Obstacle Segmentation Evaluation#
Preparation#
-
Copy sample scenario
mkdir -p ~/driving_log_replayer_data/obstacle_segmentation/sample cp -r ~/autoware/src/simulator/driving_log_replayer/sample/obstacle_segmentation/scenario.yaml ~/driving_log_replayer_data/obstacle_segmentation/sample
-
Copy bag file from dataset
mkdir -p ~/driving_log_replayer_data/obstacle_segmentation/sample/t4_dataset cp -r ~/driving_log_replayer_data/sample_dataset ~/driving_log_replayer_data/obstacle_segmentation/sample/t4_dataset
How to run#
-
Run the simulation
dlr simulation run -p obstacle_segmentation -l play_rate:=0.5
-
Check the results
Results are displayed in the terminal like below. The number of tests will vary slightly depending on PC performance and CPU load conditions, so slight differences are not a problem.
test case 1 / 1 : use case: sample_dataset -------------------------------------------------- TestResult: Failed Detection Failed: detection: 557 / 681 -> 81.79% detection_warn: 0 non_detection: 681 / 681 -> 100.00%