Recently, I want to compare my own result of HILTI 2022 and 2023 dataset and the official result in FAST-LIVO2. I can get the result from FAST-LIVO2 but the result often have large difference comparing to the ground truth. I have the following problems:
- What is the proper way for evaluation for those datasets?
- What are the corresponding launch files for 2022 and 2023 (handheld and robot) data?
- Do I need to adopt the commented
extrinsic_T and extrinsic_R for 2023 data in /condig/HILTI22.yaml?
When running HILTI 2022 dataset, I use the mapping_hesaixt32_hilti22.launch in /launch with map_sliding_en: true in /condig/HILTI22.yaml. This is the only chane in the config file. But the result is largely different from the ground truth.
Thanks!
Recently, I want to compare my own result of HILTI 2022 and 2023 dataset and the official result in FAST-LIVO2. I can get the result from FAST-LIVO2 but the result often have large difference comparing to the ground truth. I have the following problems:
extrinsic_Tandextrinsic_Rfor 2023 data in/condig/HILTI22.yaml?When running HILTI 2022 dataset, I use the
mapping_hesaixt32_hilti22.launchin/launchwithmap_sliding_en: truein/condig/HILTI22.yaml. This is the only chane in the config file. But the result is largely different from the ground truth.Thanks!