How to load .png image into object of Bitmap class?

[ArsenShnurkov]

I want to avoid unmanaged code from libgdiplus and libpng in mono.

I.e. something like http://stackoverflow.com/a/6809677/1709408

public static void LoadThroughPngcs(this Bitmap bmp, string filename)
{
    if (bmp == null) throw new ArgumentNullException("bmp");
    // read bmpWidth, bmpHeight
    var data = bmp.LockBits(
        new Rectangle(0, 0, bmpWidth, bmpHeight),
        System.Drawing.Imaging.ImageLockMode.ReadWrite,
        System.Drawing.Imaging.PixelFormat.Format32bppArgb);

    var line = data.Scan0;
    var eof = line + data.Height * data.Stride;
    while(line != eof)
    {
        var pixelAlpha = line + 3;
        var eol = pixelAlpha + data.Width * 4;
        while(pixelAlpha != eol)
        {
            System.Runtime.InteropServices.Marshal.WriteByte(
                pixelAlpha, alpha);
            pixelAlpha += 4;
        }
        line += data.Stride;
    }
    bmp.UnlockBits(data);
}

but with reading pixels through libcs

[ArsenShnurkov]

May be something should have property of type IEnumerable<KeyValuePair<Point, Color> >. KeyValuePair instead of Pair, because the former is value type, and the later is reference type...

Or it is better to have a class which converts pixel sequential number into coordinates and vice versa.

[ArsenShnurkov]

   public static Bitmap LoadPng(string fileName)
    {
        PngReader pngr = FileHelper.CreatePngReader(fileName);
        pngr.SetUnpackedMode(true);
        int channels = pngr.ImgInfo.Channels;
        int bmpWidth = pngr.ImgInfo.Cols;
        int bmpHeight = pngr.ImgInfo.Rows;
        Bitmap bmp = new Bitmap(bmpWidth, bmpHeight, PixelFormat.Format32bppArgb);
        int A = 255;
        for (int y = 0; y < bmpHeight; y++)
        {
            ImageLine line = pngr.ReadRowInt(y);
            for (int x = 0; x < bmpWidth; x++)
            {
                int startByte = x * channels;
                int R = line.Scanline[startByte + 0];
                int G = line.Scanline[startByte + 1];
                int B = line.Scanline[startByte + 2];
                if (channels == 4)
                {
                    A = line.Scanline[startByte + 3];
                }
                Color c = Color.FromArgb(A, R, G, B);
                bmp.SetPixel(x, y, c);
            }
        }
        return bmp;
    }

My picture have 3 channels. How to get global transparency info in pngcs?
PngChunkTRNS trns = pngr.GetMetadata().GetTRNS(); // transparency metadata, can be null

pngr.ImgInfo {ImageInfo [cols=16, rows=16, bitDepth=8, channels=3, bitspPixel=24, bytesPixel=3, bytesPerRow=48, samplesPerRow=48, samplesPerRowP=48, alpha=False, greyscale=False, indexed=False, packed=False]} Hjg.Pngcs.ImageInfo

pngr.GetMetadata().GetTRNS() {chunk id= tRNS (len=6 off=33) c=PngChunkTRNS} Hjg.Pngcs.Chunks.PngChunkTRNS

So, I receive image without transparency at the end (lose alpha value of image)
may be this will work, may be not:

var trns = pngr.GetMetadata ().GetTRNS ();
if (trns != null) {
    int[] rgb = trns.GetRGB ();
    var tc = Color.FromArgb (0, rgb [0], rgb [1], rgb [2]);
    bmp.MakeTransparent (tc);
}