CODEWITHSUNDEEP
c++opencv
Dzung Nguyen
Dzung Nguyen

Reputation: 3944

Compiling error cv::gpu

I'm using OpenCV master branch (3.0.0. dev) with CUDA on Ubuntu 12.04, and trying to compile the following opencv with gpu code:

#include <iostream>
#include "opencv2/opencv.hpp"
#include "opencv2/core.hpp"
#include "opencv2/highgui.hpp"
#include "opencv2/gpu/gpu.hpp"

using namespace cv;

int main (int argc, char* argv[])
{
    try
    {
        cv::Mat src_host = cv::imread("file.png", CV_LOAD_IMAGE_GRAYSCALE);
        cv::gpu::GpuMat dst, src;
        src.upload(src_host);

        cv::gpu::threshold(src, dst, 128.0, 255.0, CV_THRESH_BINARY);

        cv::Mat result_host = dst;
        cv::imshow("Result", result_host);
        cv::waitKey();
    }
    catch(const cv::Exception& ex)
    {
        std::cout << "Error: " << ex.what() << std::endl;
    }
    return 0;
}

The compiling command is:

g++ testgpu.cpp -o test `pkg-config --cflags --libs opencv` -lopencv_gpu

It has the following compiling errors:

testgpu.cpp: In function ‘int main(int, char**)’:
testgpu.cpp:13:51: error: ‘CV_LOAD_IMAGE_GRAYSCALE’ was not declared in this scope
         cv::Mat src_host = cv::imread("file.png", CV_LOAD_IMAGE_GRAYSCALE);
                                                   ^
testgpu.cpp:17:52: error: ‘CV_THRESH_BINARY’ was not declared in this scope
         cv::gpu::threshold(src, dst, 128.0, 255.0, CV_THRESH_BINARY);
                                                    ^
testgpu.cpp:19:31: error: conversion from ‘cv::gpu::GpuMat’ to non-scalar type ‘cv::Mat’ requested
         cv::Mat result_host = dst;
                           ^

It is something wrong with the installation of OpenCV, or the API change in Opencv 3.0.0?

Upvotes: 11

Views: 16302

Answers (2)

berak
berak

Reputation: 39796

Ah, they've been playing with the constants in master. Expect the CV_* prefix removed almost anywhere ( except the types, CV_8U and such are still alive).

So it's cv::THRESH_BINARY, cv::LOAD_IMAGE_GRAYSCALE, but .... cv::COLOR_BGR2GRAY (you didn't use it now, but i'll spare you the searching ;) )

Sorry, I've no experience with GPU stuff, so i can't solve the last riddle there.

Upvotes: 7

vinograd47
vinograd47

Reputation: 6420

The gpu module was redesigned in OpenCV 3.0. It was splitted onto several modules, it was renamed to cuda and gpu:: namespace was renamed to cuda::. The correct code for OpenCV 3.0:

#include <iostream>
#include "opencv2/opencv.hpp"
#include "opencv2/core.hpp"
#include "opencv2/highgui.hpp"
#include "opencv2/cudaarithm.hpp"

using namespace cv;

int main (int argc, char* argv[])
{
    try
    {
        cv::Mat src_host = cv::imread("file.png", cv::IMREAD_GRAYSCALE);
        cv::cuda::GpuMat dst, src;
        src.upload(src_host);
        cv::cuda::threshold(src, dst, 128.0, 255.0, cv::THRESH_BINARY);
        cv::Mat result_host(dst);
        cv::imshow("Result", result_host);
        cv::waitKey();
    }
    catch(const cv::Exception& ex)
    {
        std::cout << "Error: " << ex.what() << std::endl;
    }
    return 0;
}

Upvotes: 25

Related Questions