I'm reprocessing an existing database recorded with an iPhone with SuperPoint features and rtabmap-reprocess, which consistently crashes at node 92255 for me.
What's peculiar about this node is that its depth photo is completely blank. There is an RGB photo but no depth. It's likely the first such node as well, so I speculate that must be the cause of the crash.
[...]
Processed 92253/114896 nodes [id=92253 map=63 opt_graph=74]... 195ms
Processed 92254/114896 nodes [id=92254 map=63 opt_graph=75]... 95ms
Processed 92255/114896 nodes [id=92255 map=63 opt_graph=76]... 54ms
terminate called after throwing an instance of 'c10::IndexError'
what(): Dimension out of range (expected to be in range of [-1, 0], but got 1)
Exception raised from maybe_wrap_dim_slow at /pytorch/c10/core/WrapDimMinimal.cpp:23 (most recent call first):
frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >) + 0x98 (0x7e3c165785e8 in /home/torjusti/.local/lib/python3.10/site-packages/torch/lib/libc10.so)
frame #1: long c10::detail::maybe_wrap_dim_slow<long>(long, long, bool) + 0x27b (0x7e3c1655de3b in /home/torjusti/.local/lib/python3.10/site-packages/torch/lib/libc10.so)
frame #2: at::native::transpose(at::Tensor const&, long, long) + 0x63 (0x7e3bfb18bef3 in /home/torjusti/.local/lib/python3.10/site-packages/torch/lib/libtorch_cpu.so)
frame #3: <unknown function> + 0x2b54483 (0x7e3bfbf54483 in /home/torjusti/.local/lib/python3.10/site-packages/torch/lib/libtorch_cpu.so)
frame #4: at::_ops::transpose_int::redispatch(c10::DispatchKeySet, at::Tensor const&, long, long) + 0x7d (0x7e3bfbab937d in /home/torjusti/.local/lib/python3.10/site-packages/torch/lib/libtorch_cpu.so)
frame #5: <unknown function> + 0x5045dea (0x7e3bfe445dea in /home/torjusti/.local/lib/python3.10/site-packages/torch/lib/libtorch_cpu.so)
frame #6: <unknown function> + 0x5046086 (0x7e3bfe446086 in /home/torjusti/.local/lib/python3.10/site-packages/torch/lib/libtorch_cpu.so)
frame #7: at::_ops::transpose_int::redispatch(c10::DispatchKeySet, at::Tensor const&, long, long) + 0x7d (0x7e3bfbab937d in /home/torjusti/.local/lib/python3.10/site-packages/torch/lib/libtorch_cpu.so)
frame #8: <unknown function> + 0x4abf97f (0x7e3bfdebf97f in /home/torjusti/.local/lib/python3.10/site-packages/torch/lib/libtorch_cpu.so)
frame #9: <unknown function> + 0x4abfe16 (0x7e3bfdebfe16 in /home/torjusti/.local/lib/python3.10/site-packages/torch/lib/libtorch_cpu.so)
frame #10: at::_ops::transpose_int::call(at::Tensor const&, long, long) + 0x196 (0x7e3bfbb2a056 in /home/torjusti/.local/lib/python3.10/site-packages/torch/lib/libtorch_cpu.so)
frame #11: rtabmap::SPDetector::compute(std::vector<cv::KeyPoint, std::allocator<cv::KeyPoint> > const&) + 0x920 (0x7e3c199ac800 in /usr/local/lib/librtabmap_core.so.0.21)
frame #12: rtabmap::SuperPointTorch::generateDescriptorsImpl(cv::Mat const&, std::vector<cv::KeyPoint, std::allocator<cv::KeyPoint> >&) const + 0x5c (0x7e3c195e61fc in /usr/local/lib/librtabmap_core.so.0.21)
frame #13: rtabmap::Feature2D::generateDescriptors(cv::Mat const&, std::vector<cv::KeyPoint, std::allocator<cv::KeyPoint> >&) const + 0x7b (0x7e3c195e79bb in /usr/local/lib/librtabmap_core.so.0.21)
frame #14: rtabmap::Memory::createSignature(rtabmap::SensorData const&, rtabmap::Transform const&, rtabmap::Statistics*) + 0xa358 (0x7e3c194b1498 in /usr/local/lib/librtabmap_core.so.0.21)
frame #15: rtabmap::Memory::update(rtabmap::SensorData const&, rtabmap::Transform const&, cv::Mat const&, std::vector<float, std::allocator<float> > const&, rtabmap::Statistics*) + 0x1dc (0x7e3c194b94bc in /usr/local/lib/librtabmap_core.so.0.21)
frame #16: rtabmap::Rtabmap::process(rtabmap::SensorData const&, rtabmap::Transform, cv::Mat const&, std::vector<float, std::allocator<float> > const&, std::map<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >, float, std::less<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > >, std::allocator<std::pair<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const, float> > > const&) + 0x1210 (0x7e3c1943d370 in /usr/local/lib/librtabmap_core.so.0.21)
frame #17: <unknown function> + 0xee3a (0x61bdbb0a2e3a in rtabmap-reprocess)
frame #18: <unknown function> + 0x29d90 (0x7e3c17a29d90 in /lib/x86_64-linux-gnu/libc.so.6)
frame #19: __libc_start_main + 0x80 (0x7e3c17a29e40 in /lib/x86_64-linux-gnu/libc.so.6)
frame #20: <unknown function> + 0x126e5 (0x61bdbb0a66e5 in rtabmap-reprocess)
--RGBD/ProximityBySpace true --Mem/STMSize 20 --Rtabmap/MemoryThr 2000 --Kp/MaxFeatures 500 --Kp/DetectorStrategy 11 --Vis/PnPVarianceMedianRatio 2 --Vis/DepthAsMask false --SuperPoint/ModelPath superpoint.pt --SuperGlue/Path rtabmap_superglue.py --Reg/RepeatOnce false --Vis/CorGuessWinSize 0 --Vis/FeatureType 11 --Mem/UseOdomFeatures false --Vis/CorNNType 6
I'm reprocessing an existing database recorded with an iPhone with SuperPoint features and rtabmap-reprocess, which consistently crashes at node 92255 for me.
What's peculiar about this node is that its depth photo is completely blank. There is an RGB photo but no depth. It's likely the first such node as well, so I speculate that must be the cause of the crash.
I have been using the following set of parameters: