Develop Reference

WebBrowser Html Document to Image

I'm trying to make image of webpage, but some pages shows me as white page.
In Registry editor browse \HKEY_CURRENT_USER\Software\Microsoft\Internet Explorer\Main\FeatureControl\FEATURE_BROWSER_EMULATION\
and add there this:
WindowsFormsApp1.exe with decimal value 11000
WindowsFormsApp1.vshost.exe with decimal value 11000
Here is my code:
using System;
using System.Collections.Generic;
using System.Windows.Forms;
using System.Drawing;
using System.Drawing.Imaging;
using System.Runtime.InteropServices;
namespace WindowsFormsApp1
{
public partial class Form1 : Form
{
Dictionary<Uri, Bitmap> browserShots = new Dictionary<Uri, Bitmap>();
WebBrowser browser = new WebBrowser();
public Form1()
{
InitializeComponent();
browser.DocumentCompleted += new WebBrowserDocumentCompletedEventHandler(browser_DocumentCompleted);
}
//=========================================MADE BY JIMY====================================
private void browser_DocumentCompleted(object sender, WebBrowserDocumentCompletedEventArgs e)
{
var browser = sender as WebBrowser;
if (browser.ReadyState != WebBrowserReadyState.Complete) return;
var bitmap = WebBrowserExtender.DrawContent(browser);
if (bitmap != null)
{
if (!browserShots.ContainsKey(browser.Url))
browserShots.Add(browser.Url, bitmap);
else
{
browserShots[browser.Url]?.Dispose();
browserShots[browser.Url] = bitmap;
}
// Show the Bitmap in a PictureBox control, eventually
pictureBox1.Image = browserShots[browser.Url];
}
}
public class WebBrowserExtender
{
public static Bitmap DrawContent(WebBrowser browser)
{
if (browser.Document == null) return null;
Size docSize = Size.Empty;
Graphics g = null;
var hDc = IntPtr.Zero;
try
{
docSize.Height = (int)((dynamic)browser.Document.DomDocument).documentElement.scrollHeight;
docSize.Width = (int)((dynamic)browser.Document.DomDocument).documentElement.scrollWidth;
docSize.Height = Math.Max(Math.Min(docSize.Height, 32750), 1);
docSize.Width = Math.Max(Math.Min(docSize.Width, 32750), 1);
var previousSize = browser.ClientSize;
browser.ClientSize = new Size(docSize.Width, docSize.Height);
var bitmap = new Bitmap(docSize.Width, docSize.Height, PixelFormat.Format32bppArgb);
g = Graphics.FromImage(bitmap);
var rect = new RECT(0, 0, bitmap.Width, bitmap.Height);
hDc = g.GetHdc();
var view = browser.ActiveXInstance as IViewObject;
view.Draw(1, -1, IntPtr.Zero, IntPtr.Zero, IntPtr.Zero, hDc, ref rect, IntPtr.Zero, IntPtr.Zero, 0);
browser.ClientSize = previousSize;
return bitmap;
}
catch
{
// This catch block is like this on purpose: nothing to do here
return null;
}
finally
{
if (hDc != null) g?.ReleaseHdc(hDc);
g?.Dispose();
}
}
[ComImport]
[Guid("0000010D-0000-0000-C000-000000000046")]
[InterfaceType(ComInterfaceType.InterfaceIsIUnknown)]
interface IViewObject
{
void Draw(uint dwAspect, int lindex, IntPtr pvAspect, [In] IntPtr ptd,
IntPtr hdcTargetDev, IntPtr hdcDraw, ref RECT lprcBounds,
[In] IntPtr lprcWBounds, IntPtr pfnContinue, uint dwContinue);
}
[StructLayout(LayoutKind.Sequential, Pack = 4)]
struct RECT
{
public int Left;
public int Top;
public int Right;
public int Bottom;
public RECT(int left, int top, int width, int height)
{
Left = left; Top = top; Right = width; Bottom = height;
}
}
}
//=========================================MADE BY JIMY====================================}
private void button1_Click(object sender, EventArgs e)
{
browser.Navigate(textBox1.Text, null, null, "User-Agent: User agent");
}
}
}

In order to print the Html content of a WebBrowser Control, there are a few points that need to be considered:
We need to use the WebBrowser's DocumentCompleted event to determine when the current Document is loaded and rendered
A single Document may (will) contain more that one sub-Document, usually contained inside Frames/IFrames. Each IFrame contains its own Document: when a Document contained in an IFrame is loaded, the DocumentCompleted is reaised. This means that the event can and will be raised multiple times when the WebBrowser navigates to a URL.
The notes here explain more: How to get an HtmlElement value inside Frames/IFrames?
The managed properties of the WebBrowser don't always reflect the DOM's real values. For example, the actual dimensions of the Html Document, when the rendering is completed, are not reflected anywhere, so we need to get those measures from the DOM ourselves. The current DOM rendered dimensions are referenced by:
[WebBrowser].Document.DomDocument.documentElement.scrollHeight;
[WebBrowser].Document.DomDocument.documentElement.scrollWidth;
See: Measuring Element Dimension and Location with CSSOM in Windows Internet Explorer
The WebBrowser Control DrawToBitmap() method is derived from Control but it's not actually implemented as we could expect. The same applies to other Controls: the RichTextBox is known to print blank content when this method is used.
A Html Document may be larger than the maximum Size supported by a Bitmap. There is also a more subtle memory limit: the Bitmap object needs to store its content in a contiguous memory space, so the limit in Size of a Bitmap is actually hard to pre-determine and may cause exceptions when we might not expect it.
The WebBrowser control's Emulation Feature must be set to Internet Explorer 11. See:
How can I get the WebBrowser control to show modern contents?
Web browser control emulation issue (FEATURE_BROWSER_EMULATION)
To proceed, first subscribe to DocumentCompleted event of the WebBrowser Control.
A Dictionary<Uri, Bitmap> is used here to store the Bitmap representing the Html content of URLs visited in a session.
When the DocumentCompleted event is raised, we add a new element to the Dictionary when the current URL has never been visited before.
If the Uri is already stored, we updated the related Bitmap object, so only the most recent snapshot of a Html Document is present in the collection.
I'm using a support class to handle the Bitmaps creation and to declare the native COM Interface used to generate the Bitmap from the current ISurfacePresenter.
Since the WebBrowser control is forced to use VIEW_OBJECT_COMPOSITION_MODE_LEGACY as the CompositionMode for all sites, the internal GetPrintBitmap method calls the IViewObject Interface Draw() method in this situation, so do we.
To print the content (all the content) of the current Html Document, call the
DrawContent(WebBrowser browser) static method of the WebBrowserExtender class:
Dictionary<Uri, Bitmap> browserShots = new Dictionary<Uri, Bitmap>();
private void browser_DocumentCompleted(object sender, WebBrowserDocumentCompletedEventArgs e)
{
var browser = sender as WebBrowser;
if (browser.ReadyState != WebBrowserReadyState.Complete) return;
var bitmap = WebBrowserExtender.DrawContent(browser);
if (bitmap != null) {
if (!browserShots.ContainsKey(browser.Url)) {
browserShots.Add(browser.Url, bitmap);
}
else {
browserShots[browser.Url]?.Dispose();
browserShots[browser.Url] = bitmap;
}
// Show the Bitmap in a PictureBox control, eventually
[PictureBox].Image = browserShots[browser.Url];
}
}
The WebBrowserExtender support class:
using System.Drawing;
using System.Drawing.Imaging;
using System.Runtime.InteropServices;
using System.Windows.Forms;
public class WebBrowserExtender
{
public static Bitmap DrawContent(WebBrowser browser)
{
if (browser.Document == null) return null;
Size docSize = Size.Empty;
Graphics g = null;
var hDc = IntPtr.Zero;
try {
docSize.Height = (int)((dynamic)browser.Document.DomDocument).documentElement.scrollHeight;
docSize.Width = (int)((dynamic)browser.Document.DomDocument).documentElement.scrollWidth;
var screenWidth = Screen.FromHandle(browser.Handle).Bounds.Width;
docSize.Width = Math.Max(Math.Min(docSize.Width, screenWidth), 1);
docSize.Height = Math.Max(Math.Min(docSize.Height, 32750), 1);
var previousSize = browser.ClientSize;
browser.ClientSize = new Size(docSize.Width, docSize.Height);
var bitmap = new Bitmap(docSize.Width, docSize.Height, PixelFormat.Format32bppArgb);
g = Graphics.FromImage(bitmap);
var rect = new RECT(0, 0, bitmap.Width, bitmap.Height);
hDc = g.GetHdc();
var view = browser.ActiveXInstance as IViewObject;
view.Draw(1, -1, IntPtr.Zero, IntPtr.Zero, IntPtr.Zero, hDc, ref rect, IntPtr.Zero, IntPtr.Zero, 0);
browser.ClientSize = previousSize;
return bitmap;
}
catch {
// This catch block is like this on purpose: nothing to do here
return null;
}
finally {
if (hDc != null) g?.ReleaseHdc(hDc);
g?.Dispose();
}
}
[ComImport]
[Guid("0000010D-0000-0000-C000-000000000046")]
[InterfaceType(ComInterfaceType.InterfaceIsIUnknown)]
interface IViewObject
{
void Draw(uint dwAspect, int lindex, IntPtr pvAspect, [In] IntPtr ptd,
IntPtr hdcTargetDev, IntPtr hdcDraw, ref RECT lprcBounds,
[In] IntPtr lprcWBounds, IntPtr pfnContinue, uint dwContinue);
}
[StructLayout(LayoutKind.Sequential, Pack = 4)]
struct RECT
{
public int Left;
public int Top;
public int Right;
public int Bottom;
public RECT(int left, int top, int width, int height)
{
Left = left; Top = top; Right = width; Bottom = height;
}
}
}
This is how it works:
The full Document is captured. Of course, the Bitmap can also be limited to a specific maximum/minimum size, to capture just a section of the Html Document.
:
Sample WinForms Project on Google Drive.

try to set User Agent like this
browser.Navigate(url, null, null, "User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:74.0) Gecko/20100101 Firefox/74.0");

How to save drew Graphics of "Paint()" into image using c#?

I actually wanted to Convert RTF into Image so after googling a lot I've got a code that does it by Paint() Event of Picturebox1 and it works perfectly :
private void pictureBox1_Paint(object sender, PaintEventArgs e)
{
e.Graphics.Clear(richTextBox1.BackColor);
e.Graphics.DrawRtfText(this.richTextBox1.Rtf, this.pictureBox1.ClientRectangle);
base.OnPaint(e);
// below code just create an empty image file
Bitmap newBitmap = new Bitmap(pictureBox1.Width, pictureBox1.Height);
e.Graphics.DrawImage(newBitmap, new Rectangle(0, 0, pictureBox1.Width, pictureBox1.Height), new Rectangle(0, 0, pictureBox1.Width, pictureBox1.Height), GraphicsUnit.Pixel);
newBitmap.Save(#"c:\adv.jpg");
}
in the picture above the left is my richTextBox and the right is a Picturebox.
the ISSUE is I don't know how to save Paint() drew graphic into a file because the 3 last lines of my code just save an empty image.
UPDATE #1:
g.SmoothingMode = SmoothingMode.AntiAlias;
g.TextRenderingHint = System.Drawing.Text.TextRenderingHint.ClearTypeGridFit;
g.InterpolationMode = InterpolationMode.HighQualityBicubic;
g.PixelOffsetMode = PixelOffsetMode.HighQuality;
g.Clear(richTextBox1.BackColor);
g.DrawRtfText(this.richTextBox1.Rtf, this.pictureBox1.ClientRectangle);
by changing the graphics from e.graphics to g the issue is resolved but with one other issue that the quality of bitmap is too low. I've Added this bunch of code but I've got same result, the quality is too low!
Any suggestions?
UPDATE #2
here is the Graphics_DrawRtfText class that does the conversion :
public static class Graphics_DrawRtfText
{
private static RichTextBoxDrawer rtfDrawer;
public static void DrawRtfText(this Graphics graphics, string rtf, Rectangle layoutArea)
{
if (Graphics_DrawRtfText.rtfDrawer == null)
{
Graphics_DrawRtfText.rtfDrawer = new RichTextBoxDrawer();
}
Graphics_DrawRtfText.rtfDrawer.Rtf = rtf;
Graphics_DrawRtfText.rtfDrawer.Draw(graphics, layoutArea);
}
private class RichTextBoxDrawer : RichTextBox
{
//Code converted from code found here: http://support.microsoft.com/kb/812425/en-us
//Convert the unit used by the .NET framework (1/100 inch)
//and the unit used by Win32 API calls (twips 1/1440 inch)
private const double anInch = 14.4;
protected override CreateParams CreateParams
{
get
{
CreateParams createParams = base.CreateParams;
if (SafeNativeMethods.LoadLibrary("msftedit.dll") != IntPtr.Zero)
{
createParams.ExStyle |= SafeNativeMethods.WS_EX_TRANSPARENT; // transparent
createParams.ClassName = "RICHEDIT50W";
}
return createParams;
}
}
public void Draw(Graphics graphics, Rectangle layoutArea)
{
//Calculate the area to render.
SafeNativeMethods.RECT rectLayoutArea;
rectLayoutArea.Top = (int)(layoutArea.Top * anInch);
rectLayoutArea.Bottom = (int)(layoutArea.Bottom * anInch);
rectLayoutArea.Left = (int)(layoutArea.Left * anInch);
rectLayoutArea.Right = (int)(layoutArea.Right * anInch);
IntPtr hdc = graphics.GetHdc();
SafeNativeMethods.FORMATRANGE fmtRange;
fmtRange.chrg.cpMax = -1; //Indicate character from to character to
fmtRange.chrg.cpMin = 0;
fmtRange.hdc = hdc; //Use the same DC for measuring and rendering
fmtRange.hdcTarget = hdc; //Point at printer hDC
fmtRange.rc = rectLayoutArea; //Indicate the area on page to print
fmtRange.rcPage = rectLayoutArea; //Indicate size of page
IntPtr wParam = IntPtr.Zero;
wParam = new IntPtr(1);
//Get the pointer to the FORMATRANGE structure in memory
IntPtr lParam = IntPtr.Zero;
lParam = Marshal.AllocCoTaskMem(Marshal.SizeOf(fmtRange));
Marshal.StructureToPtr(fmtRange, lParam, false);
SafeNativeMethods.SendMessage(this.Handle, SafeNativeMethods.EM_FORMATRANGE, wParam, lParam);
//Free the block of memory allocated
Marshal.FreeCoTaskMem(lParam);
//Release the device context handle obtained by a previous call
graphics.ReleaseHdc(hdc);
}
#region SafeNativeMethods
private static class SafeNativeMethods
{
[DllImport("USER32.dll")]
public static extern IntPtr SendMessage(IntPtr hWnd, int msg, IntPtr wp, IntPtr lp);
[DllImport("kernel32.dll", CharSet = CharSet.Auto)]
public static extern IntPtr LoadLibrary(string lpFileName);
[StructLayout(LayoutKind.Sequential)]
public struct RECT
{
public int Left;
public int Top;
public int Right;
public int Bottom;
}
[StructLayout(LayoutKind.Sequential)]
public struct CHARRANGE
{
public int cpMin; //First character of range (0 for start of doc)
public int cpMax; //Last character of range (-1 for end of doc)
}
[StructLayout(LayoutKind.Sequential)]
public struct FORMATRANGE
{
public IntPtr hdc; //Actual DC to draw on
public IntPtr hdcTarget; //Target DC for determining text formatting
public RECT rc; //Region of the DC to draw to (in twips)
public RECT rcPage; //Region of the whole DC (page size) (in twips)
public CHARRANGE chrg; //Range of text to draw (see earlier declaration)
}
public const int WM_USER = 0x0400;
public const int EM_FORMATRANGE = WM_USER + 57;
public const int WS_EX_TRANSPARENT = 0x20;
}
#endregion
}
}

Disclaimer: I don't have the time to dig into the posted extension method but it is interesting and works well, at least when drawing onto a control surface.
But I could reproduce how bad the results are when drawing into a bitmap..
But: When done right the saved results are excellent!
So here here are a few things to keep in mind:
Saving in the Paint event is a bad idea, as this event will be triggered by the system whenever it needs to redraw the control; test by doing a minimize/maximize cycle.
In addition the DrawRtfText semms to create a double-vision effect when drawing into a bitmap.
So make sure you use DrawToBitmap to grab the results. For this you need to place the call to DrawRtfText in the Paint event of a control!
Also make sure to have large enough resolutions both in the control (pixel size) and the Bitmap (dpi) to get nice, crispy and (if needed) printable results.
Do not save to jpg as this is bound to result in blurry text! Png is the format of choice!
Here is a Paint event:
private void panel1_Paint(object sender, PaintEventArgs e)
{
e.Graphics.Clear(richTextBox1.BackColor);
e.Graphics.TextRenderingHint = System.Drawing.Text.TextRenderingHint.AntiAlias;
e.Graphics.SmoothingMode = SmoothingMode.AntiAlias;
Padding pad = new Padding(120, 230, 10, 30); // pick your own numbers!
Size sz = panel1.ClientSize;
Rectangle rect = new Rectangle(pad.Left, pad.Top,
sz.Width - pad.Horizontal, sz.Height - pad.Vertical);
e.Graphics.DrawRtfText(this.richTextBox1.Rtf, rect);
}
Note that it pays to improve on the default quality settings; if you don't the text in the resulting file will break apart when zooming in..
Here is a Save button click:
private void button1_Click(object sender, EventArgs e)
{
Size sz = panel1.ClientSize;
// first we (optionally) create a bitmap in the original panel size:
Rectangle rect1 = panel1.ClientRectangle;
Bitmap bmp = new Bitmap(rect1.Width, rect1.Height);
panel1.DrawToBitmap(bmp, rect);
bmp.Save("D:\\rtfImage1.png", ImageFormat.Png);
// now we create a 4x larger one:
Rectangle rect2 = new Rectangle(0, 0, sz.Width * 4, sz.Height * 4);
Bitmap bmp2 = new Bitmap(rect2.Width, rect2.Height);
// we need to temporarily enlarge the panel:
panel1.ClientSize = rect2.Size;
// now we can let the routine draw
panel1.DrawToBitmap(bmp2, rect2);
// and before saving we optionally can set the dpi resolution
bmp2.SetResolution(300, 300);
// optionally make background transparent:
bmp2.MakeTransparent(richTextBox1.BackColor);
UnSemi(bmp2); // see the link in the comment!
// save text always as png; jpg is only for fotos!
bmp2.Save("D:\\rtfImage2.png", ImageFormat.Png);
// restore the panels size
panel1.ClientSize = sz;
}
I found the result to be really good.
Note that DrawToBitmap will internally trigger the Paint event to grab the drawn graphics.
Of course you don't need both parts - use only the one you want (.e. skip the 1st part, between first and now ) and do use your own numbers. It helps to know what the output shall be and calculate the necessary sizes and resolutions backward from there.
I added the enlarged version because usually the monitor resolution, which is what the controls all have, is rather limited, around 75-100dpi, while print quality starts only at 150dpi..
Here is a link to the UnSemi function

Your code produces an empty image file because you are not drawing anything onto 'newBitmap'.
If you want to draw anything onto 'newBitmap' you need to create a Graphics object from it. As I do not know where 'DrawRtfText' comes from and how it works my guess would be:
private void pictureBox1_Paint(object sender, PaintEventArgs e)
{
//...
Bitmap newBitmap = new Bitmap(pictureBox1.Width, pictureBox1.Height);
Graphics g = Graphics.FromImage(newBitmap);
g.DrawRtfText(this.richTextBox1.Rtf, this.pictureBox1.ClientRectangle);
newBitmap.Save(#"d:\adv.jpg");
}

Bitmap gets mangled when I converted from System.Drawing.Bitmap object to cv::Mat

I have a WPF application that takes a screen shot of the running Handbrake executable using a class called ScreenCapture that I copied from stack overflow.
public class ScreenCapture
{
[DllImport("user32.dll")]
static extern int GetWindowRgn(IntPtr hWnd, IntPtr hRgn);
//Region Flags - The return value specifies the type of the region that the function obtains. It can be one of the following values.
const int ERROR = 0;
const int NULLREGION = 1;
const int SIMPLEREGION = 2;
const int COMPLEXREGION = 3;
[DllImport("user32.dll")]
[return: MarshalAs(UnmanagedType.Bool)]
static extern bool GetWindowRect(HandleRef hWnd, out RECT lpRect);
[DllImport("gdi32.dll")]
static extern IntPtr CreateRectRgn(int nLeftRect, int nTopRect, int nRightRect, int nBottomRect);
[DllImport("user32.dll", SetLastError = true)]
[return: MarshalAs(UnmanagedType.Bool)]
static extern bool PrintWindow(IntPtr hwnd, IntPtr hDC, uint nFlags);
[StructLayout(LayoutKind.Sequential)]
public struct RECT
{
public int Left, Top, Right, Bottom;
public RECT(int left, int top, int right, int bottom)
{
Left = left;
Top = top;
Right = right;
Bottom = bottom;
}
public RECT(System.Drawing.Rectangle r) : this(r.Left, r.Top, r.Right, r.Bottom) { }
public int X
{
get { return Left; }
set { Right -= (Left - value); Left = value; }
}
public int Y
{
get { return Top; }
set { Bottom -= (Top - value); Top = value; }
}
public int Height
{
get { return Bottom - Top; }
set { Bottom = value + Top; }
}
public int Width
{
get { return Right - Left; }
set { Right = value + Left; }
}
public System.Drawing.Point Location
{
get { return new System.Drawing.Point(Left, Top); }
set { X = value.X; Y = value.Y; }
}
public System.Drawing.Size Size
{
get { return new System.Drawing.Size(Width, Height); }
set { Width = value.Width; Height = value.Height; }
}
public static implicit operator System.Drawing.Rectangle(RECT r)
{
return new System.Drawing.Rectangle(r.Left, r.Top, r.Width, r.Height);
}
public static implicit operator RECT(System.Drawing.Rectangle r)
{
return new RECT(r);
}
public static bool operator ==(RECT r1, RECT r2)
{
return r1.Equals(r2);
}
public static bool operator !=(RECT r1, RECT r2)
{
return !r1.Equals(r2);
}
public bool Equals(RECT r)
{
return r.Left == Left && r.Top == Top && r.Right == Right && r.Bottom == Bottom;
}
public override bool Equals(object obj)
{
if (obj is RECT)
return Equals((RECT)obj);
else if (obj is System.Drawing.Rectangle)
return Equals(new RECT((System.Drawing.Rectangle)obj));
return false;
}
public override int GetHashCode()
{
return ((System.Drawing.Rectangle)this).GetHashCode();
}
public override string ToString()
{
return string.Format(System.Globalization.CultureInfo.CurrentCulture, "{{Left={0},Top={1},Right={2},Bottom={3}}}", Left, Top, Right, Bottom);
}
}
public Bitmap GetScreenshot(IntPtr ihandle)
{
IntPtr hwnd = ihandle;//handle here
RECT rc;
GetWindowRect(new HandleRef(null, hwnd), out rc);
Bitmap bmp = new Bitmap(rc.Right - rc.Left, rc.Bottom - rc.Top, PixelFormat.Format32bppArgb);
Graphics gfxBmp = Graphics.FromImage(bmp);
IntPtr hdcBitmap;
try
{
hdcBitmap = gfxBmp.GetHdc();
}
catch
{
return null;
}
bool succeeded = PrintWindow(hwnd, hdcBitmap, 0);
gfxBmp.ReleaseHdc(hdcBitmap);
if (!succeeded)
{
gfxBmp.FillRectangle(new SolidBrush(Color.Gray), new Rectangle(Point.Empty, bmp.Size));
}
IntPtr hRgn = CreateRectRgn(0, 0, 0, 0);
GetWindowRgn(hwnd, hRgn);
Region region = Region.FromHrgn(hRgn);//err here once
if (!region.IsEmpty(gfxBmp))
{
gfxBmp.ExcludeClip(region);
gfxBmp.Clear(Color.Transparent);
}
gfxBmp.Dispose();
return bmp;
}
public void WriteBitmapToFile(string filename, Bitmap bitmap)
{
bitmap.Save(filename, ImageFormat.Bmp);
}
So when the button click handler below is called a screenshot of the handbrake window is taken.
I write it to the harddrive to make sure its ok:
handbrake screen shot.
I create an instance of a CLR class library ClassLibrary1::Class1 and call the method "DoSomething" passing it the System.Drawing.Bitmap object.
private void button4_Click(object sender, RoutedEventArgs e)
{
string wName = "HandBrake";
IntPtr hWnd = IntPtr.Zero;
foreach (Process pList in Process.GetProcesses())
{
if (pList.MainWindowTitle.Contains(wName))
{
hWnd = pList.MainWindowHandle;
var sc = new ScreenCapture();
SetForegroundWindow(hWnd);
var bitmap = sc.GetScreenshot(hWnd);
sc.WriteBitmapToFile("handbrake.bmp", bitmap);
Bitmap image1 = (Bitmap)System.Drawing.Image.FromFile("handbrake.bmp", true);
ClassLibrary1.Class1 opencv = new ClassLibrary1.Class1();
opencv.DoSomething(image1);
}
}
}
Inside DoSomething I attempt to convert the System.Drawing.Bitmap to a OpenCV class cv::Mat. I call cv::imwrite to make sure the bitmap is still ok, unfortunately somethings gone wrong: mangled handbrake screenshot
void Class1::DoSomething(Bitmap ^mybitmap)
{
cv::Mat *imgOriginal;
// Lock the bitmap's bits.
Rectangle rect = Rectangle(0, 0, mybitmap->Width, mybitmap->Height);
Imaging::BitmapData^ bmpData = mybitmap->LockBits(rect, Imaging::ImageLockMode::ReadWrite, mybitmap->PixelFormat);
try
{
// Get the address of the first line.
IntPtr ptr = bmpData->Scan0;
// Declare an array to hold the bytes of the bitmap.
// This code is specific to a bitmap with 24 bits per pixels.
int bytes = Math::Abs(bmpData->Stride) * mybitmap->Height;
array<Byte>^rgbValues = gcnew array<Byte>(bytes);
// Copy the RGB values into the array.
System::Runtime::InteropServices::Marshal::Copy(ptr, rgbValues, 0, bytes);
imgOriginal = new cv::Mat(mybitmap->Height, mybitmap->Width, CV_8UC3, (void *)ptr, std::abs(bmpData->Stride));
}
finally { mybitmap->UnlockBits(bmpData); }//Remember to unlock!!!
cv::imwrite("from_mat.bmp", *imgOriginal);
}
Can anybody spot my error?

Since your image is stretched horizontally, I'm betting that you have the wrong pixel format selected. (It's not stretched vertically, nor skewed diagonally, so the stride is correct.) CV_8UC3 specifies 24 bits per pixel, but I think that your BMP file is using 32 bits per pixel.
Switch your pixel format to CV_8UC4, or better yet, read the number of bits per pixel from the image and select the correct CV format based on that.
Side note: Since you're doing sc.WriteBitmapToFile() followed by opencv.DoSomething(Image.FromFile(), the entire bit about how you're capturing the screenshot is irrelevant. You're reading the bitmap from a file; that's all that matters.

Simulate Windows Dragging Effect in a Winform FlowLayoutPanel

I'm currently simulating the windows multiple selection rectangle when the user is dragging the mouse. To synchronize our understanding, this picture shows the effect I want to simulate:
Now I want to simulate this effect on a FlowLayoutPanel with some controls inside.
So far I am managed to get the effect almost done:
What I did here was putting a unfocused border-less semi-transparent (half the opacity) form on top the main form. To get the border simulated, I handled SizeChanged and Paint to draw the border.
However, this solution sometimes flickers, as in the owner border couldn't get cleared on-time:
I have tried using double buffering on the cover form by setting DoubleBuffer to true, and override CreateParam to set WM_EX_COMPOSITED, but neither works.
My question is: How to reduce this artifact?
Thanks a lot!
My code:
For the cover form:
public partial class CoverForm : Form
{
public CoverForm()
{
InitializeComponent();
BackColor = Color.CadetBlue;
FormBorderStyle = FormBorderStyle.None;
SizeChanged += (s, e) => Invalidate();
Paint += (s, e) =>
{
e.Graphics.Clear(BackColor);
using (var pen = new Pen(Color.DodgerBlue))
{
e.Graphics.DrawRectangle(pen, 1, 1, Size.Width - 2, Size.Height - 2);
}
};
}
protected override bool ShowWithoutActivation
{
get { return true; }
}
}
For the main form:
public Form1()
{
InitializeComponent();
// mainPanel is the panel that simulates the dragging effect
mainPanel.MouseDown += (s, e) =>
{
_isMouseDown = true;
_startPosition = e.Location;
coverForm.Location = mainPanel.PointToScreen(e.Location);
coverForm.Show();
};
mainPanel.MouseUp += (s, e) =>
{
_isMouseDown = false;
coverForm.Hide();
};
mainPanel.MouseMove += CoverPanelMouseMoveHandler;
DoubleBuffered = true;
}
~Form1()
{
if (coverForm != null && !coverForm.IsDisposed)
{
coverForm.Dispose();
}
}
# region Dragging Effect
private void CoverPanelMouseMoveHandler(object sender, MouseEventArgs e)
{
if (_isMouseDown)
{
_curPosition = e.Location;
// find the dragging rectangle
var rect = CreateRect(_curPosition, _startPosition);
coverForm.Size = rect.Size;
coverForm.Location = mainPanel.PointToScreen(rect.Location);
foreach (Control control in mainPanel.Controls)
{
// logic to get button backcolor changed
}
mainPanel.Invalidate(true);
}
}
Update
I have tried to override OnPaint and put my drawing there, but it gave even worse result: the old paints wouldn't get erased:
Code I modified for cover form:
public partial class CoverForm : Form
{
public CoverForm()
{
InitializeComponent();
BackColor = Color.CadetBlue;
FormBorderStyle = FormBorderStyle.None;
}
protected override void OnPaint(PaintEventArgs e)
{
base.OnPaint(e);
e.Graphics.Clear(BackColor);
using (var pen = new Pen(Color.FromArgb(255, 0, 0, 255)))
{
e.Graphics.DrawRectangle(pen, 0, 0, Size.Width - 1, Size.Height - 1);
}
}
protected override bool ShowWithoutActivation
{
get { return true; }
}
}
Update 2
Actually the problem I am facing is about drawing above a FlowLayoutPanel, not a normal Panel. The reason I put Panel before was I was seeking answer for my flickering 2-layers design. But since someone approach the problem by adding control to the panel to get it drawn above all controls, I would like to point this out: adding control to a panel would be trivial, but FlowLayoutPanel will auto-align the newly added control to the next available position, which may screw up the expected effect.

Video Demo of the Solution: Remember to Switch to 1080p
Recorded in a VM on a crappy machine. So kinda slow.
You are getting those artifacts because you're doing a combination of 3 things all at once.
The two big ones are moving the form to another location and resizing the form. It also doesn't help if the form is semi transparent :) To get a better understanding of what I mean, just open VS2013 up and resize the window very quickly (at the top-left corner, and run in random directions really fast), you will see that around the edges it can't keep up. And yes, you will get different results when you're resizing from a different position around the window (just think about it for a minute and you will figure it out).
Aybe, provided a pretty clever solution but it doesn't allow you to see through it or see if any updates to the panel....since it basically just copies the last output to a bitmap and uses that as a back buffer (much like what you assume someone might do when doing the selection thing in a paint program).
If you really want to do it with an overlay form and keep it semi-transparent then you will need to eliminate those three things if you don't want artifacts.
The code requires quite a bit of WIN32 knowledge.... lucky for you Microsoft has already done the hard part. We are going to enable per pixel transparency in your cover frame by using the PerPixelAlphaForm by Microsoft (you can google it) I will paste the code here. It basically just creates a Window with a Style of WS_EX_LAYERED. Keeps a Backbuffer which is AlphaBlended with the screen (simple huh?).
/******************************** Module Header ********************************\
Module Name: PerPixelAlphaForm.cs
Project: CSWinFormLayeredWindow
Copyright (c) Microsoft Corporation.
This source is subject to the Microsoft Public License.
See http://www.microsoft.com/opensource/licenses.mspx#Ms-PL.
All other rights reserved.
THIS CODE AND INFORMATION IS PROVIDED "AS IS" WITHOUT WARRANTY OF ANY KIND,
EITHER EXPRESSED OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE IMPLIED WARRANTIES
OF MERCHANTABILITY AND/OR FITNESS FOR A PARTICULAR PURPOSE.
\*******************************************************************************/
#region Using directives
using System;
using System.Collections.Generic;
using System.ComponentModel;
using System.Data;
using System.Drawing;
using System.Text;
using System.Windows.Forms;
using System.Drawing.Imaging;
using System.Runtime.InteropServices;
#endregion
namespace CSWinFormLayeredWindow
{
public partial class PerPixelAlphaForm : Form
{
public PerPixelAlphaForm()
{
InitializeComponent();
}
protected override CreateParams CreateParams
{
get
{
// Add the layered extended style (WS_EX_LAYERED) to this window.
CreateParams createParams = base.CreateParams;
createParams.ExStyle |= WS_EX_LAYERED;
return createParams;
}
}
/// <summary>
/// Let Windows drag this window for us (thinks its hitting the title
/// bar of the window)
/// </summary>
/// <param name="message"></param>
protected override void WndProc(ref Message message)
{
if (message.Msg == WM_NCHITTEST)
{
// Tell Windows that the user is on the title bar (caption)
message.Result = (IntPtr)HTCAPTION;
}
else
{
base.WndProc(ref message);
}
}
/// <summary>
///
/// </summary>
/// <param name="bitmap"></param>
public void SelectBitmap(Bitmap bitmap)
{
SelectBitmap(bitmap, 255);
}
/// <summary>
///
/// </summary>
/// <param name="bitmap">
///
/// </param>
/// <param name="opacity">
/// Specifies an alpha transparency value to be used on the entire source
/// bitmap. The SourceConstantAlpha value is combined with any per-pixel
/// alpha values in the source bitmap. The value ranges from 0 to 255. If
/// you set SourceConstantAlpha to 0, it is assumed that your image is
/// transparent. When you only want to use per-pixel alpha values, set
/// the SourceConstantAlpha value to 255 (opaque).
/// </param>
public void SelectBitmap(Bitmap bitmap, int opacity)
{
// Does this bitmap contain an alpha channel?
if (bitmap.PixelFormat != PixelFormat.Format32bppArgb)
{
throw new ApplicationException("The bitmap must be 32bpp with alpha-channel.");
}
// Get device contexts
IntPtr screenDc = GetDC(IntPtr.Zero);
IntPtr memDc = CreateCompatibleDC(screenDc);
IntPtr hBitmap = IntPtr.Zero;
IntPtr hOldBitmap = IntPtr.Zero;
try
{
// Get handle to the new bitmap and select it into the current
// device context.
hBitmap = bitmap.GetHbitmap(Color.FromArgb(0));
hOldBitmap = SelectObject(memDc, hBitmap);
// Set parameters for layered window update.
Size newSize = new Size(bitmap.Width, bitmap.Height);
Point sourceLocation = new Point(0, 0);
Point newLocation = new Point(this.Left, this.Top);
BLENDFUNCTION blend = new BLENDFUNCTION();
blend.BlendOp = AC_SRC_OVER;
blend.BlendFlags = 0;
blend.SourceConstantAlpha = (byte)opacity;
blend.AlphaFormat = AC_SRC_ALPHA;
// Update the window.
UpdateLayeredWindow(
this.Handle, // Handle to the layered window
screenDc, // Handle to the screen DC
ref newLocation, // New screen position of the layered window
ref newSize, // New size of the layered window
memDc, // Handle to the layered window surface DC
ref sourceLocation, // Location of the layer in the DC
0, // Color key of the layered window
ref blend, // Transparency of the layered window
ULW_ALPHA // Use blend as the blend function
);
}
finally
{
// Release device context.
ReleaseDC(IntPtr.Zero, screenDc);
if (hBitmap != IntPtr.Zero)
{
SelectObject(memDc, hOldBitmap);
DeleteObject(hBitmap);
}
DeleteDC(memDc);
}
}
#region Native Methods and Structures
const Int32 WS_EX_LAYERED = 0x80000;
const Int32 HTCAPTION = 0x02;
const Int32 WM_NCHITTEST = 0x84;
const Int32 ULW_ALPHA = 0x02;
const byte AC_SRC_OVER = 0x00;
const byte AC_SRC_ALPHA = 0x01;
[StructLayout(LayoutKind.Sequential)]
struct Point
{
public Int32 x;
public Int32 y;
public Point(Int32 x, Int32 y)
{ this.x = x; this.y = y; }
}
[StructLayout(LayoutKind.Sequential)]
struct Size
{
public Int32 cx;
public Int32 cy;
public Size(Int32 cx, Int32 cy)
{ this.cx = cx; this.cy = cy; }
}
[StructLayout(LayoutKind.Sequential, Pack = 1)]
struct ARGB
{
public byte Blue;
public byte Green;
public byte Red;
public byte Alpha;
}
[StructLayout(LayoutKind.Sequential, Pack = 1)]
struct BLENDFUNCTION
{
public byte BlendOp;
public byte BlendFlags;
public byte SourceConstantAlpha;
public byte AlphaFormat;
}
[DllImport("user32.dll", CharSet = CharSet.Auto, SetLastError = true)]
[return: MarshalAs(UnmanagedType.Bool)]
static extern bool UpdateLayeredWindow(IntPtr hwnd, IntPtr hdcDst,
ref Point pptDst, ref Size psize, IntPtr hdcSrc, ref Point pprSrc,
Int32 crKey, ref BLENDFUNCTION pblend, Int32 dwFlags);
[DllImport("gdi32.dll", CharSet = CharSet.Auto, SetLastError = true)]
static extern IntPtr CreateCompatibleDC(IntPtr hDC);
[DllImport("user32.dll", CharSet = CharSet.Auto, SetLastError = true)]
static extern IntPtr GetDC(IntPtr hWnd);
[DllImport("user32.dll", CharSet = CharSet.Auto, SetLastError = true)]
static extern int ReleaseDC(IntPtr hWnd, IntPtr hDC);
[DllImport("gdi32.dll", CharSet = CharSet.Auto, SetLastError = true)]
[return: MarshalAs(UnmanagedType.Bool)]
static extern bool DeleteDC(IntPtr hdc);
[DllImport("gdi32.dll", CharSet = CharSet.Auto, SetLastError = true)]
static extern IntPtr SelectObject(IntPtr hDC, IntPtr hObject);
[DllImport("gdi32.dll", CharSet = CharSet.Auto, SetLastError = true)]
[return: MarshalAs(UnmanagedType.Bool)]
static extern bool DeleteObject(IntPtr hObject);
#endregion
}
}
OK, that should eliminate your semi-transparent problem. Remember to get rid of the override of the WndProc (you won't need it). Set Double-Buffer to false and TopMost to true.
Now to eliminate the other two problems. I hope you thought of a way of doing it....but I will give you my solution. Always keep the PerPixelAlphaForm the size of your MainForm. Same location, Same SIZE. :) And resize the PerPixelAlphaForm's backbuffer bitmap to the same size as well. When you do it this way, all you have to do is redraw the Selection Rectangle. Why? because it overlays the entire MainForm perfectly.
So basically
`OnMouseDown` = Save initial point of mouse, show the Cover layer
`OnMouseMove` = clear the PerPixelAlphaForm bitmap, draw your rectangle
call SelectBitmap again update the form
`OnMouseUp` = hide the Cover layer (or whatever you want to do)
I personally have all this hook up to the Control-Key
To clear the PerPixelAlphaForm we need to do in a certain way. Give all values an Alpha of 0.
public void ClearBackbuffer()
{
Graphics g = Graphics.FromImage(_reference_to_your_backbuffer_);
g.CompositingMode = System.Drawing.Drawing2D.CompositingMode.SourceCopy;
SolidBrush sb = new SolidBrush(Color.FromArgb(0x00, 0x00, 0x00, 0x00));
g.FillRectangle(sb, this.ClientRectangle);
sb.Dispose();
g.Dispose();
}
Video Demo of the Solution: Remember to Switch to 1080p
If you need more help, let me know I can find some time to rip the code out of the larger program. But it seems to me you're the kind of person that likes tinkering with stuff :D

EDIT : using an additional PictureBox and Bitmap makes the whole thing working
The following Panel draws a rectangle without flickering:
internal sealed class MyPanel : Panel
{
private readonly PictureBox _pictureBox;
private Bitmap _bitmapContent;
private Bitmap _bitmapForeground;
private Point? _point1;
private Point? _point2;
public MyPanel()
{
DoubleBuffered = true;
_pictureBox = new PictureBox();
}
protected override void OnSizeChanged(EventArgs e)
{
if (_bitmapForeground != null) _bitmapForeground.Dispose();
_bitmapForeground = new Bitmap(Size.Width, Size.Height);
if (_bitmapContent != null) _bitmapContent.Dispose();
_bitmapContent = new Bitmap(Size.Width, Size.Height);
_pictureBox.Size = Size;
_pictureBox.Image = _bitmapForeground;
base.OnSizeChanged(e);
}
protected override void OnMouseDown(MouseEventArgs e)
{
_point1 = e.Location;
DrawToBitmap(_bitmapContent, new Rectangle(0, 0, Size.Width, Size.Height));
SetControlsVisibility(false);
Controls.Add(_pictureBox);
base.OnMouseDown(e);
}
private void SetControlsVisibility(bool visible)
{
IEnumerable<Control> ofType = Controls.OfType<Control>();
foreach (Control control in ofType)
{
control.Visible = visible;
}
}
protected override void OnMouseUp(MouseEventArgs e)
{
Controls.Remove(_pictureBox);
SetControlsVisibility(true);
_point1 = null;
_point2 = null;
Refresh();
base.OnMouseUp(e);
}
protected override void OnMouseMove(MouseEventArgs e)
{
if (_point1 != null)
{
_point2 = e.Location;
if (_point1 != null && _point2 != null)
{
Point p1 = _point1.Value;
Point p2 = _point2.Value;
int x1 = p1.X;
int y1 = p1.Y;
int x2 = p2.X;
int y2 = p2.Y;
int xmin = Math.Min(x1, x2);
int ymin = Math.Min(y1, y2);
int xmax = Math.Max(x1, x2);
int ymax = Math.Max(y1, y2);
using (Graphics graphics = Graphics.FromImage(_bitmapForeground))
{
graphics.DrawImageUnscaled(_bitmapContent, 0, 0, _bitmapContent.Width, _bitmapContent.Height);
graphics.DrawRectangle(Pens.Red, new Rectangle(xmin, ymin, xmax - xmin, ymax - ymin));
}
_pictureBox.Refresh();
}
}
base.OnMouseMove(e);
}
}
However, the rectangle will be below the controls, not sure why ...

C# - Capturing the Mouse cursor image

BACKGROUND
I am writing a screen capture application
My code is based derived from this project: http://www.codeproject.com/KB/cs/DesktopCaptureWithMouse.aspx?display=Print
Note that the code captures the the mouse cursor also (which is desirable for me)
MY PROBLEM
Code works fine when the mouse cursor is the normal pointer or hand icon - the mouse is rendered correctly on the screenshot
However, when the mouse cursor is changed to the insertion point (the "I-beam" cursor) - for example typing in NOTEPAD - then code doesn't work - the result is that I get a faint image of the cursor - like a very translucent (gray) version of it instead of the blank & white one would expect.
MY QUESTION
How can I capture the mouse cursor image when the image is one of these "I-beam"-type images
NOTE: If you click on the original article someone offers a suggestion - it doesn't work
SOURCE
This is from the original article.
static Bitmap CaptureCursor(ref int x, ref int y)
{
Bitmap bmp;
IntPtr hicon;
Win32Stuff.CURSORINFO ci = new Win32Stuff.CURSORINFO();
Win32Stuff.ICONINFO icInfo;
ci.cbSize = Marshal.SizeOf(ci);
if (Win32Stuff.GetCursorInfo(out ci))
{
if (ci.flags == Win32Stuff.CURSOR_SHOWING)
{
hicon = Win32Stuff.CopyIcon(ci.hCursor);
if (Win32Stuff.GetIconInfo(hicon, out icInfo))
{
x = ci.ptScreenPos.x - ((int)icInfo.xHotspot);
y = ci.ptScreenPos.y - ((int)icInfo.yHotspot);
Icon ic = Icon.FromHandle(hicon);
bmp = ic.ToBitmap();
return bmp;
}
}
}
return null;
}

While I can't explain exactly why this happens, I think I can show how to get around it.
The ICONINFO struct contains two members, hbmMask and hbmColor, that contain the mask and color bitmaps, respectively, for the cursor (see the MSDN page for ICONINFO for the official documentation).
When you call GetIconInfo() for the default cursor, the ICONINFO struct contains both valid mask and color bitmaps, as shown below (Note: the red border has been added to clearly show the image boundaries):
Default Cursor Mask Bitmap
Default Cursor Color Bitmap
When Windows draws the default cursor, the mask bitmap is first applied with an AND raster operation, then the color bitmap is applied with an XOR raster operation. This results in an opaque cursor and a transparent background.
When you call GetIconInfo() for the I-Beam cursor, though, the ICONINFO struct only contains a valid mask bitmap, and no color bitmap, as shown below (Note: again, the red border has been added to clearly show the image boundaries):
I-Beam Cursor Mask Bitmap
According to the ICONINFO documentation, the I-Beam cursor is then a monochrome cursor. The top half of the mask bitmap is the AND mask, and the bottom half of the mask bitmap is the XOR bitmap. When Windows draws the I-Beam cursor, the top half of this bitmap is first drawn over the desktop with an AND raster operation. The bottom half of the bitmap is then drawn over top with an XOR raster operation. Onscreen, The cursor will appear as the inverse of the content behind it.
One of the comments for the original article that you linked mentions this. On the desktop, since the raster operations are applied over the desktop content, the cursor will appear correct. However, when the image is drawn over no background, as in your posted code, the raster operations that Windows performs result in a faded image.
That being said, this updated CaptureCursor() method will handle both color and monochrome cursors, supplying a plain black cursor image when the cursor is monochrome.
static Bitmap CaptureCursor(ref int x, ref int y)
{
Win32Stuff.CURSORINFO cursorInfo = new Win32Stuff.CURSORINFO();
cursorInfo.cbSize = Marshal.SizeOf(cursorInfo);
if (!Win32Stuff.GetCursorInfo(out cursorInfo))
return null;
if (cursorInfo.flags != Win32Stuff.CURSOR_SHOWING)
return null;
IntPtr hicon = Win32Stuff.CopyIcon(cursorInfo.hCursor);
if (hicon == IntPtr.Zero)
return null;
Win32Stuff.ICONINFO iconInfo;
if (!Win32Stuff.GetIconInfo(hicon, out iconInfo))
return null;
x = cursorInfo.ptScreenPos.x - ((int)iconInfo.xHotspot);
y = cursorInfo.ptScreenPos.y - ((int)iconInfo.yHotspot);
using (Bitmap maskBitmap = Bitmap.FromHbitmap(iconInfo.hbmMask))
{
// Is this a monochrome cursor?
if (maskBitmap.Height == maskBitmap.Width * 2)
{
Bitmap resultBitmap = new Bitmap(maskBitmap.Width, maskBitmap.Width);
Graphics desktopGraphics = Graphics.FromHwnd(Win32Stuff.GetDesktopWindow());
IntPtr desktopHdc = desktopGraphics.GetHdc();
IntPtr maskHdc = Win32Stuff.CreateCompatibleDC(desktopHdc);
IntPtr oldPtr = Win32Stuff.SelectObject(maskHdc, maskBitmap.GetHbitmap());
using (Graphics resultGraphics = Graphics.FromImage(resultBitmap))
{
IntPtr resultHdc = resultGraphics.GetHdc();
// These two operation will result in a black cursor over a white background.
// Later in the code, a call to MakeTransparent() will get rid of the white background.
Win32Stuff.BitBlt(resultHdc, 0, 0, 32, 32, maskHdc, 0, 32, Win32Stuff.TernaryRasterOperations.SRCCOPY);
Win32Stuff.BitBlt(resultHdc, 0, 0, 32, 32, maskHdc, 0, 0, Win32Stuff.TernaryRasterOperations.SRCINVERT);
resultGraphics.ReleaseHdc(resultHdc);
}
IntPtr newPtr = Win32Stuff.SelectObject(maskHdc, oldPtr);
Win32Stuff.DeleteObject(newPtr);
Win32Stuff.DeleteDC(maskHdc);
desktopGraphics.ReleaseHdc(desktopHdc);
// Remove the white background from the BitBlt calls,
// resulting in a black cursor over a transparent background.
resultBitmap.MakeTransparent(Color.White);
return resultBitmap;
}
}
Icon icon = Icon.FromHandle(hicon);
return icon.ToBitmap();
}
There are some issues with the code that may or may not be a problem.
The check for a monochrome cursor simply tests whether the height is twice the width. While this seems logical, the ICONINFO documentation does not mandate that only a monochrome cursor is defined by this.
There is probably a better way to render the cursor that the BitBlt() - BitBlt() - MakeTransparent() combination of method calls I used.

[StructLayout(LayoutKind.Sequential)]
struct CURSORINFO
{
public Int32 cbSize;
public Int32 flags;
public IntPtr hCursor;
public POINTAPI ptScreenPos;
}
[StructLayout(LayoutKind.Sequential)]
struct POINTAPI
{
public int x;
public int y;
}
[DllImport("user32.dll")]
static extern bool GetCursorInfo(out CURSORINFO pci);
[DllImport("user32.dll")]
static extern bool DrawIcon(IntPtr hDC, int X, int Y, IntPtr hIcon);
const Int32 CURSOR_SHOWING = 0x00000001;
public static Bitmap CaptureScreen(bool CaptureMouse)
{
Bitmap result = new Bitmap(Screen.PrimaryScreen.Bounds.Width, Screen.PrimaryScreen.Bounds.Height, PixelFormat.Format24bppRgb);
try
{
using (Graphics g = Graphics.FromImage(result))
{
g.CopyFromScreen(0, 0, 0, 0, Screen.PrimaryScreen.Bounds.Size, CopyPixelOperation.SourceCopy);
if (CaptureMouse)
{
CURSORINFO pci;
pci.cbSize = System.Runtime.InteropServices.Marshal.SizeOf(typeof(CURSORINFO));
if (GetCursorInfo(out pci))
{
if (pci.flags == CURSOR_SHOWING)
{
DrawIcon(g.GetHdc(), pci.ptScreenPos.x, pci.ptScreenPos.y, pci.hCursor);
g.ReleaseHdc();
}
}
}
}
}
catch
{
result = null;
}
return result;
}

Here's a modified version of Dimitar's response (using DrawIconEx) that worked for me on multiple screens:
public class ScreenCapturePInvoke
{
[StructLayout(LayoutKind.Sequential)]
private struct CURSORINFO
{
public Int32 cbSize;
public Int32 flags;
public IntPtr hCursor;
public POINTAPI ptScreenPos;
}
[StructLayout(LayoutKind.Sequential)]
private struct POINTAPI
{
public int x;
public int y;
}
[DllImport("user32.dll")]
private static extern bool GetCursorInfo(out CURSORINFO pci);
[DllImport("user32.dll", SetLastError = true)]
static extern bool DrawIconEx(IntPtr hdc, int xLeft, int yTop, IntPtr hIcon, int cxWidth, int cyHeight, int istepIfAniCur, IntPtr hbrFlickerFreeDraw, int diFlags);
private const Int32 CURSOR_SHOWING = 0x0001;
private const Int32 DI_NORMAL = 0x0003;
public static Bitmap CaptureFullScreen(bool captureMouse)
{
var allBounds = Screen.AllScreens.Select(s => s.Bounds).ToArray();
Rectangle bounds = Rectangle.FromLTRB(allBounds.Min(b => b.Left), allBounds.Min(b => b.Top), allBounds.Max(b => b.Right), allBounds.Max(b => b.Bottom));
var bitmap = CaptureScreen(bounds, captureMouse);
return bitmap;
}
public static Bitmap CapturePrimaryScreen(bool captureMouse)
{
Rectangle bounds = Screen.PrimaryScreen.Bounds;
var bitmap = CaptureScreen(bounds, captureMouse);
return bitmap;
}
public static Bitmap CaptureScreen(Rectangle bounds, bool captureMouse)
{
Bitmap result = new Bitmap(bounds.Width, bounds.Height);
try
{
using (Graphics g = Graphics.FromImage(result))
{
g.CopyFromScreen(bounds.Location, Point.Empty, bounds.Size);
if (captureMouse)
{
CURSORINFO pci;
pci.cbSize = Marshal.SizeOf(typeof (CURSORINFO));
if (GetCursorInfo(out pci))
{
if (pci.flags == CURSOR_SHOWING)
{
var hdc = g.GetHdc();
DrawIconEx(hdc, pci.ptScreenPos.x-bounds.X, pci.ptScreenPos.y-bounds.Y, pci.hCursor, 0, 0, 0, IntPtr.Zero, DI_NORMAL);
g.ReleaseHdc();
}
}
}
}
}
catch
{
result = null;
}
return result;
}
}

Based on the other answers I made a version without all the Windows API stuff (for the monochrome part) because the solutions did not work for all monochrome cursors. I create the cursor from the mask by combining the two mask parts.
My solution:
Bitmap CaptureCursor(ref Point position)
{
CURSORINFO cursorInfo = new CURSORINFO();
cursorInfo.cbSize = Marshal.SizeOf(cursorInfo);
if (!GetCursorInfo(out cursorInfo))
return null;
if (cursorInfo.flags != CURSOR_SHOWING)
return null;
IntPtr hicon = CopyIcon(cursorInfo.hCursor);
if (hicon == IntPtr.Zero)
return null;
ICONINFO iconInfo;
if (!GetIconInfo(hicon, out iconInfo))
return null;
position.X = cursorInfo.ptScreenPos.x - iconInfo.xHotspot;
position.Y = cursorInfo.ptScreenPos.y - iconInfo.yHotspot;
using (Bitmap maskBitmap = Bitmap.FromHbitmap(iconInfo.hbmMask))
{
// check for monochrome cursor
if (maskBitmap.Height == maskBitmap.Width * 2)
{
Bitmap cursor = new Bitmap(32, 32, PixelFormat.Format32bppArgb);
Color BLACK = Color.FromArgb(255, 0, 0, 0); //cannot compare Color.Black because of different names
Color WHITE = Color.FromArgb(255, 255, 255, 255); //cannot compare Color.White because of different names
for (int y = 0; y < 32; y++)
{
for (int x = 0; x < 32; x++)
{
Color maskPixel = maskBitmap.GetPixel(x, y);
Color cursorPixel = maskBitmap.GetPixel(x, y + 32);
if (maskPixel == WHITE && cursorPixel == BLACK)
{
cursor.SetPixel(x, y, Color.Transparent);
}
else if (maskPixel == BLACK)
{
cursor.SetPixel(x, y, cursorPixel);
}
else
{
cursor.SetPixel(x, y, cursorPixel == BLACK ? WHITE : BLACK);
}
}
}
return cursor;
}
}
Icon icon = Icon.FromHandle(hicon);
return icon.ToBitmap();
}

This is the patched version with all fixes for the bugs presented on this page:
public static Bitmap CaptureImageCursor(ref Point point)
{
try
{
var cursorInfo = new CursorInfo();
cursorInfo.cbSize = Marshal.SizeOf(cursorInfo);
if (!GetCursorInfo(out cursorInfo))
return null;
if (cursorInfo.flags != CursorShowing)
return null;
var hicon = CopyIcon(cursorInfo.hCursor);
if (hicon == IntPtr.Zero)
return null;
Iconinfo iconInfo;
if (!GetIconInfo(hicon, out iconInfo))
{
DestroyIcon(hicon);
return null;
}
point.X = cursorInfo.ptScreenPos.X - iconInfo.xHotspot;
point.Y = cursorInfo.ptScreenPos.Y - iconInfo.yHotspot;
using (var maskBitmap = Image.FromHbitmap(iconInfo.hbmMask))
{
//Is this a monochrome cursor?
if (maskBitmap.Height == maskBitmap.Width * 2 && iconInfo.hbmColor == IntPtr.Zero)
{
var final = new Bitmap(maskBitmap.Width, maskBitmap.Width);
var hDesktop = GetDesktopWindow();
var dcDesktop = GetWindowDC(hDesktop);
using (var resultGraphics = Graphics.FromImage(final))
{
var resultHdc = resultGraphics.GetHdc();
BitBlt(resultHdc, 0, 0, final.Width, final.Height, dcDesktop, (int)point.X + 3, (int)point.Y + 3, CopyPixelOperation.SourceCopy);
DrawIconEx(resultHdc, 0, 0, cursorInfo.hCursor, 0, 0, 0, IntPtr.Zero, 0x0003);
//TODO: I have to try removing the background of this cursor capture.
//Native.BitBlt(resultHdc, 0, 0, final.Width, final.Height, dcDesktop, (int)point.X + 3, (int)point.Y + 3, Native.CopyPixelOperation.SourceErase);
resultGraphics.ReleaseHdc(resultHdc);
ReleaseDC(hDesktop, dcDesktop);
}
DeleteObject(iconInfo.hbmMask);
DeleteDC(dcDesktop);
DestroyIcon(hicon);
return final;
}
DeleteObject(iconInfo.hbmColor);
DeleteObject(iconInfo.hbmMask);
DestroyIcon(hicon);
}
var icon = Icon.FromHandle(hicon);
return icon.ToBitmap();
}
catch (Exception ex)
{
//You should catch exception with your method here.
//LogWriter.Log(ex, "Impossible to get the cursor.");
}
return null;
}
This version works with:
I-Beam cursors.
Black cursors.
Normal cursors.
Inverted cursors.
See working, here: https://github.com/NickeManarin/ScreenToGif/blob/master/ScreenToGif/Util/Native.cs#L991

Your description of a translucent 'gray' version of the I-beam cursor makes me wonder if you're encountering an issue with image scaling or mispositioning of the cursor.
One of the people posting on that site provided a (broken) link to a report with peculiar behavior that I've tracked down to: http://www.efg2.com/Lab/Graphics/CursorOverlay.htm
The examples on that page are not in C# but the author of the codeproject solution may have been doing something similar and I know I've screwed up my scaling when using the graphics object on plenty of occassions myself:
In any ImageMouseDown event once an
image is loaded, the CusorBitmap is
drawn with transparency on top of the
bitmap using the Canvas.Draw method.
Note some coordinate adjustments
(rescaling) are needed in case the
bitmap is stretched to fit in the
TImage.