This is a package special for kitti data,

including tasks for exploring-data-analysis, sparse-to-dense estimation,

road segmentation and lane line detection.

to quickly get access to data, please use ./utils/data_provider.py or ./utils/velo_2_cam.py

• subfolder and files illustration

• kitti-tools
  ├─ data [source data]
  │ ├─ bin [lidar point clouds]
  │ ├─ calib [calibration files]
  │ ├─ img [RGB images]
  │ ├─ pcd [generated pcd point clouds]
  │ └─ readme.md
  ├─ dense_estimation [sparse-to-dense estimation]
  │ └─ points_estimation.py
  ├─ evaluation [general evalutation code]
  │ └─ data_similarity.py
  ├─ lane_detection [lane line detection]
  ├─ requirements.txt
  ├─ result [general results]
  ├─ road_segmentation [road segmentation]
  ├─ utils [general tools]
  │ ├─ canny.py
  │ ├─ config.py
  │ ├─ data_augmentation.py
  │ ├─ data_provider.py
  │ ├─ show_lidar.py
  │ ├─ velo_2_cam.py
  │ └─ velo_2_cam_origin.py
  

TODO:

• load data as .pcd format
• add pcl processing operation
• create folder and load image via opencv-python
• add projected lidar pixels to image
• fix modified projection code
• add lidar estimation code: reflectance
• complete diastance calculation for images
• add canny detecion for points
• complete data_augmentation code
• add road segmentation with RANSAC on point clouds