MCPcopy Create free account

hub / github.com/adityamwagh/SuperSLAM / functions

Functions291 in github.com/adityamwagh/SuperSLAM

↓ 134 callersMethodsize
include/GlobalPoseGraph.h:47
↓ 52 callersMethodempty
deleter releases `slot` back to the pool
include/DescriptorPool.h:19
↓ 16 callersMethodadd_keyframe
src/SparseMap.cc:5
↓ 14 callersMethodtrack
src/VoEstimator.cc:203
↓ 11 callersFunctiondesc
Deterministic unit-ish descriptor: a one-hot-ish vector with a small offset so two "places" are near-duplicates when seed matches and distinct when it
tests/test_place_recognizer.cc:13
↓ 10 callersMethodevaluateError
include/PoseOptimizationFactors.h:83
↓ 9 callersMethodadd
tests/test_loop_closer.cc:52
↓ 9 callersMethodvote
src/PlaceRecognizer.cc:54
↓ 8 callersMethodacquire
include/DescriptorPool.h:31
↓ 8 callersFunctionjitter
Deterministic pseudo-noise so the regression is reproducible across runs.
tests/test_pose_optimization_factors.cc:180
↓ 8 callersFunctionyaml_to_mat
Parse a {rows, cols, data: [...]} node into a CV_64F cv::Mat (returns empty on a missing/malformed node, matching the old FileStorage >> behaviour).
examples/stereo/euroc.cc:76
↓ 6 callersMethodadd
src/PlaceRecognizer.cc:21
↓ 6 callersMethodin_use
include/DescriptorPool.h:39
↓ 6 callersMethodinitialize
src/Logging.cpp:13
↓ 6 callersFunctionmakeFrame
Build a StereoFrame by projecting the world from camInWorld (Twc).
tests/test_vo_loop_closure.cc:74
↓ 6 callersFunctionmake_keyframe
Build a keyframe record observing `lms` (camera-frame points) from camera pose `camInWorld` relative to the candidate frame; here we just project from
tests/test_loop_closer.cc:69
↓ 6 callersMethodoptimize
src/WindowSmoother.cc:38
↓ 6 callersMethodpose_of
src/WindowSmoother.cc:120
↓ 6 callersMethodquery
tests/test_loop_closer.cc:53
↓ 6 callersFunctionshould_insert_keyframe
Covisibility-based keyframe gate (scale-invariant): insert a keyframe when the fraction of the reference keyframe's features tracked drops below covis
include/KeyframeGate.h:11
↓ 5 callersFunction_human
(n: float)
scripts/datasets/_common.py:32
↓ 5 callersMethodenable_loop_closure
src/VoEstimator.cc:59
↓ 5 callersMethodloop_closure_count
src/SuperSLAM.cc:187
↓ 5 callersFunctionmakeK
tests/test_vo_loop_closure.cc:61
↓ 5 callersFunctionparse_example_args
Parse `<settings.yaml> <sequence_dir> [--no-viewer|--viewer]`. Returns false (and prints usage) on malformed input.
examples/example_common.h:52
↓ 5 callersMethodshutdown
include/SuperSLAM.h:38
↓ 5 callersFunctionworld
Landmarks visible from every pose along the (small) loop.
tests/test_vo_loop_closure.cc:66
↓ 4 callersMethodbackproject
src/StereoFrame.cc:5
↓ 4 callersMethodcloud
src/SparseMap.cc:9
↓ 4 callersFunctiondiagSigmaUR
Extract the uR (disparity) sigma from a stereo_diag_noise model.
tests/test_pose_optimization_factors.cc:266
↓ 4 callersFunctiondrawMatches
tests/test_superpoint_cosine_matching.cc:77
↓ 4 callersFunctionenv_double
src/VoEstimator.cc:15
↓ 4 callersMethodlog
src/LightGlue.cc:16
↓ 4 callersFunctionmake
StereoFrame of `lms` as seen from camera at Twc `pose`.
tests/test_vo_convention.cc:40
↓ 4 callersFunctionmakeRec
tests/test_keyframe_database.cc:10
↓ 4 callersMethodoptimize_and_get_all
src/GlobalPoseGraph.cc:68
↓ 4 callersFunctionpace_to_timestamp
Sleep so a visualized run plays back near the dataset frame rate. `dt` is the inter-frame time (s); `track_ms` is the time already spent tracking this
examples/example_common.h:89
↓ 4 callersFunctionplaceDesc
One-hot global descriptor for "place" p.
tests/test_vo_loop_closure.cc:92
↓ 4 callersMethodprocess
src/RgbdFrontEnd.cc:23
↓ 4 callersFunctionreport_timing
examples/example_common.h:99
↓ 4 callersMethodtrack_stereo
src/SuperSLAM.cc:155
↓ 3 callersMethodadd_odometry
src/GlobalPoseGraph.cc:33
↓ 3 callersMethodclose
src/RerunViewer.cc:71
↓ 3 callersMethodcorrected_trajectory
src/VoEstimator.cc:181
↓ 3 callersFunctiondataset_dir
Target directory for one dataset, lowercased (e.g. ~/datasets/kitti).
scripts/datasets/_common.py:25
↓ 3 callersFunctiondownload
(names, tag: str | None = None, dest: Path | None = None)
scripts/models/_release.py:29
↓ 3 callersFunctionextract
Extract a .zip/.tar.*/.tgz archive into dest (idempotent-ish).
scripts/datasets/_common.py:105
↓ 3 callersFunctionfinite
src/GlobalPoseGraph.cc:16
↓ 3 callersMethodhas
include/GlobalPoseGraph.h:48
↓ 3 callersMethodinfer
Host path: run inference, copy all outputs to the host, CPU grid-sample in postprocess.
src/SuperPoint.cc:322
↓ 3 callersFunctionmakeK
tests/test_loop_closer.cc:62
↓ 3 callersMethodmatch
src/LightGlue.cc:285
↓ 3 callersFunctionodomNoise
tests/test_global_pose_graph.cc:13
↓ 3 callersFunctionpercentile
examples/stereo/benchmark.cc:37
↓ 3 callersMethodrelease
src/RerunViewer.cc:115
↓ 3 callersMethodsave_trajectory_kitti
src/SuperSLAM.cc:238
↓ 3 callersFunctionstereo_diag_noise
Diagonal stereo measurement noise over (uL, uR, v). uL and v keep the octave-scaled reprojection floor sigma_px (the SuperPoint and LightGlue matching
include/PoseOptimizationFactors.h:127
↓ 3 callersFunctionstore_floats
Store float values into a tensor's host buffer respecting its dtype (fp16 engines take fp16 inputs).
src/LightGlue.cc:231
↓ 2 callersMethodadd_loop
src/GlobalPoseGraph.cc:41
↓ 2 callersMethodanchors
src/VoEstimator.cc:175
↓ 2 callersFunctioncosineMatch
Match descriptors using cosine similarity
tests/test_superpoint_cosine_matching.cc:37
↓ 2 callersMethoddetect
src/LoopCloser.cc:105
↓ 2 callersFunctiondownload
Download url to dest with resume support; skips if already complete.
scripts/datasets/_common.py:40
↓ 2 callersFunctiondrawKeypoints
Draw keypoints on image
tests/test_superpoint_cosine_matching.cc:12
↓ 2 callersMethoddraw_frame
src/RerunViewer.cc:141
↓ 2 callersFunctionenv_double
src/LoopCloser.cc:11
↓ 2 callersFunctionfetch
Download (and optionally verify + extract) a single file into dest_dir.
scripts/datasets/_common.py:118
↓ 2 callersMethodin_window
src/WindowSmoother.cc:126
↓ 2 callersMethodkeyframe_count
include/SparseMap.h:19
↓ 2 callersFunctionload_list
examples/rgbd/tum.cc:42
↓ 2 callersFunctionmakeLandmarks
tests/test_loop_closer.cc:84
↓ 2 callersFunctionnormalizedRow
Return a 1xD CV_32F row, L2-normalized.
src/PlaceRecognizer.cc:10
↓ 2 callersMethodplot
src/RerunViewer.cc:89
↓ 2 callersFunctionpose_to_cv_tcw
src/SuperSLAM.cc:48
↓ 2 callersMethodsave_trajectory
src/SuperSLAM.cc:191
↓ 2 callersMethodstop_loop_worker
src/VoEstimator.cc:48
↓ 2 callersFunctionto_numpy
(t)
utils/convert_eigenplaces_to_onnx.py:21
↓ 2 callersFunctionverify
(path: Path, expected: str | None)
scripts/datasets/_common.py:94
↓ 2 callersMethodverify
src/LoopCloser.cc:44
↓ 1 callersFunctionTwoViewStereoBAIsFinite
Run a 2-KF stereo BA with KITTI-like geometry and ~noisePx measurement noise, anchoring KF0 with the given prior sigma. Returns true if every optimize
tests/test_pose_optimization_factors.cc:189
↓ 1 callersMethod__init__
(self, descriptor_dim: int = 256)
utils/convert_superpoint_to_onnx.py:33
↓ 1 callersFunction_cumulative_lengths
(xyz: np.ndarray)
scripts/benchmarks/_eval_common.py:75
↓ 1 callersFunction_last_frame_for_length
(dist: np.ndarray, first: int, length: float)
scripts/benchmarks/_eval_common.py:80
↓ 1 callersMethodadd
include/Profiling.h:24
↓ 1 callersFunctionapply_tuning_overrides
Bridge YAML tuning knobs to the SUPERSLAM_* env vars their components read. setenv with overwrite=0 keeps the precedence env, then YAML, then default.
src/SuperSLAM.cc:19
↓ 1 callersFunctionbackproject_cam
Stereo backprojection into the camera frame (identity pose). Same math as StereoFrame::backproject without the Twc lift; verification works in the can
src/LoopCloser.cc:19
↓ 1 callersFunctionbuild_rectify_maps
Build left/right rectification remap tables from LEFT.*/RIGHT.* in settings.
examples/stereo/euroc.cc:89
↓ 1 callersMethodcompute_global_descriptor
src/EigenPlaces.cc:145
↓ 1 callersFunctioncosineSimilarity
Compute cosine similarity between two descriptors
tests/test_superpoint_cosine_matching.cc:28
↓ 1 callersFunctiondatasets_root
Base directory for all datasets (~/datasets by default).
scripts/datasets/_common.py:19
↓ 1 callersMethoddescriptors_to_host
FP16 device slot to host CV_32F (rows L2-normalized by the gather kernel).
src/LightGlue.cc:460
↓ 1 callersFunctiondisp_sigma_px
Base disparity measurement precision (px). The SuperPoint and LightGlue stereo-match disparity floor is approx 8px (not subpixel).
include/PoseOptimizationFactors.h:106
↓ 1 callersFunctiondrawKeypoints
tests/test_superpoint_only.cc:11
↓ 1 callersFunctionelement_size
src/EigenPlaces.cc:25
↓ 1 callersMethodencode
Run the shared encoder and return the stride-8 feature map.
utils/convert_superpoint_to_onnx.py:51
↓ 1 callersMethodextract
superslam::IFeatureExtractor. Scores live in keypoint.response.
src/SuperPoint.cc:895
↓ 1 callersFunctionextract_sequences
(bundle_path, dest)
scripts/datasets/download_euroc.py:31
↓ 1 callersMethodextract_stereo
Batched stereo: one {2,1,H,W} infer for L and R.
src/SuperPoint.cc:902
↓ 1 callersMethodlast_loop_rejected
include/GlobalPoseGraph.h:50
↓ 1 callersFunctionload_images
examples/stereo/kitti.cc:44
next →1–100 of 291, ranked by callers