SiftMatching

SiftMatching.num_threads

default: -1
Number of threads for feature matching and geometric verification.

SiftMatching.use_gpu

default: 1 (enabled)
Whether to use the GPU for feature matching.

SiftMatching.gpu_index

default: -1
Index of the GPU used for feature matching. For multi-GPU matching, you should separate multiple GPU indices by comma, e.g. "0,1,2,3". See: Multi-GPU support in feature extraction/matching

SiftMatching.max_ratio

default: 0.8
Maximum distance ratio between first and second best match.

SiftMatching.max_distance

default: 0.7
Maximum distance to best match.

SiftMatching.cross_check

default: 1 (enabled)
Whether to enable cross checking in matching.

SiftMatching.max_error

default: 4
Maximum epipolar error in pixels for geometric verification.

SiftMatching.max_num_matches

default: 32768
Maximum number of matches.

SiftMatching.confidence

default: 0.999
Confidence threshold for geometric verification.

SiftMatching.max_num_trials

default: 10000
Maximum number of RANSAC iterations. Note that this option overrules the SiftMatching.min_inlier_ratio option.

SiftMatching.min_inlier_ratio

default: 0.25
A priori assumed minimum inlier ratio, which determines the maximum number of iterations.

SiftMatching.min_num_inliers

default: 15
Minimum number of inliers for an image pair to be considered as geometrically verified.

SiftMatching.multiple_models

default: 0 (disabled)
Whether to attempt to estimate multiple geometric models per image pair.

SiftMatching.guided_matching

default: 0 (disabled)
Whether to perform guided matching, if geometric verification succeeds.