The aim of the project is to create a powerful desktop application to make debugging of SLAM frontend algorithms easier.
This tool will provide following features:
- Visualization of landmarks of three types (points, lines, planes)
- Mapping of the same landmarks between different images based on SLAM data association
- Editing and fine-tuning landmarks positions and association
Follow instructions in the User Guide
- Create the
Images
folder, in which put all the frames from the dataset. Images must match the following criteria:- same format for all images
- jpg or png extension
- file name is timestamp in Long format
- For each used landmark detection algorithm create a separate folder. You need to name it in the
format
name_algorithm_landmark
(example:alg1_line
,alg1_keypoint
,alg1_plane
) - Put files with markup in each of the folders. They should be named according to the frame they belong to (timestamp)
- Points are represented in .csv file with (uid,x,y) columns, where uid is id of landmark and x,y are coordinates of
keypoint.
Example
- Lines are represented in .csv file with (uid,x0,y0,x1,y1) columns, where uid is id of landmark and x0,y0,x1,y1 are coordinates of the
beginning and end of the
line.
Example
- The planes are represented as a png file, where
the points belonging to the planes are shown in different
pixel colors. For points that do not belong
to any plane, black is
used.
Example
- Points are represented in .csv file with (uid,x,y) columns, where uid is id of landmark and x,y are coordinates of
keypoint.
Ensure you have JDK 17 to build or run this project.
Build: ./gradlew assemble
Run: ./gradlew run