GithubHelp home page GithubHelp logo

Comments (7)

taihup avatar taihup commented on August 11, 2024

That means that the cameras are not rectified. At least, in the cameraInfo messages the rectification does not appear. If the cameras are rectified the intrinsic parameters inside the P matrix of cameraInfo messages should be the same.

from sptam.

breyner-posso avatar breyner-posso commented on August 11, 2024

Hi! first of all, thanks a lot for your help. I fixed the problem and now I'm running S-PTAM with my own dataset, but I'm getting this:

acfr@acfr-protea:~/catkin_ws$ roslaunch sptam acfr.launch
... logging to /home/acfr/.ros/log/d3f4d29a-48e7-11e8-a52a-b8ca3a8203b2/roslaunch-acfr-protea-5190.log
Checking log directory for disk usage. This may take awhile.
Press Ctrl-C to interrupt
Done checking log file disk usage. Usage is <1GB.

started roslaunch server http://acfr-protea:34699/

SUMMARY

PARAMETERS

  • /rosdistro: kinetic
  • /rosversion: 1.12.13
  • /sptam/BundleAdjustmentActiveKeyframes: 10
  • /sptam/DescriptorExtractor/Name: BRIEF
  • /sptam/DescriptorExtractor/bytes: 32
  • /sptam/DescriptorMatcher/Name: BruteForce-Hamming
  • /sptam/DescriptorMatcher/crossCheck: False
  • /sptam/EpipolarDistance: 0
  • /sptam/FeatureDetector/Name: GFTT
  • /sptam/FeatureDetector/minDistance: 15.0
  • /sptam/FeatureDetector/nfeatures: 2000
  • /sptam/FeatureDetector/qualityLevel: 0.01
  • /sptam/FeatureDetector/useHarrisDetector: False
  • /sptam/FrustumFarPlaneDist: 10000.0
  • /sptam/FrustumNearPlaneDist: 0.1
  • /sptam/LoopDetectorVocabulary: /home/acfr/catkin...
  • /sptam/MatchingCellSize: 15
  • /sptam/MatchingDistance: 25
  • /sptam/MatchingNeighborhood: 2
  • /sptam/approximate_sync: False
  • /sptam/base_frame: left_camera
  • /sptam/camera_frame: left_camera
  • /sptam/minimumTrackedPointsRatio: 0.9
  • /sptam/publish_on_fail: True
  • /sptam/publish_transform: True
  • /sptam/reference_frame: left_camera
  • /sptam/use_prediction: False
  • /use_sim_time: True

NODES
/stereo/
stereo_image_proc (stereo_image_proc/stereo_image_proc)
/
sptam (sptam/sptam_node)
sptam_path (ros_utils/pose_to_path)

ROS_MASTER_URI=http://localhost:11311

process[stereo/stereo_image_proc-1]: started with pid [5208]
process[sptam_path-2]: started with pid [5209]
process[sptam-3]: started with pid [5217]
[ INFO] [1524702130.917324400]: Initializing nodelet with 8 worker threads.
crossCheck: 0
OpenCV: threads set to -1
nfeatures: 2000
qualityLevel: 0.01
minDistance: 15
blockSize: 3
useHarrisDetector: 0
k: 0.04
nfeatures: 2000
qualityLevel: 0.01
minDistance: 15
blockSize: 3
useHarrisDetector: 0
k: 0.04
bytes: 32
use_orientation: 0
bytes: 32
use_orientation: 0
[ INFO] [1524702131.078567789]: Loop Detector initializing, loading vocabulary
[ INFO] [1524702139.255362110]: S-PTAM stereo node initialized.
[ INFO] [1524702148.291382720]: init calib
[ INFO] [1524702148.316342967]: baseline: 0.584967
[ INFO] [1524702148.937731023]: Map initialized with 34 points.
[ WARN] [1524702153.929021746]: Not enough points for tracking.
ERROR: There is not any covisibility keyframes observing the tracked points
[ WARN] [1524702169.995521925]: Not enough points for tracking.
ERROR: There is not any covisibility keyframes observing the tracked points
[ WARN] [1524702173.697949973]: Not enough points for tracking.
ERROR: There is not any covisibility keyframes observing the tracked points
[ WARN] [1524702181.509765843]: Not enough points for tracking.
ERROR: There is not any covisibility keyframes observing the tracked points
[ WARN] [1524702183.504044707]: Not enough points for tracking.
ERROR: There is not any covisibility keyframes observing the tracked points
[ WARN] [1524702192.387926727]: Not enough points for tracking.
ERROR: There is not any covisibility keyframes observing the tracked points
[ WARN] [1524702198.697397908]: Not enough points for tracking.
ERROR: There is not any covisibility keyframes observing the tracked points
[ WARN] [1524702199.844781791]: Not enough points for tracking.
ERROR: There is not any covisibility keyframes observing the tracked points
[ WARN] [1524702211.024708136]: Not enough points for tracking.
ERROR: There is not any covisibility keyframes observing the tracked points
[ WARN] [1524702217.886760256]: Not enough points for tracking.
ERROR: There is not any covisibility keyframes observing the tracked points
[ WARN] [1524702219.562904172]: Not enough points for tracking.
ERROR: There is not any covisibility keyframes observing the tracked points
[ WARN] [1524702228.745723087]: Not enough points for tracking.
ERROR: There is not any covisibility keyframes observing the tracked points
[ WARN] [1524702236.601386300]: Not enough points for tracking.
ERROR: There is not any covisibility keyframes observing the tracked points
[ WARN] [1524702239.989759320]: Not enough points for tracking.
ERROR: There is not any covisibility keyframes observing the tracked points
[ WARN] [1524702255.421210900]: Not enough points for tracking.

And at the end I didn't get any path.

What could I do?

Thaks a lot for all your support.

from sptam.

taihup avatar taihup commented on August 11, 2024

Which camera are you using? What is the camera baseline (distance between left and right camera)? Send me some sample images? Try to leave still the camera observing a high textured area. Check the issue #27, #24 and #16

from sptam.

breyner-posso avatar breyner-posso commented on August 11, 2024

Hi, tks a lot for your answer.

  1. The cameras I'm using are SF3322 - Sekonix (100° FOV) (1920x1208)
  2. Camera baseline is aprox. 58cm.
  3. These are some sample images:

Left and Right Images (raw)
image

Left and Right undistorted and rectified (stereo_before_refinement)
image

Left and Right undistorted and rectified (stereo_after_refinement)
image

Another scene: Left and Right undistorted and rectified (stereo_before_refinement)
image

I noticed looking at rviz that sptam publications are not smooth (as looking at raw), therefore each stereo frame is far from the previous one. I don't know if that is only visualization or if that sequence of frames is the one used for mapping.

I'm gonna collect mmore drives and test.

Tanks a lot for your advice!

from sptam.

taihup avatar taihup commented on August 11, 2024

Maybe S-PTAM is not running well. Depending on the configuration (features detector + descriptor, and other parameters), it will require a high computing power. Also, you can try to run the rosbag with a lower frame rate (rosbag play --clock --rate 0.5 rosbag_file.bag). What computer are you using? Check if the calibration is alright.

from sptam.

breyner-posso avatar breyner-posso commented on August 11, 2024

Hi, how can I check if the calibration is alright? When I see the horizontal lines on both images they look well rectified (aligned) but I don't know if that is enough.
My computer has a Corei7 Vpro processor and 8GB of RAM.
Tks a lot for your advice!

from sptam.

taihup avatar taihup commented on August 11, 2024

The calibration looks fine, but could be wrong. Try to calibrate the camera with ROS calibration package.

You can change epipolardistance parameter to 2. It will triangulate more points.

Try to run s-ptam as nodelet instead of node. It is prepare for that. It will run much faster.

Disable Loop Closure module.

from sptam.

Related Issues (20)

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.