Return the html index file for this folder (if appropriate)

Implement option --stretch

Intersect the ray from the given pixel with surface at this height above the datum. Use Newton's method. The obtained intersection point must project

Given the albedo image, and the valid mask, in-fill from valid pixels to invalid ones using 3x3 neighborhood averaging. Do a single pass. This is most

Given a set of integers in increasing order, with each assigned to a group, return the index of the current integer in its group.

For the cameras that are of frame type, copy the initial values of camera position and orientation to the vector of variables that we will optimize. H

Apply the input adjustments to the CSM cameras. Resample linescan models. Get pointers to the underlying CSM cameras, as need to manipulate those dire

Find the best-fit affine transform, that will serve as initial guess for RPC

Initialize the position and orientation of a pinhole camera model using GCP. It invokes OpenCV's PnP functionality.

Create the param storage. Collect in it any input adjustments and initial transform. Return a copy of the cameras having these adjustments applied to

Set all camera position and pose values to zero.

TODO: Share more code with the similar pinhole case. Specialization for CSM

TODO: Share more code with the similar pinhole case. Specialization for optical bar cameras.

Specialization for pinhole cameras

Initialize the position and orientation of each pinhole camera model using a least squares error transform to match the provided camera positions. Thi

Compute an initial source to reference transform based on tie points (interest point matches).

Initializations needed for various modes

This is kind of like part 2 of the constructor - This function finalizes the spacing and generates a spacing-snapped BBox.

A helper function to do interpolation. Will not interpolate when exactly on the grid. This does not interpolate when the weight is 0, as that suggests

Given a map from timestamps to frame camera indices in csm_models, do linear interpolation in time. Return false if out of bounds. TODO(oalexan1): Mus

Given a set of poses indexed by time, interpolate or extrapolate (within range of bracket_len) at a set of target timestamps. Go forward in time both

Get positions. Based on the UsgsAstroLsSensorModel code. TODO(oalexan1): Move this to a new CsmModelUtils.cc file and

See if a given image has only invalid pixels

Filter IP using a given DEM and max height difference. Assume that the interest points have alignment applied to them (either via a transform or from

Return true if a file is a recognized DEM.

Return true if the input value can be converted to a float.

Return true if the file is an input (not converted) lidar format

A little helper function

If --trans-crop-win is in the input arguments, this is means that we are running stereo for a tile

Fetch metadata from an ISIS cube for saving to the output CSM camera.

Run one pass of solving for jitter. At each pass the cameras we have so far are used to triangulate the points and the DEM constraint is refreshed if

Create keypoint_vec from keypoint_map. That just reorganizes the data to the format expected later.

We will fetch a chunk of the las file of area ASP_POINT_CLOUD_TILE_LEN x ASP_POINT_CLOUD_TILE_LEN, split it into bins of spatially close points, and w

Compute alignment using least squares

Generate a DEM from a lidar file in the given region (plus a buffer)

Fetch only lidar files for the given frame range. Do that as follows. For each ortho frame in [startFrame, stopFrame], find the lidar fi

Solves for Affine Fundamental Matrix as per instructions in Multiple View Geometry. Outlier elimination happens later.

Returns the number of lines is the ISIS cube

Get quaternions (only for linescan cameras)

Load cameras and sun positions

Find the bounding box of all DEMs in the projected space. - mosaic_bbox is the output bounding box in projected space - dem_proj_bboxes and dem_pixel_

Load DEMs that intersect with the current tile batch. This filters the input DEMs to only those needed for the tiles being processed, and prepares the

Looks in the input camera position file to generate a GCC position for each input camera. - If no match is found, the coordinate is (0,0,0)

Read the georeference for the DEM and albedo

Load masked images and compute blending weights. These weights will be adjusted later.

Load all of the reference disparities specified in the input text file and store them in the vectors. Return the number loaded.

Read the reference terrain. Ensure enough reference points are loaded that project in the camera boxes. Do two attempts. TODO(oalexan1): This will nee

Validate the bundle_adjust options. This modifies the options structure.

Load the D_sub file fully in memory

Load xyz points from disk into a matrix with 4 columns. Last column is just ones.

For each option, the string must include a subset of the entries: "focal_length, optical_center, distortion_params" - Need the extra boolean to handle

Try to load at least this many points from the LAS file. TODO(oalexan1): This function should reduce the number of points if they are too many.

Set up the next interpolator - Returns false if the end of the file was reached

Loads in the points from a text file

Load a PeruSat linescan camera file

Load PeruSat camera as a CSM linescan model. Resampling is done here (not in PeruSatXML) so the XML class returns raw data. Positions/velocities: Lagr

Load the DEM and adjust some options depending on DEM statistics.

Algorithm to perform local alignment. Approach: - Given the global interest points and the left crop window, find the right crop window. - Crop the gl

Identify the appropriate inputs in the ASTER directory

A function to extract poses, filenames, and timestamps from read data.

Look up images for a set of rigs. This requires looking up images for individual rigs, then concatenating the results and adjusting the book-keeping.

Look up each ref cam image by timestamp, with the rig assumption. In between any two ref cam timestamps, which are no further from each other than the

Assuming that the rig constraint is not used, initialize the 'cams' structure by copying each image and its other data in that structure as expected b

Look up images, with or without the rig constraint. See individual functions below for more details.