I am trying to implement cv::cuda::warpPerspective in python2, there is a very sweet post about how to do that here: link. I followed the instruction as described in that post, however, I got Segmentation fault (core dumped) error.
I was able to allocate the error in GpuWrapper.pyx file line 11:
pyopencv_to(<PyObject*> _src, src_mat) It seems that it fails to convert numpy array to opencv Mat.
I am not sure where is wrong and how to fix it.
The python script that got Segmentation fault (core dumped) error is in below:
import cv2 import numpy as np import csv import timeit import GpuWrapper while (1): start = timeit.default_timer() eo_img = cv2.imread('./sample/eo.png', 1) nir_img = cv2.imread('./sample/nir.png', 0) with open('./sample/reg.csv', 'rb') as f: line = csv.reader(f) reg_line = list(line) reg = np.array(reg_line[0], dtype=np.float32) new_reg = reg.reshape((3,3)) print nir_img.shape dist = GpuWrapper.cudaWarpPerspectiveWrapper(nir_img, new_reg, (2448,2048)) cv2.imwrite('./sample/result.png', dist) end = timeit.default_timer() print end-start Another thing to mention, when compiling cython code, it has some warnings:
[1/1] Cythonizing GpuWrapper.pyx running build_ext building 'GpuWrapper' extension x86_64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -fno-strict-aliasing -Wdate-time -D_FORTIFY_SOURCE=2 -g -fstack-protector-strong -Wformat -Werror=format-security -fPIC -I/home/xinyao/Tensorflow/local/lib/python2.7/site-packages/numpy/core/include -I-I/usr/local/include/opencv -I-I/usr/local/include -I/usr/include/python2.7 -c GpuWrapper.cpp -o build/temp.linux-x86_64-2.7/GpuWrapper.o cc1plus: warning: command line option ‘-Wstrict-prototypes’ is valid for C/ObjC but not for C++ In file included from /home/xinyao/Tensorflow/local/lib/python2.7/site-packages/numpy/core/include/numpy/ndarraytypes.h:1809:0, from /home/xinyao/Tensorflow/local/lib/python2.7/site-packages/numpy/core/include/numpy/ndarrayobject.h:18, from pyopencv_converter.cpp:2, from GpuWrapper.cpp:544: /home/xinyao/Tensorflow/local/lib/python2.7/site-packages/numpy/core/include/numpy/npy_1_7_deprecated_api.h:15:2: warning: #warning "Using deprecated NumPy API, disable it by " "#defining NPY_NO_DEPRECATED_API NPY_1_7_API_VERSION" [-Wcpp] #warning "Using deprecated NumPy API, disable it by " \ ^ c++ -pthread -shared -Wl,-O1 -Wl,-Bsymbolic-functions -Wl,-Bsymbolic-functions -Wl,-z,relro -fno-strict-aliasing -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -Wdate-time -D_FORTIFY_SOURCE=2 -g -fstack-protector-strong -Wformat -Werror=format-security -Wl,-Bsymbolic-functions -Wl,-z,relro -Wdate-time -D_FORTIFY_SOURCE=2 -g -fstack-protector-strong -Wformat -Werror=format-security build/temp.linux-x86_64-2.7/GpuWrapper.o -o /home/xinyao/projects/image_reg/GpuWrapper.so -L/usr/local/lib -lopencv_cudabgsegm -lopencv_cudaobjdetect -lopencv_cudastereo -lopencv_dnn -lopencv_ml -lopencv_shape -lopencv_stitching -lopencv_cudafeatures2d -lopencv_superres -lopencv_cudacodec -lopencv_videostab -lopencv_cudaoptflow -lopencv_cudalegacy -lopencv_calib3d -lopencv_features2d -lopencv_highgui -lopencv_videoio -lopencv_photo -lopencv_imgcodecs -lopencv_cudawarping -lopencv_cudaimgproc -lopencv_cudafilters -lopencv_video -lopencv_objdetect -lopencv_imgproc -lopencv_flann -lopencv_cudaarithm -lopencv_viz -lopencv_core -lopencv_cudev Which seems to be fine according to here.
Environment: CUDA 8.0 OpenCV 3.3 without extra modules Ubuntu 16.04 Python 2.7 Cython 0.26
I have tried cv::cuda::warpPerspective in C++ and it works fine with GPU support. I have also tried cv2.warpPerspective in python and without any issue(so my opencv for python was compiled correctly).
Another thing is I cannot compile opencv C++ code directly with g++, I need to add extra flags to do it, like in this way: g++ -o main main.cc `pkg-config opencv --cflags --libs Otherwise I will have issues on finding opencv:
main.cc:(.text+0x45): undefined reference to `cv::cuda::GpuMat::defaultAllocator()' main.cc:(.text+0x9b): undefined reference to `cv::imread(cv::String const&, int)' main.cc:(.text+0x104): undefined reference to `cv::imread(cv::String const&, int)' main.cc:(.text+0x140): undefined reference to `cv::cuda::GpuMat::defaultAllocator()' main.cc:(.text+0x1a5): undefined reference to `cv::Mat::eye(int, int, int)' main.cc:(.text+0x31f): undefined reference to `cv::cuda::Stream::Null()' main.cc:(.text+0x3d6): undefined reference to `cv::cuda::warpPerspective(cv::_InputArray const&, cv::_OutputArray const&, cv::_InputArray const&, cv::Size_<int>, int, int, cv::Scalar_<double>, cv::cuda::Stream&)' main.cc:(.text+0x487): undefined reference to `cv::imwrite(cv::String const&, cv::_InputArray const&, std::vector<int, std::allocator<int> > const&)' Not sure if that is a factor, but I am not able to fix it as well.
Any help will be really appreciated.
UPDATE: I was able to allocate the bug by using pdb:
>$gdb --args python opencv_reg_cuda.py >$run Starting program: /usr/bin/python opencv_reg_cuda.py [Thread debugging using libthread_db enabled] Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1". [New Thread 0x7fffc6b83700 (LWP 10933)] [New Thread 0x7fffc4382700 (LWP 10934)] [New Thread 0x7fffc1b81700 (LWP 10935)] [New Thread 0x7fffc1380700 (LWP 10936)] [New Thread 0x7fffbcb7f700 (LWP 10937)] [New Thread 0x7fffba37e700 (LWP 10938)] [New Thread 0x7fffb7b7d700 (LWP 10939)] (2048, 2448) Thread 1 "python" received signal SIGSEGV, Segmentation fault. pyopencv_to (o=o@entry=0x7fffb37e3b70, m=..., info=...) at pyopencv_converter.cpp:210 210 if( !PyArray_Check(o) ) >$bt #0 pyopencv_to (o=o@entry=0x7fffb37e3b70, m=..., info=...) at pyopencv_converter.cpp:210 #1 0x00007fffb33c56f5 in pyopencv_to<cv::Mat> (name=0x7fffb33c98d8 "<unknown>", m=..., o=0x7fffb37e3b70) at pyopencv_converter.cpp:349 #2 __pyx_pf_10GpuWrapper_cudaWarpPerspectiveWrapper (__pyx_v(short, long double, char)=__pyx_v(short, long double, char)@entry=0x7fffb37e3b70, __pyx_v__M=__pyx_v__M@entry=0x7fffb37e3c60, __pyx_v__size_tuple=__pyx_v__size_tuple@entry=0x7ffff7e80b90, __pyx_v__flags=<optimized out>, __pyx_self=<optimized out>) at GpuWrapper.cpp:1796 #3 0x00007fffb33c76a2 in __pyx_pw_10GpuWrapper_1cudaWarpPerspectiveWrapper (__pyx_self=<optimized out>, __pyx_args=<optimized out>, __pyx_kwds=<optimized out>) at GpuWrapper.cpp:1737 #4 0x00000000004c468a in PyEval_EvalFrameEx () #5 0x00000000004c2765 in PyEval_EvalCodeEx () #6 0x00000000004c2509 in PyEval_EvalCode () #7 0x00000000004f1def in ?? () #8 0x00000000004ec652 in PyRun_FileExFlags () #9 0x00000000004eae31 in PyRun_SimpleFileExFlags () #10 0x000000000049e14a in Py_Main () #11 0x00007ffff7810830 in __libc_start_main (main=0x49dab0 <main>, argc=2, argv=0x7fffffffde38, init=<optimized out>, fini=<optimized out>, rtld_fini=<optimized out>, stack_end=0x7fffffffde28) at ../csu/libc-start.c:291 #12 0x000000000049d9d9 in _start () >list 205 } 206 } 207 return true; 208 } 209 210 if( !PyArray_Check(o) ) 211 { 212 failmsg("%s is not a numpy array, neither a scalar", info.name); 213 return false; 214 } It seems like it cannot pass o to PyArray_Check, but I have no idea what was going on and how to fix it.