Configuration Guide

To ensure proper setup, modifications in the yaml file are required. Here are the details on the relevant configurations:

pcd_path: This represents the path to your point cloud data. Ensure you've correctly specified the full path to your point cloud file or directory.
bag_path: This is the path to your data. Make sure to correctly provide the path to where your data is stored.
start_time: This is the starting time, which is used to get a better initial pose. Specifically, it represents the timestamp of the first shot when building the map. We will provide a series of start_time values for your reference.

Run Canny-EVT

To run Canny-EVT, follow the steps below (We take VECtor dataset as example):

Starting roscore:

roscore

Starting rosrun:

rosrun evit_demo run_VECtor /path/to/config.yaml

Terminating the program:

killall -9 run_VECtor

Ensure you have the necessary permissions and paths set up before executing the commands.

Mapping

Mapping is a refined version of the open-source project "Incremental 3D Line Segment Extraction from Semi-dense SLAM". Our enhancement includes the integration of 3D gradients. To view the related code modifications, please navigate to the mapping branch.

We also provide a script named ExtractBag.py that converts ROS bag files to the TUM format, extracts images, and generates an rgb.txt file.

To use the script, execute the following command:

python ExtractBag.py

Compilation and Execution

The procedure for compilation and execution aligns with the original source code (one can refer to the README in the mapping branch). After a successful run, you will produce a semi_pointcloud.obj file.

File Conversion

To transition the semi_pointcloud.obj file to a .pcd format, employ the following command with the provided script:

python3 readobj_pc_normal_flow.py (todo: might change the file name)

Testing

To improve testing of localization accuracy, we use result.txt provided by the mapping phase as ground truth (GT). The assessment method is as follows:

Place gt.txt and the output result.txt into the /scripts/for_evaluation directory.

Run the evaluation script with the following command:

python compare_scale.sh