我正在使用OpenCVSharp(在Visual Studio 2015中运行并通过NuGet安装的OpenCvSharp3-AnyCPU版本3.0.0.20150823)从C#访问OpenCV,但是当调用Cv2.PCACompute
时,我得到了一个通用的OpenCVException,指出我的输入输出数组不受支持格式。
我的目标是使用PCA查找像素斑点的主轴。这是我目前的代码(分解后的代码):
using OpenCvSharp;
public struct point2D
{
public int X;
public int Y;
public point2D(int X, int Y)
{
this.X = X;
this.Y = Y;
}
}
public static void PCA2D()
{
int height = 5;
int width = 5;
int[] image = new int[]
{
0,0,0,0,1,
0,0,0,1,0,
0,0,1,0,0,
0,1,0,0,0,
1,0,0,0,0,
}
// extract the datapoints
List<point2D> dataPoints = new List<point2D>();
for (int row = 0; row < height; ++row)
{
for (int col = 0; col < width; ++col)
{
if (image[row * width + col] == 1)
{
dataPoints.Add(new point2D(col, row));
}
}
}
// create the input matrix
Mat input = new Mat(dataPoints.Length, 2, MatType.CV_32SC1);
for (int i = 0; i < dataPoints.Length; ++i)
{
input.Set(i, 0, dataPoints[i].X);
input.Set(i, 1, dataPoints[i].Y);
}
Mat mean = new Mat();
Mat eigenvectors = new Mat();
// OpenCVException occurs here: unsupported combination of input and output array formats
Cv2.PCACompute(input, mean, eigenvectors);
// Code to get orientation from the eigenvectors
}
我没有找到有关如何初始化均值和特征 vector Mats的文档,或者我调用PCACompute的方法是否正确。对使用PCACompute的正确过程的一些了解将非常有帮助。
最佳答案
因此,事实证明dataPoints
不能是MatType.CV_32SC1
。将代码更改为以下代码使其可以工作:
// create the input matrix
Mat input = new Mat(dataPoints.Length, 2, MatType.CV_32FC1);
for (int i = 0; i < dataPoints.Length; ++i)
{
input.Set(i, 0, (float)dataPoints[i].X);
input.Set(i, 1, (float)dataPoints[i].Y);
}