opencv gpu cuda项目整理

opencv cuda编译：

Window编译Opencv CUDA_jacke121的专栏-CSDN博客_opencv编译cuda

// C++
net.setPreferableBackend(cv::dnn::DNN_BACKEND_CUDA);
net.setPreferableTarget(cv::dnn::DNN_TARGET_CUDA);

...

# Python
net.setPreferableBackend(cv.dnn.DNN_BACKEND_CUDA)
net.setPreferableTarget(cv.dnn.DNN_TARGET_CUDA)
or
net.setPreferableTarget(cv.dnn.DNN_TARGET_CUDA_FP16)

 

https://answers.opencv.org/question/201456/how-to-run-opencv-dnn-on-nvidia-gpu/

光流算法：

GPU based optical flow extraction in OpenCV

https://github.com/feichtenhofer/gpu_flow

不是gpu的光流算法：

https://github.com/gurkirt/optical-flow

gpu测试：

https://github.com/ariandyy/gpu_testing

多尺度检测：

Python OpenCV GPU Face Detection

视频稳像：

https://github.com/alexanderkoumis/opencv-gpu-py

cpu/gpu性能比较

https://github.com/rsteiger/opencv_surf_performance

 +=====================================================+
  | CPU                                                 |
  +=====================================================+
  | Total Feature Detection Time:              13.81 s  |
  | Average Feature Detection Time:           575.26 ms |
  | Total Descriptor Extraction Time:          23.05 s  |
  | Average Descriptor Extraction Time:       960.42 ms |
  | Total Calculation Time:                    36.86 s  |
  | Average Calculation Time:                1535.69 ms |
  +=====================================================+
  | GPU                                                 |
  +=====================================================+
  | Total Feature Detection Time:               2.12 s  |
  | Average Feature Detection Time:            88.38 ms |
  | Total Descriptor Extraction Time:           1.09 s  |
  | Average Descriptor Extraction Time:        45.23 ms |
  | Total Calculation Time (including bus):    19.03 s  |
  | Average Calculation Time:                 792.97 ms |
  +=====================================================+
  | Comparison                                          |
  +=====================================================+
  | GPU Feature Detection Speedup:              6.51 x  |
  | GPU Descriptor Extraction Speedup:         21.24 x  |
  | GPU Combined Speedup:                      11.49 x  |
  | GPU Overall Speedup:                        1.94 x  |
  +=====================================================+

实时对数极性滤波

https://github.com/tremalnaik/real_time_log_polar_filtering

OpenCV GPU Module Benchmark

https://github.com/Yougmark/opencv_gpu_benchmark

这个有makefile

 

基于OpenCV的图像边缘检测，用于测试使用gpu能多大程度上加快图像处理

https://github.com/SongWalker/testCanny

object-detect

https://github.com/bokiscout/object-detect

Image Processing on GPU using OpenCV

https://github.com/tejusp/Image-Processing-on-GPU-using-OpenCV