Skip to content

Obstacle Segmentation Evaluation#

Preparation#

  1. 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

  2. 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#

  1. Run the simulation

    dlr simulation run -p obstacle_segmentation  -l play_rate:=0.5

    obstacle_segmentation

  2. 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%