cbeall3
|
29ad9478f7
|
Move noise model initialization from add function into constructor. Can only have one noise model per factor anyway
|
2015-08-26 13:02:39 -04:00 |
Luca
|
cd6c5ca0bd
|
using parameters in smart projection factors constructors.. breaking the API, but now is way more elegant
|
2015-06-19 18:09:39 -04:00 |
Luca
|
78c8160dc5
|
all tests pass and it compiles (yuppii!), but if I make check I get errors with isManifold and something that seems unrelated to smart factors. going to merge with develop
|
2015-06-19 12:06:45 -04:00 |
Luca
|
fb7bc12c84
|
most unit tests fixed - 2 to go, now sensor_T_body is in the base class, probably not the best choice
|
2015-06-19 11:42:51 -04:00 |
Luca
|
756d1d29b7
|
fixing key unit tests - still failures in the optimization
|
2015-06-18 17:23:39 -04:00 |
Luca
|
66083f5e18
|
included calibration in constructor SmartProjectionPoseFactor
|
2015-06-18 16:33:17 -04:00 |
Luca
|
391386a654
|
revived SmartProjectionPoseFactor. Compiles now, going to fix unit tests now
|
2015-06-18 16:00:54 -04:00 |
dellaert
|
3675754816
|
Fixed two more tests: down to two !
|
2015-03-12 10:00:29 -07:00 |
dellaert
|
695d080e4d
|
DegeneracyMode
|
2015-03-12 07:56:59 -07:00 |
dellaert
|
3558da9c4c
|
Fixed SVD factor
|
2015-03-11 07:04:02 -07:00 |
dellaert
|
62868b9071
|
Added linear error checks
|
2015-03-11 06:50:15 -07:00 |
dellaert
|
9991240ae7
|
Removed selective linearization mess - it was non-functional anyway, we can aleays re-add it if needed.
|
2015-03-10 15:24:00 -07:00 |
dellaert
|
a0470b7e1c
|
Moved all linearize logic into SmartProjectionFactor and removed two subclasses
|
2015-03-10 12:44:19 -07:00 |
dellaert
|
1eec6748cb
|
Removed timing, added new SUMMARY stats on failing tests
|
2015-03-09 18:00:47 -07:00 |
dellaert
|
0ab1b8631a
|
Fixed compilation issues due to renaming
|
2015-03-09 17:43:00 -07:00 |
dellaert
|
daf16acdfa
|
Get rid of debugging fluff and more copy/paste
|
2015-03-04 23:56:18 -08:00 |
dellaert
|
7f27a594d2
|
Fixed degeernate test-case
|
2015-03-04 23:45:27 -08:00 |
dellaert
|
5dda36a4d6
|
Fixed 3 more test cases
|
2015-03-04 23:38:55 -08:00 |
dellaert
|
3b144a9cab
|
Fixed priors
|
2015-03-04 23:34:43 -08:00 |
dellaert
|
485fabeae6
|
Fixed another test case
|
2015-03-04 23:04:32 -08:00 |
dellaert
|
26f2b33c47
|
Migrated to non-keyed Fblocks
|
2015-03-04 23:01:47 -08:00 |
dellaert
|
23cf0a49f5
|
Fixed another test case
|
2015-03-04 22:53:17 -08:00 |
dellaert
|
a95e5c7e05
|
Fixed another test
|
2015-03-04 17:21:11 -08:00 |
dellaert
|
48d8de50d0
|
Fixed SVD unit test
|
2015-03-04 16:38:02 -08:00 |
dellaert
|
f28b083c14
|
Merge remote-tracking branch 'origin/develop' into feature/SmartFactors3
Conflicts:
gtsam/slam/tests/testSmartProjectionPoseFactor.cpp
|
2015-03-04 16:02:25 -08:00 |
dellaert
|
3ab4c329e8
|
Check explicit poses rather than printing them
|
2015-03-04 12:10:43 -08:00 |
dellaert
|
cc4083d33e
|
Fixed global replace error
|
2015-03-04 10:28:26 -08:00 |
dellaert
|
216b5ae62b
|
Fixed a large number of issues in test with switch to PinholePose
|
2015-03-04 10:21:58 -08:00 |
dellaert
|
6fc208b8d2
|
Comment out linear solve (throws)
|
2015-03-03 22:27:23 -08:00 |
dellaert
|
8e818a4cfd
|
Fixed types in debug mode
|
2015-03-03 22:24:48 -08:00 |
dellaert
|
ca18c13cd2
|
Most factors now work out for sigma=1 and sigma=0.1
|
2015-03-02 08:53:07 -08:00 |
dellaert
|
cee64e3853
|
Made sure all refer to the same information matrix
|
2015-03-01 14:28:31 +01:00 |
dellaert
|
6a024259e5
|
Switched to (i,j) instead of (i1,i2)
|
2015-03-01 14:28:13 +01:00 |
dellaert
|
2cc0223624
|
Made tests compile with TriangulationResult changes
|
2015-03-01 12:06:46 +01:00 |
dellaert
|
a375e7b5be
|
RADICAL2: The SmartProjectionPoseFactor (soon to be renamed SmartPinholePoseFactor, if it survives at all) now no longer stores shared calibrations. Values expect to contain PinholePoses not Pose3s now. The current state of affairs was simply a bug: one pose could be optimized for several different calibrations. It relied on the user to make sure all measurements for a specific pose to optimize were all given the same shared calibration, which was then stored *millions of times* in the pose factors. Instead, there is now *one* shared calibration per PinholePose unknown.
|
2015-02-26 13:55:16 +01:00 |
dellaert
|
a132d959f5
|
RADICAL: Got rid of sensor_pose. In the new PinholePose philosophy, this is entirely the responsibility of the CAMERA. Just like PinholePose, we can have a camera class that has a shared extra transform: it is part of the calibration. PinholePose itself is not able to do this, as the calibration is assumed 2D only, but it's easy to create a class and have the correct derivatives in place.
|
2015-02-26 12:06:43 +01:00 |
dellaert
|
fd71974ff3
|
Got mostly rid of computeEP: just a call to cameras.project2
|
2015-02-25 20:52:16 +01:00 |
dellaert
|
850470ef06
|
rename of computeJacobians overloads to better reflect functionality
|
2015-02-25 19:30:17 +01:00 |
dellaert
|
eb28d0ffa8
|
Restored reprojectionErrors -> reprojectionError
|
2015-02-24 14:09:35 +01:00 |
dellaert
|
ae1f534e66
|
Now second test uses common header as well.
|
2015-02-23 23:58:25 +01:00 |
dellaert
|
d6f54475c3
|
BIG change: SmartFactorBase and SmartProjectionFactor now templated on CAMERA
|
2015-02-23 12:43:43 +01:00 |
dellaert
|
cdaaee6fce
|
Move Errors to CameraSet
|
2015-02-22 22:05:42 +01:00 |
dellaert
|
241366a6e6
|
Check augmented information matrices
|
2015-02-22 17:30:08 +01:00 |
dellaert
|
772db8850a
|
SVD and HessianFactor checked using same information
|
2015-02-22 07:05:11 +01:00 |
dellaert
|
d09c7aa105
|
Test createJacobianSVDFactor
|
2015-02-22 06:48:34 +01:00 |
dellaert
|
8b914189cb
|
Check JacobianFactorQ
|
2015-02-22 06:33:05 +01:00 |
dellaert
|
b0974d6f41
|
Remove duplicate test
|
2015-02-22 06:21:09 +01:00 |
dellaert
|
64bb6b77d7
|
Merged in feature/SmartCT (pull request #107)
Refactoring of Smart Factors
|
2015-02-22 06:14:19 +01:00 |
dellaert
|
3ccd02e8b1
|
Two factors unit tested
|
2015-02-21 17:09:43 +01:00 |
dellaert
|
96e63fb480
|
Added test for RegularHessianFactor creation
|
2015-02-21 14:32:36 +01:00 |