Open yuchenweng-private opened 2 years ago
Could you elaborate what data were you running ESVO on? Have you obtained normally good Time Surfaces?
I run ESVO on VECtor Benchmark(https://star-datasets.github.io/vector/), I have tested time surface node, it works well.
Then you may need to tune the depth perception range according to the dataset used. Please refer to this issue for further detail. Hopefully it would help.
When I try to test esvo on other dataset, it always shows that initialization failed once, and then whole system will reset. The terminal always shows like: I0730 15:50:41.684104 17011 esvo_Mapping.cpp:481] ** Initialization (SGM) returns 0 points. I0730 15:50:41.684176 17011 esvo_Mapping.cpp:243] Initialization fails once. I have checked my calibration.