fabianschenk / reslam Goto Github PK

Hi dear experts
Honor to write a letter to you.
My question is what is the meaning of "python associate.py DATASET_XXX / rgb.txt DATASET_XXX / depth.txt > associate.txt" ?
I downloaded the associate.py, but can not find rgb.txt and depth.txt in the directory RESLAM.

thank you
Best regards
william

here the pdf is a wrong file when I download.

Hi,Fabian!
Thank you very much for sharing some related work on EDGE-SLAM. The running effect is very exciting. But I currently have a problem. The algorithm can run normally under the TUM dataset, but it cannot extract the depth information normally under the ICL dataset, that is to say, it cannot really run.
I have set InputDepthScaleFactor: 1000.0 and 5000.0,but it didn't work.
What is the problem?
Thank you very much!

Hi, Fabian,
Thank you very much for sharing this work. I have a small question, can RESLAM be run in real-time with the Orbbec Astra Pro camera? Because I didn't find a configuration file like orbbec_pro.yaml inside RESLAM. I look forward to your reply. thank you very much!

Hello, I'm following your job. I find a potential bug in the System.cpp file on line 311 and line 315 to 321.
Because we have considered the situation of TrackerTrackFromFrameToKf in mMapper.computePosesToTry() function , we needn't consider once more in line 315 to 321, right?
Thanks for you excellent work.

I have Segmentation fault when run:
build/RESLAM config_files/reslam_settings.yaml config_files/dataset_tum1.yaml

test_error.txt

yhzhao@yhzhao:~/test_git/RESLAM$ pkg-config --modversion eigen3
3.2.92

yhzhao@yhzhao:~/test_git/RESLAM$ build/RESLAM config_files/reslam_settings.yaml config_files/dataset_tum1.yaml
INF|System.cpp[readSettings,144]: EnableLocalMapping: 1
INF|System.cpp[readSettings,158]: SystemMultiThreading: 0
INF|System.cpp[readSettings,165]: Read: [1, 1, 1.25, 1.5] into 1 1 1.25 1.5
INF|System.cpp[readSensorSettings,90]: InputDatasetFile1: associate.txt
INF|System.cpp[readSensorSettings,92]: InputDatasetFile2: associate.txt
INF|System.cpp[readSensorSettings,106]: InputSkipFirstNFrames: 0
INF|System.cpp[readSensorSettings,112]: InputDepthMin: 0.1 InputDepthMax: 5.2
INF|LocalMapper.cpp[LocalMapper,55]: setThreadReduce: 0xb2c100
INF|Tracker.cpp[Tracker,42]: mCameraMatrix: 517.306 516.469
INF|fernrelocaliser.cpp[FernRelocaliser,297]: imgSize / (1 << levels): [40 x 30] height: 480 width: 640
INF|FernConservatory.cpp[FernConservatory,47]: Generating FernConservatory with: 500 4 [40 x 30] [0, 255] 4 width: 40x h:30
INF|fernrelocaliser.cpp[FernRelocaliser,300]: Creating RelocDatabase!
INF|Tracker.cpp[Tracker,42]: mCameraMatrix: 517.306 516.469
INF|Input.cpp[readGTTrajectory,50]: Reading: /home/yhzhao/dataSets/vslam/tum/rgbd_dataset_freiburg1_desk2/groundtruth.txt
INF|Input.cpp[InputDatasetReader,104]: Reading: /home/yhzhao/dataSets/vslam/tum/rgbd_dataset_freiburg1_desk2/associate.txt
START PANGOLIN!
Input.cpp[readImages,192]: readImages loop!
INF|Input.cpp[getImages,128]: getimages!0
INF|Input.cpp[getImages,132]: mNbReceivedImages!1
INF|System.cpp[startSystem,229]: input is still active!

INF|ceresloopcloser.cpp[BuildOptimizationProblem,97]: Fixing: 2 flagFixPose: 1
INF|ceresloopcloser.cpp[BuildOptimizationProblem,94]: p: 3
INF|ceresloopcloser.cpp[BuildOptimizationProblem,94]: p: 4
INF|ceresloopcloser.cpp[BuildOptimizationProblem,94]: p: 5
INF|ceresloopcloser.cpp[BuildOptimizationProblem,94]: p: 6
INF|ceresloopcloser.cpp[BuildOptimizationProblem,94]: p: 7
INF|ceresloopcloser.cpp[BuildOptimizationProblem,97]: Fixing: 7 flagFixPose: 1
INF|ceresloopcloser.cpp[BuildOptimizationProblem,94]: p: 8
INF|ceresloopcloser.cpp[BuildOptimizationProblem,94]: p: 9
INF|ceresloopcloser.cpp[BuildOptimizationProblem,94]: p: 10
INF|ceresloopcloser.cpp[BuildOptimizationProblem,94]: p: 11
INF|ceresloopcloser.cpp[BuildOptimizationProblem,94]: p: 12
INF|ceresloopcloser.cpp[BuildOptimizationProblem,97]: Fixing: 12 flagFixPose: 1
INF|ceresloopcloser.cpp[BuildOptimizationProblem,94]: p: 13
INF|ceresloopcloser.cpp[BuildOptimizationProblem,94]: p: 14
INF|ceresloopcloser.cpp[BuildOptimizationProblem,97]: Fixing: 14 flagFixPose: 1
INF|ceresloopcloser.cpp[BuildOptimizationProblem,94]: p: 15
INF|ceresloopcloser.cpp[BuildOptimizationProblem,97]: Fixing: 15 flagFixPose: 1
INF|ceresloopcloser.cpp[BuildOptimizationProblem,94]: p: 16
INF|ceresloopcloser.cpp[BuildOptimizationProblem,97]: Fixing: 16 flagFixPose: 1
INF|ceresloopcloser.cpp[BuildOptimizationProblem,103]: nFixedPoses: 7
Segmentation fault (core dumped)

I have errors while running the command 'make -j', the error messages are listed below.

Error messages

Error 1

~/RESLAM/System/System.cpp:97:71: error: no matching function for call to ‘read(cv::FileNode, std::__cxx11::string&, const char [1])’ cv::read(fs["InputDatasetConfigFileName"],DatasetConfigFileName,"");

Error 2

~/RESLAM/System/System.cpp:97:47: error: cannot bind non-const lvalue reference of type ‘cv::String&’ to an rvalue of type ‘cv::String’ cv::read(fs["InputDatasetConfigFileName"],DatasetConfigFileName,"");

Opinions

As for the error 2, someone on the internet said that adding a 'const' before the parameter in the function definition and declaration would fixed the error. I tried it ,modifying the opencv source code, and more errors came out. Obviously, when error 2 was fixed, the error 1 would disappear. But I have no idea how to fixed it.
The opencv version on my ubuntu 18.04 is 3.2.0.

Thank you very much for your open source code, which has greatly inspired me. When I use the default setting file dataset_ Icl.yaml and reslam_ Settings.yaml tests the two sequences lr-kt0 and lr-kt1 of ICL, and the experimental results are as follows:
ATE of VO+LM
lr-kt0: 4.5cm lr-kt1:1.8cm

ATE of VO+LM+LC
lr-kt0: 7.5cm lr-kt1:1.8cm

However, the experimental results in your paper are
ATE of VO+LM
lr-kt0: 3.2cm lr-kt1:1.8cm

ATE of VO+LM+LC
lr-kt0: 1.9cm lr-kt1:1.7cm

Do I need to change some parameters of the system to achieve the same results as in the paper?

I am deeply attracted by your work，and thank you very much for your open source. But when I test your code , I encounter a program error “Segmentation fault (core dumped)”

Hello, thanks for your contribution for this edge-based slam. It's really a good innovation to use such a kind of feature in slam.
I have tested it on openloris market dataset. From the result, it seems that RESLAM outputs a wrong scale. But as I have known, depth data could help recover the scale. Maybe I have set a wrong config(Maybe there's is a parameter for this), or RESLAM really has the scale problem? Thank you, really hope to get your response!

Best regards,
Slamer