n in s.nodes:
    print(n.x, n.y, n.yaw)
  #+end_src


CONTRIBUTING

No CONTRIBUTING.md found.

Repository Summary

Checkout URI https://github.com/safijari/sba_python.git
VCS Type git
VCS Version python-devel
Last Updated 2019-10-16
Dev Status DEVELOPED
CI status No Continuous Integration
Released UNRELEASED
Package Tags No category tags.
Contributing Help Wanted (0)
Good First Issues (0)
Pull Requests to Review (0)

Packages

Name Version
sba_python 0.4.2

README

* Sparse Bundle Adjustment Library
  Originally developed at Willow Garage as part of the vslam stack, 
  this library is currently used by =open_karto=. The python wrapper
  currently only supports 2D mode and is being used as a backend for
  [[https://github.com/safijari/mp-slam][=mp-slam=]].

* Simple usecase example
  #+begin_src python
from sba_cpp import SPA2d, Node2d
import numpy as np

s = SPA2d()

"""
Test prep

start at x = 0, y = 0, yaw = 0 (0, 0, 0, 0)
start at x = 1, y = 1, yaw = 0 (1, 1.1, 1.1, 0)
start at x = 0, y = 1, yaw = 0 (2, 0.1, 1.1, 0)
"""

# x, y, yaw, node_id
s.add_node(0, 0, 0, 0)
s.add_node(1.1, 1.1, 0, 1)
s.add_node(0.1, 1.1, 0, 2)

# inverse of covariance
precision_matrix = np.identity(3)

# from_node_id, to_node_id, xdiff, ydiff, yawdiff, precision matrix
s.add_constraint(0, 1, 1.1, 1.1, 0, precision_matrix.tolist())
s.add_constraint(1, 2, -1.1, 0.1, 0, precision_matrix.tolist())
s.add_constraint(2, 0, -0.1, -1.1, 0, precision_matrix.tolist())

# the parameters below are
# max iterations, diagonal augmentation for LM
# use CSparse (set to True for the really fast version)
# initial tolerance for CG
# max iterations for some step in LM
s.compute(100, 1.0e-4, True, 1.0e-8, 50)

# The above values are defaults in the original c++ code and they work
# well there

for n in s.nodes:
    print(n.x, n.y, n.yaw)
  #+end_src


CONTRIBUTING

No CONTRIBUTING.md found.

Repository Summary

Checkout URI https://github.com/safijari/sba_python.git
VCS Type git
VCS Version indigo-devel
Last Updated 2018-08-19
Dev Status DEVELOPED
CI status No Continuous Integration
Released UNRELEASED
Package Tags No category tags.
Contributing Help Wanted (0)
Good First Issues (0)
Pull Requests to Review (0)

Packages

Name Version
sparse_bundle_adjustment 0.3.2

README

Sparse Bundle Adjustment Library
================================

Originally developed at Willow Garage as part of the vslam stack.


CONTRIBUTING

No CONTRIBUTING.md found.