Public Member Functions | |
def | __init__ |
def | build_sparsity_dict |
def | compute_cov |
def | compute_expected |
def | compute_expected_J |
def | compute_marginal_gamma_sqrt |
def | compute_residual |
def | compute_residual_scaled |
def | get_measurement |
def | get_residual_length |
def | update_config |
Public Attributes | |
sensor_id | |
sensor_type | |
terms_per_sample | |
Private Member Functions | |
def | _compute_expected |
Private Attributes | |
_camera | |
_chain | |
_config_dict | |
_M_cam | |
_M_chain |
Definition at line 94 of file camera_chain_sensor.py.
def pr2_calibration_estimation::sensors::camera_chain_sensor::CameraChainSensor::__init__ | ( | self, | ||
config_dict, | ||||
M_cam, | ||||
M_chain | ||||
) |
Generates a single sensor block for a single configuration Inputs: - config_dict: The configuration for this specific sensor. This looks exactly like a single element from the valid_configs list passed into CameraChainBundler.__init__ - M_cam: The camera measurement of type calibration_msgs/CameraMeasurement - M_chain: The chain measurement of type calibration_msgs/ChainMeasurement
Definition at line 95 of file camera_chain_sensor.py.
def pr2_calibration_estimation::sensors::camera_chain_sensor::CameraChainSensor::_compute_expected | ( | self, | ||
chain_state, | ||||
target_pts | ||||
) | [private] |
Compute what measurement we would expect to see, based on the current system parameters and the specified target point locations. Inputs: - chain_state: The joint angles of this sensor's chain of type sensor_msgs/JointState. - target_pts: 4XN matrix, storing features point locations in world cartesian homogenous coordinates. Returns: target points projected into pixel coordinates, in a Nx2 matrix
Definition at line 194 of file camera_chain_sensor.py.
def pr2_calibration_estimation::sensors::camera_chain_sensor::CameraChainSensor::build_sparsity_dict | ( | self | ) |
Build a dictionary that defines which parameters will in fact affect this measurement. Returned dictionary should be of the following form: transforms: my_transformA: [1, 1, 1, 1, 1, 1] my_transformB: [1, 1, 1, 1, 1, 1] dh_chains: my_chain: dh: - [1, 1, 1, 1] - [1, 1, 1, 1] | - [1, 1, 1, 1] gearing: [1, 1, ---, 1] rectified_cams: my_cam: baseline_shift: 1 f_shift: 1 cx_shift: 1 cy_shift: 1
Definition at line 276 of file camera_chain_sensor.py.
def pr2_calibration_estimation::sensors::camera_chain_sensor::CameraChainSensor::compute_cov | ( | self, | ||
target_pts | ||||
) |
Computes the measurement covariance in pixel coordinates for the given set of target points (target_pts) Input: - target_pts: 4xN matrix, storing N feature points of the target, in homogeneous coords
Definition at line 237 of file camera_chain_sensor.py.
def pr2_calibration_estimation::sensors::camera_chain_sensor::CameraChainSensor::compute_expected | ( | self, | ||
target_pts | ||||
) |
Compute the expected pixel coordinates for a set of target points. target_pts: 4xN matrix, storing feature points of the target, in homogeneous coords Returns: target points projected into pixel coordinates, in a Nx2 matrix
Definition at line 186 of file camera_chain_sensor.py.
def pr2_calibration_estimation::sensors::camera_chain_sensor::CameraChainSensor::compute_expected_J | ( | self, | ||
target_pts | ||||
) |
The output Jacobian J shows how moving target_pts in cartesian space affects the expected measurement in (u,v) camera coordinates. For n points in target_pts, J is a 2nx3n matrix Note: This doesn't seem to be used anywhere, except maybe in some drawing code
Definition at line 213 of file camera_chain_sensor.py.
def pr2_calibration_estimation::sensors::camera_chain_sensor::CameraChainSensor::compute_marginal_gamma_sqrt | ( | self, | ||
target_pts | ||||
) |
Calculates the square root of the information matrix for the measurement of the current set of system parameters at the passed in set of target points.
Definition at line 153 of file camera_chain_sensor.py.
def pr2_calibration_estimation::sensors::camera_chain_sensor::CameraChainSensor::compute_residual | ( | self, | ||
target_pts | ||||
) |
Computes the measurement residual for the current set of system parameters and target points Input: - target_pts: 4XN matrix, storing features point locations in world cartesian homogenous coordinates. Output: - r: 2N long vector, storing pixel residuals for the target points in the form [u1, v1, u2, v2, ..., uN, vN]
Definition at line 126 of file camera_chain_sensor.py.
def pr2_calibration_estimation::sensors::camera_chain_sensor::CameraChainSensor::compute_residual_scaled | ( | self, | ||
target_pts | ||||
) |
Computes the residual, and then scales it by sqrt(Gamma), where Gamma is the information matrix for this measurement (Cov^-1).
Definition at line 143 of file camera_chain_sensor.py.
def pr2_calibration_estimation::sensors::camera_chain_sensor::CameraChainSensor::get_measurement | ( | self | ) |
Get the target's pixel coordinates as measured by the actual sensor
Definition at line 179 of file camera_chain_sensor.py.
def pr2_calibration_estimation::sensors::camera_chain_sensor::CameraChainSensor::get_residual_length | ( | self | ) |
Definition at line 174 of file camera_chain_sensor.py.
def pr2_calibration_estimation::sensors::camera_chain_sensor::CameraChainSensor::update_config | ( | self, | ||
robot_params | ||||
) |
On each optimization cycle the set of system parameters that we're optimizing over will change. Thus, after each change in parameters we must call update_config to ensure that our calculated residual reflects the newest set of system parameters.
Definition at line 116 of file camera_chain_sensor.py.
Definition at line 122 of file camera_chain_sensor.py.
Definition at line 112 of file camera_chain_sensor.py.
Definition at line 108 of file camera_chain_sensor.py.
Definition at line 109 of file camera_chain_sensor.py.
Definition at line 110 of file camera_chain_sensor.py.
Definition at line 106 of file camera_chain_sensor.py.
Definition at line 105 of file camera_chain_sensor.py.
Definition at line 114 of file camera_chain_sensor.py.