ktro2828 / ktro2828/perception_eval_rs
[TODO] add support of loading config from scenario
- Dominant language
- Rust
- Stars
- 1
- Forks
- 0
- PR merge metrics
- No merged PRs in 30d
Description
## What
Add support of loading `PerceptionEvaluationConfig` from scenario file in yaml format.
As a first step, we start from support of the scenario file like as below.
```yaml
ScenarioFormatVersion: 3.0.0
ScenarioName: sample_perception_evaluation
ScenarioDescription: sample evaluation for perception
SensorModel: sample_sensor_kit
VehicleModel: sample_vehicle
Evaluation:
UseCaseName: perception
UseCaseFormatVersion: 0.4.0
Datasets:
- tests/sample_data:
Version: annotation
VehicleId: default # Specify VehicleId for each data set.
LaunchSensing: false # Specifies whether the sensing module should be activated for each dataset. if false, use concatenated/pointcloud in bag
LocalMapPath: $HOME/autoware_map/sample-map-planning # Specify LocalMapPath for each data set.
Conditions:
PassRate: 99.0 # How much (%) of the evaluation attempts are considered successful.
PerceptionEvaluationConfig:
evaluation_config_dict:
evaluation_task: detection
frame_id: base_link
target_labels: [car, bicycle, pedestrian, motorbike]
max_x_position: 100.0
max_y_position: 100.0
min_point_number: 0
target_uuids: null
center_distance_threshold: 1.0
plane_distance_threshold: 2.0
iou_2d_threshold: 0.5
iou_3d_threshold: 0.5
```
This is a minimum implementation of scenario, and also a little bit different from original one.
Now I am working on this in https://github.com/ktro2828/perception_eval_rs/tree/feat/config-from-yaml.
## TODO list
- [x] load scenario from yaml.
- [x] instantiate `PerceptionEvaluationConfig` from loaded scenario.
- [ ] refactoring codes.
Contributor guide
No contributing guide indexed for this repository
Assessment
This issue has not been assessed yet.