Skip to content

Latest commit

 

History

History

helper_scripts

README.md

This folder contains helper scripts and Jupyter notebooks. They can be divided into 2 categories: CARLA-related and Thesis-related. Their usages are briefly described here.

CARLA-related:

  • explor_map.py: Use this script to manually wander around a map with a randomly spawned car. It is modified from a CARLA's example which is under the MIT license. The modification extends the visualization to also show the 4 lane boundaries in the neightborhood. This is intend to be a tool to find inconsistencies between the lane boundaries in the underlying map and the rendering. Run the script with the flag -h to see descriptions of available options.

  • show_all_landmarks.py: As its names suggests, it visualizes all landmark objects in CARLA's rendering. A landmark object corresponds to a traffic sign defined in the underlying map.

  • show_waypoints.py: Shows all waypoints in CARLA's rendering.

Thesis-related:

These scripts or notebooks are to generate images for thesis' use. When a manual setting of TEST_NAME, NOISE_LEVEL, or SW_CONFIG (SW stands for slidig window) is required, refer to scenarios.yaml for the configuration names. Such settings can be found under the comment title ############### Set directories manually ###############. The order of image labels (if required) should follow the order of configurations defined in scenarios.yaml.

  • gen_error_box_plots.py: Generates error box plots of a test.

  • gen_error_box_plots_null_test.py: Generates error box plots of null hypothesis test in the thesis. It is separated from the above one because the labelling logic is different.

  • gen_error_histo_plots.py: Generates error histograms of a test.

  • gen_error_plots.py: Generates color-coded error plots of a test.

  • gen_gt_plots.py: Generates ground truth path plots of a scenario recording.

  • gen_plot_for_sem_test.py: Generates a comparison plot of a semantic information test.

The following two Jupyter notebooks relies on the urban scenario to be recorded and saved first. They are ugly modified from the original detection simulation modules, but they get the job done.

  • visualize_lane_detection.ipynb: Generate visualizations of lane detection process.

  • visualize_pole_detection.ipynb: Generate visualizations of pole detection process.

Misc:

  • show_range_of_pxs.ipynb: This notebook is used to investigate the relationship between the pixels along the vertical center line in the image and their corresponding depth (x-coordinate in the front bumper frame).