Pi3DET-Visualization

December 22, 2025 · View on GitHub

Visualization toolkit for Pi3DET-Dataset. Hope you can enjoy it.

Getting Start

Data Preparation

Down the dataset from Huggingface. And then set your own path at DATA_PATH

The file-tree is like:

M3ED
├── processed
│   ├── Car
│   │   ├── Urban_Day
│   │   │   ├── car_urban_day_city_hall
│   │   │   ├── penno_big_loop
│   │   │   ├── rittenhouse
│   │   │   └── ucity_small_loop
│   │   └── Urban_Night
│   │       ├── car_urban_night_city_hall
│   │       ├── penno_big_loop
│   │       ├── rittenhouse
│   │       └── ucity_small_loop
│   ├── Falcon
│   │   ├── Outdoor_Day
│   │   │   ├── fast_flight_2
│   │   │   ├── penno_cars
│   │   │   ├── penno_parking_1
│   │   │   ├── penno_parking_2
│   │   │   ├── penno_plaza
│   │   │   └── penno_trees
│   │   └── Outdoor_Night
│   │       ├── high_beams
│   │       ├── penno_parking_1
│   │       └── penno_parking_2
│   └── Spot
│       ├── Outdoor_Day
│       │   ├── art_plaza_loop
│       │   ├── penno_short_loop
│       │   ├── rocky_steps
│       │   ├── skatepark_1
│       │   ├── skatepark_2
│       │   ├── srt_green_loop
│       │   ├── srt_under_bridge_1
│       │   └── srt_under_bridge_2
│       └── Outdoor_Night
│           ├── penno_plaza_lights
│           └── penno_short_loop

Visualization

Launch Engine: python tools/active_window.py
Select one Platform --> Terrian --> Sequence, and then click Start Main Window button
Load Dataset: Click the Load Dataset button
Visualization LiDAR: Click the Open 3D Viewer button
Check Tracking information: Select one category and the load track file.