I want to process image with pointer to make it faster.At first,I do it like this
unsafe
{
double a = 288 / 55;
double b = -215 * 288 / 55;
MIplImage ss = diff.MIplImage;
for (int i = 0; i < ss.height; i++)
{
IntPtr ptr = ss.imageData + i * ss.widthStep;
for (int j = 0; j < ss.width; j++)
{
if (a * j - i + b > 0)
((byte*)(ptr))[j] = 0;
}
}
}
and it performs well(0.5ms while 4ms without pointer).
Then I find that the managed objects should be fixed to prevent relocation.So I think it would be like this
Image<Gray, byte> diff = new Image<Gray, byte>(frame.Width,frame.Height);
fixed (void* p_temp=diff.Ptr.ToPointer()){}
or
fixed (byte* p_temp=(byte*)temp.Ptr){}
But it is wrong.So how to fix the Image? And how to use pointer with emgucv? I am really confused.Thanks!
unsafe
{
var data = diff.Data;
int stride = diff.MIplImage.widthStep;
byte* p;
fixed (byte* pData = data)
{
for (int i = 0; i < diff.Height; i++)
{
p = pData + i * stride;
for (int j = 0; j < diff.Width; j++)
{
if (a * j - i + b > 0)
*(p + j) = (byte)0;
}
}
}
}
I just made a quick demo (not production quality code) to give you a starting working example on unsafe emgu usage.
Image<Gray, Byte> img = new Image<Gray, byte>(510, 510);
// Fill image with random values
img.SetRandUniform(new MCvScalar(), new MCvScalar(255));
// Show Image
ImageViewer.Show(img);
var data = img.Data;
int stride = img.MIplImage.widthStep;
fixed (byte* pData = data)
{
for (int i = 0; i < 255 * stride; i++)
*(pData + i) = (byte)(i % stride);
}
// Show Image Again
ImageViewer.Show(img);
First I fill image with random noise and then fill some rows with a gradient :-)
The technical post webpages of this site follow the CC BY-SA 4.0 protocol. If you need to reprint, please indicate the site URL or the original address.Any question please contact:yoyou2525@163.com.