Given all the merged and filtered tracks in pid_cid_fid, for each image pair cid1 and cid2 with cid1 < cid2 < cid1 + num_overlaps + 1, save the matche

Save the measured and computed intensity images

Create from scratch and and save an ISIS cnet from a given control network and latest param values.

Save mapprojected matches offsets for each image pair having matches

Once bundle adjustment is done, export the outlier list, camera poses, triangulated points, and optical offsets. This is needed for saving an NVM file

Write an optical bar camera file to disk after updating the intrinsics and extrinsics. Return the path to the saved file.

Find convergence angles between every pair of images and save to disk their percentiles assumed that the cameras in world_to_cam are up-to-date given

Write a pinhole camera file to disk after updating the intrinsics and extrinsics. Return the path to the saved file.

Save some info that will be useful for peeking at a run

Save the points as a shapefile

Save the depth clouds and optimized transforms needed to create a mesh with voxblox (if depth clouds exist).

Apply a transform to a point cloud to make it go from camera coordinates to world coordinates and save it as a ply file.

Write an updated csm camera state file to disk. Assumes no intrinsics are optimized.

Update an ISIS cnet with the latest info on triangulated points and outliers, and write it to disk at <outputPrefix>.net. We do not change here if a p

Write an updated RPC camera file to disk. Assumes no intrinsics are optimized.

Write an image with 3 floats per pixel. OpenCV's imwrite() cannot do that.

Save the DEM

Save the intersection error

A small function to save the l-r-diff file

Form the mtl file having the given png texture

Save the orthoimage. The texture comes form L.tif instead of the heights.

Save a pinhole camera, in tsai or csm format

Save horizontal and vertical stddev

Save an XML file having all RPC information

Save an XML file having all RPC information

Based on tabulated satellite position and quaternion covariance for each DG camera, find the interpolated covariances for cam1 at pix1 (6 for position

Given two DG cameras and a pixel in each camera image, consider the following transform. Go from the perturbed joint vector of satellite positions and

Estimate the search range by finding the median disparity and creating a box of given dimensions around it. This assumes aligned interest points. Note

Convert from integer number to HHMMSS string.

Set the Isis cnet target based on the image name. If that fails, as for Earth, set it to the datum name.

Load the camera model from a model state written to disk. A model state is obtained from an ISD model by pre-processing and combining its data in a fo

Set the projection based on options. By now opt.proj_lon and opt.proj_lat should have been set.

Copy weights from separate weight image into the alpha channel of the DEM.

A function to set the environmental variables ISISROOT, QT_PLUGIN_PATH, GDAL_DATA, and PROJ_LIB. In packaged build mode, set these with the help of th

Set the focal length

Update the values for a point

If the DEM spacing is set to zero, we compute a DEM with approximately the same pixel dimensions as the input image. Note, however, that this could le

Load a target location vector.

Allows us to set the time offset

Set up block sizes for various cost functions

Bring crops in memory. It greatly helps with multi-threading speed. This function is used only for small tiles, to avoid running out of memory (which

Prepare a DEM file that encompasses a given image and with a given height.

Heuristics for setting up the no-data value

Choose the images to match and load them. It is assumed that in image_files we have the images from the first and then he second maps to merge.

Read the mapprojected data, if having the option --mapprojected-data.

Set up output mosaic geometry: georef, dimensions (cols/rows), spacing. Handles projection changes, resolution, alignment (tap/gdal-tap), and projwin

Prepare a DEM file that encompasses a given image and with a given height, or reuse the one already available if agrees with what is intended.

If --ref-map is passed in, crop to its extent. This will create opt.crop_win.

Initalize the reflectance parameters based on user input

Set up the optimization problem for rig calibration

Setup tile grid parameters and validate configuration. Determines number of tiles, checks for single-file output mode, validates tile index.

TODO(oalexan1): This should be modularized

Warp the source DEM to the reference DEM's georef, while applying a translation defined in the projected space of the reference DEM.

Given the direction in orbit from first_proj to last_proj (these determine the orbit geometry), find the signed distance along the orbit (in ECEF) fro

A function that first increases slowly, then very fast, but without being numerically unstable for very small or very large values of the input.

Compute the n-weight slerp, analogous to the linear combination W[0]*Q[0] + ... + W[n-1]*Q[n-1]. This is experimental. We assume the sum of weights is

Snap the coordinates of a BBox2 to a grid spacing. Use floor for min and ceil for max, so the box always covers at least the original extent.

Some preliminary work before solving for intrinsics. Here we look up the default calibration file, and generate an RPC approximation of its di

Create a camera model with optimized intrinsics. By now we processed a bunch of images and created bundle-adjusted and pc_aligned cameras and

Solve for a camera translation/rotation based on pixel/gcc coordinate pairs.