How to create thrust device_vector from opencv gpumat and vice-versa

asked 2018-10-14 12:43:43 -0500

shome gravatar image

updated 2018-10-14 12:48:51 -0500

I need to convert a opencv::gpumat to a thrust::device_vector, then perform a thrust::copy_if , then copy the resultant thrust::device_vector to a cv::gpumat again.

How to convert from cv::gpumat to thrust::device_vector & vice-versa? I am omitting the thrust::copy_if statements for clarity in my code below.

My code(below) gives compilation error: error: expected a type specifier

gcc (Ubuntu 5.4.0-6ubuntu1~16.04.10) 5.4.0 20160609

nvcc: Cuda compilation tools, release 8.0, V8.0.44

Thrust v1.8

ubuntu 16.04

My code:

#include <thrust/random.h>

#include <thrust/device_ptr.h>
#include <thrust/device_vector.h>

#include <opencv2/core/cuda.hpp>

int main(void)
{
    cv::Mat input_host(1,100,CV_32F);
    cv::Mat output_host(1,100,CV_32F);

    cv::randu(input_host, cv::Scalar(-10), cv::Scalar(10));
    cv::cuda::GpuMat input_device(1, 100, CV_32F);
    input_device.upload(input_host);

    thrust::device_ptr<CV_32F> input_ptr(input_device.ptr());

    thrust::device_vector<CV_32F>   input_device_vector  (input_ptr,100);
    return 0;
}
edit retag flag offensive close merge delete

Comments

I need to copy the non zero elements of a gpumat to a new gpumat with corresponding indices. I intend to use thrust iterators for the same.

I tried to use the gpuMat-thrust-interop sample code from opencv. But it gave several(46) compilation errors as pointed out in my other post compilation error with opencv gpu-thrust-interop sample code

shome gravatar imageshome ( 2018-10-14 13:16:43 -0500 )edit