mirror of
https://github.com/zebrajr/opencv.git
synced 2025-12-06 12:19:50 +01:00
ba19416730
### Pull Request Readiness Checklist resolves #16295 ``` docker run --gpus 0 -v ~/opencv:/opencv -v ~/opencv_contrib:/opencv_contrib -it nvidia/cuda:12.8.1-cudnn-devel-ubuntu22.04 apt-get update && apt-get install -y cmake python3-dev python3-pip python3-venv && python3 -m venv .venv && source .venv/bin/activate && pip install -U pip && pip install -U numpy && pip install torch --index-url https://download.pytorch.org/whl/cu128 && cmake \ -DWITH_OPENCL=OFF \ -DCMAKE_BUILD_TYPE=Release \ -DBUILD_DOCS=OFF \ -DWITH_CUDA=ON \ -DOPENCV_DNN_CUDA=ON \ -DOPENCV_EXTRA_MODULES_PATH=/opencv_contrib/modules \ -DBUILD_LIST=ts,cudev,python3 \ -S /opencv -B /opencv_build && cmake --build /opencv_build -j16 export PYTHONPATH=/opencv_build/lib/python3/:$PYTHONPATH ``` See details at https://github.com/opencv/opencv/wiki/How_to_contribute#making-a-good-pull-request - [x] I agree to contribute to the project under Apache 2 License. - [x] To the best of my knowledge, the proposed patch is not based on a code under GPL or another license that is incompatible with OpenCV - [x] The PR is proposed to the proper branch - [x] There is a reference to the original bug report and related work - [x] There is accuracy test, performance test and test data in opencv_extra repository, if applicable Patch to opencv_extra has the same branch name. - [x] The feature is well documented and sample code can be built with the project CMake
163 lines
6.6 KiB
Python
163 lines
6.6 KiB
Python
#!/usr/bin/env python
|
|
|
|
'''
|
|
CUDA-accelerated Computer Vision functions
|
|
'''
|
|
|
|
# Python 2/3 compatibility
|
|
from __future__ import print_function
|
|
|
|
import numpy as np
|
|
import cv2 as cv
|
|
import os
|
|
|
|
from tests_common import NewOpenCVTests, unittest
|
|
|
|
class cuda_test(NewOpenCVTests):
|
|
def setUp(self):
|
|
super(cuda_test, self).setUp()
|
|
if not cv.cuda.getCudaEnabledDeviceCount():
|
|
self.skipTest("No CUDA-capable device is detected")
|
|
|
|
def test_cuda_upload_download(self):
|
|
npMat = (np.random.random((128, 128, 3)) * 255).astype(np.uint8)
|
|
cuMat = cv.cuda_GpuMat()
|
|
cuMat.upload(npMat)
|
|
|
|
self.assertTrue(np.allclose(cuMat.download(), npMat))
|
|
|
|
def test_cuda_upload_download_stream(self):
|
|
stream = cv.cuda_Stream()
|
|
npMat = (np.random.random((128, 128, 3)) * 255).astype(np.uint8)
|
|
cuMat = cv.cuda_GpuMat(128,128, cv.CV_8UC3)
|
|
cuMat.upload(npMat, stream)
|
|
npMat2 = cuMat.download(stream=stream)
|
|
stream.waitForCompletion()
|
|
self.assertTrue(np.allclose(npMat2, npMat))
|
|
|
|
def test_cuda_interop(self):
|
|
npMat = (np.random.random((128, 128, 3)) * 255).astype(np.uint8)
|
|
cuMat = cv.cuda_GpuMat()
|
|
cuMat.upload(npMat)
|
|
self.assertTrue(cuMat.cudaPtr() != 0)
|
|
cuMatFromPtrSz = cv.cuda.createGpuMatFromCudaMemory(cuMat.size(),cuMat.type(),cuMat.cudaPtr(), cuMat.step)
|
|
self.assertTrue(cuMat.cudaPtr() == cuMatFromPtrSz.cudaPtr())
|
|
cuMatFromPtrRc = cv.cuda.createGpuMatFromCudaMemory(cuMat.size()[1],cuMat.size()[0],cuMat.type(),cuMat.cudaPtr(), cuMat.step)
|
|
self.assertTrue(cuMat.cudaPtr() == cuMatFromPtrRc.cudaPtr())
|
|
stream = cv.cuda_Stream()
|
|
self.assertTrue(stream.cudaPtr() != 0)
|
|
streamFromPtr = cv.cuda.wrapStream(stream.cudaPtr())
|
|
self.assertTrue(stream.cudaPtr() == streamFromPtr.cudaPtr())
|
|
asyncstream = cv.cuda_Stream(1) # cudaStreamNonBlocking
|
|
self.assertTrue(asyncstream.cudaPtr() != 0)
|
|
|
|
def test_cuda_buffer_pool(self):
|
|
cv.cuda.setBufferPoolUsage(True)
|
|
cv.cuda.setBufferPoolConfig(cv.cuda.getDevice(), 1024 * 1024 * 64, 2)
|
|
stream_a = cv.cuda.Stream()
|
|
pool_a = cv.cuda.BufferPool(stream_a)
|
|
cuMat = pool_a.getBuffer(1024, 1024, cv.CV_8UC3)
|
|
cv.cuda.setBufferPoolUsage(False)
|
|
self.assertEqual(cuMat.size(), (1024, 1024))
|
|
self.assertEqual(cuMat.type(), cv.CV_8UC3)
|
|
|
|
def test_cuda_release(self):
|
|
npMat = (np.random.random((128, 128, 3)) * 255).astype(np.uint8)
|
|
cuMat = cv.cuda_GpuMat()
|
|
cuMat.upload(npMat)
|
|
cuMat.release()
|
|
self.assertTrue(cuMat.cudaPtr() == 0)
|
|
self.assertTrue(cuMat.step == 0)
|
|
self.assertTrue(cuMat.size() == (0, 0))
|
|
|
|
@unittest.skip("failed test")
|
|
def test_cuda_convertTo(self):
|
|
# setup
|
|
npMat_8UC4 = (np.random.random((128, 128, 4)) * 255).astype(np.uint8)
|
|
npMat_32FC4 = npMat_8UC4.astype(np.single)
|
|
new_type = cv.CV_32FC4
|
|
|
|
# sync
|
|
# in/out
|
|
cuMat_8UC4 = cv.cuda_GpuMat(npMat_8UC4)
|
|
cuMat_32FC4 = cv.cuda_GpuMat(cuMat_8UC4.size(), new_type)
|
|
cuMat_32FC4_out = cuMat_8UC4.convertTo(new_type, cuMat_32FC4)
|
|
self.assertTrue(cuMat_32FC4.cudaPtr() == cuMat_32FC4_out.cudaPtr())
|
|
npMat_32FC4_out = cuMat_32FC4.download()
|
|
self.assertTrue(np.array_equal(npMat_32FC4, npMat_32FC4_out))
|
|
# out
|
|
cuMat_32FC4_out = cuMat_8UC4.convertTo(new_type)
|
|
npMat_32FC4_out = cuMat_32FC4.download()
|
|
self.assertTrue(np.array_equal(npMat_32FC4, npMat_32FC4_out))
|
|
|
|
# async
|
|
stream = cv.cuda.Stream()
|
|
cuMat_32FC4 = cv.cuda_GpuMat(cuMat_8UC4.size(), new_type)
|
|
cuMat_32FC4_out = cuMat_8UC4.convertTo(new_type, cuMat_32FC4)
|
|
# in/out
|
|
cuMat_32FC4_out = cuMat_8UC4.convertTo(new_type, 1, 0, stream, cuMat_32FC4)
|
|
self.assertTrue(cuMat_32FC4.cudaPtr() == cuMat_32FC4_out.cudaPtr())
|
|
npMat_32FC4_out = cuMat_32FC4.download(stream)
|
|
stream.waitForCompletion()
|
|
self.assertTrue(np.array_equal(npMat_32FC4, npMat_32FC4_out))
|
|
# out
|
|
cuMat_32FC4_out = cuMat_8UC4.convertTo(new_type, 1, 0, stream)
|
|
npMat_32FC4_out = cuMat_32FC4.download(stream)
|
|
stream.waitForCompletion()
|
|
self.assertTrue(np.array_equal(npMat_32FC4, npMat_32FC4_out))
|
|
|
|
@unittest.skip("failed test")
|
|
def test_cuda_copyTo(self):
|
|
# setup
|
|
npMat_8UC4 = (np.random.random((128, 128, 4)) * 255).astype(np.uint8)
|
|
|
|
# sync
|
|
# in/out
|
|
cuMat_8UC4 = cv.cuda_GpuMat(npMat_8UC4)
|
|
cuMat_8UC4_dst = cv.cuda_GpuMat(cuMat_8UC4.size(), cuMat_8UC4.type())
|
|
cuMat_8UC4_out = cuMat_8UC4.copyTo(cuMat_8UC4_dst)
|
|
self.assertTrue(cuMat_8UC4_out.cudaPtr() == cuMat_8UC4_dst.cudaPtr())
|
|
npMat_8UC4_out = cuMat_8UC4_out.download()
|
|
self.assertTrue(np.array_equal(npMat_8UC4, npMat_8UC4_out))
|
|
# out
|
|
cuMat_8UC4_out = cuMat_8UC4.copyTo()
|
|
npMat_8UC4_out = cuMat_8UC4_out.download()
|
|
self.assertTrue(np.array_equal(npMat_8UC4, npMat_8UC4_out))
|
|
|
|
# async
|
|
stream = cv.cuda.Stream()
|
|
# in/out
|
|
cuMat_8UC4 = cv.cuda_GpuMat(npMat_8UC4)
|
|
cuMat_8UC4_dst = cv.cuda_GpuMat(cuMat_8UC4.size(), cuMat_8UC4.type())
|
|
cuMat_8UC4_out = cuMat_8UC4.copyTo(cuMat_8UC4_dst, stream)
|
|
self.assertTrue(cuMat_8UC4_out.cudaPtr() == cuMat_8UC4_out.cudaPtr())
|
|
npMat_8UC4_out = cuMat_8UC4_dst.download(stream)
|
|
stream.waitForCompletion()
|
|
self.assertTrue(np.array_equal(npMat_8UC4, npMat_8UC4_out))
|
|
# out
|
|
cuMat_8UC4_out = cuMat_8UC4.copyTo(stream)
|
|
npMat_8UC4_out = cuMat_8UC4_out.download(stream)
|
|
stream.waitForCompletion()
|
|
self.assertTrue(np.array_equal(npMat_8UC4, npMat_8UC4_out))
|
|
|
|
def test_cuda_denoising(self):
|
|
self.assertEqual(True, hasattr(cv.cuda, 'fastNlMeansDenoising'))
|
|
self.assertEqual(True, hasattr(cv.cuda, 'fastNlMeansDenoisingColored'))
|
|
self.assertEqual(True, hasattr(cv.cuda, 'nonLocalMeans'))
|
|
|
|
def test_dlpack_GpuMat(self):
|
|
for dtype in [np.int8, np.uint8, np.int16, np.uint16, np.float16, np.int32, np.float32, np.float64]:
|
|
for channels in [2, 3, 5]:
|
|
ref = (np.random.random((64, 128, channels)) * 255).astype(dtype)
|
|
src = cv.cuda_GpuMat()
|
|
src.upload(ref)
|
|
dst = cv.cuda_GpuMat.from_dlpack(src)
|
|
test = dst.download()
|
|
equal = np.array_equal(ref, test)
|
|
if not equal:
|
|
print(f"Failed test with dtype {dtype} and {channels} channels")
|
|
self.assertTrue(equal)
|
|
|
|
if __name__ == '__main__':
|
|
NewOpenCVTests.bootstrap()
|