我正在尝试计算仅由一行组成的矩阵的直方图。我使用OpenCV3.1在VS2015中编写了以下代码:

//Create matrix stromal pop
cv::Mat stromalHistogram(1, i_stromalIID.size(), CV_32F);
float * ref_ptr = stromalHistogram.ptr<float>(0);
std::copy(i_stromalIID.begin(),
          i_stromalIID.end(),
          stdext::checked_array_iterator<float *>(ref_ptr, stromalHistogram.cols));

#ifdef _DEBUG
std::cout << "Matrix =" << std::endl << stromalHistogram << std::endl;
#endif

// calculate histogram for stromal pop
auto  itRangesRef     = std::minmax_element(i_stromalIID.begin(), i_stromalIID.end());
float stromalRanges[] = {*itRangesRef.first, *itRangesRef.second};
const float *      rangesRef[] = {stromalRanges};
const int          channel     = 0;
const int          histSize    = 256;
std::vector<float> hist_out;
cv::calcHist(
  &stromalHistogram, 1, &channel, cv::Mat(), hist_out, 1, &histSize, rangesRef);
i_stromalIID是从外部传递的std::vector<double>。正确填写了cv::Mat stromalHistogram,因为在我打印时,一切都与我预期的一样(1行1203列)。但是,当程序运行cv::calcHist时,出现以下错误:



我尝试调试OpenCV代码,尝试执行时错误在cv::calcHist中:
void cv::calcHist( const Mat* images, int nimages, const int* channels,
                   InputArray _mask, OutputArray _hist, int dims, const int* histSize,
                   const float** ranges, bool uniform, bool accumulate )
{
      .....
      .....

      _hist.create(dims, histSize, CV_32F);
}

然后在matrix.cpp内部调用:
void _OutputArray::create(int d, const int* sizes, int mtype, int i,
                          bool allowTransposed, int fixedDepthMask) const
{
      .....
      .....

      if( k == STD_VECTOR || k == STD_VECTOR_VECTOR )
      {
        CV_Assert( d == 2 && (sizes[0] == 1 || sizes[1] == 1 || sizes[0]*sizes[1] == 0) );
        .....
        .....
      }
}

并且此断言失败,因为在我的情况下d等于1。

我究竟做错了什么?

最佳答案

将Mat大小更改为源图像。

10-07 14:05