Compute the depth-map for the un-rectified image from the given disparity-map of the rectified image

converts the given disparity at the rectified image coordinates into the distance from the camera position to the corresponding 3D point; note: the di

Compute search range from the given disparity-map and setup pixel-map at twice the scale; the validity mask-map is considered as well and upscaled in

export disparity-map as an image (red - maximum disparity, blue - minimum disparity)

extract disparity and range from the pixel-map

compute the distance from the camera to the given 3D point

Query as a double. See IntValue()

Save image as raw data.

Delete also the pointers (take care to use this function only if the elements are pointers).

Enable OpenGL debug output (requires OpenGL 4.3+ debug context)

remove all estimates with too big score and invert confidence map

Return the errorID.

static*/

Return the line where the error occurred, or zero if unknown.

estimate confidence map from depth-map variation in a window; the estimated confidence is the mean of the depth differences to the 3 closer pixel to t

estimate confidence map from normal variance in a window

run propagation and random refinement cycles

estimate the ground-plane as the plane agreeing with most points - planeThreshold: threshold used to estimate the ground plane (0 - auto)

create a virtual point-cloud to be used to initialize the neighbor view from image pair points at the intersection of the viewing directions

Robustly estimate the plane that fits best the given points

EstimatePlane Robustly estimate the plane that fits best the given points, making sure the first point is part of the solution (if any)

EstimatePlaneLockFirstPoint Robustly estimate the plane that fits best the given points using a known threshold

EstimatePlaneTh Robustly estimate the plane that fits best the given points using a known threshold, making sure the first point is part of the soluti

estimate the colors of the given dense point-cloud

estimates the normals through PCA over the K nearest neighbors

estimate robustly the rotation that best maps srcRots to dstRots (dstR = srcR * alignR)

Checks whether the specified key exists is in the hash

export confidence map as an image (dark - low confidence, light - high confidence)

- IDs are the reference view ID and neighbor view IDs used to estimate the depth-map (global ID)

DepthMap2Image

ExportDisparityDataRaw same as above, but exports also the empty border