Image from a bytearray failure

Question

I have a function called: DisplayAndSaveImageFromByteArray.

by its name you probably understand what i am trying to do. In the bytearray the values are pixeldata. so like this 255,220,130,0, etc..

The size of the byte is the Width and the Height of the image times 4. because it works with strides.

public void DisplayAndSaveImageFromByteArray(byte[] byteArray)
{
    try{
        byte[] data = new byte[width * height * 4];
        int o = 0;

        for (int io = 0; io < width * height; io++){
            byte value = byteArray[io];

            data[o++] = value;
            data[o++] = value;
            data[o++] = value;
            data[o++] = 0;
        }
        unsafe
        {
            fixed (byte* ptr = data)
            {
                using (image = new Bitmap((int)width, (int)height, (int)width * 4,System.Drawing.Imaging.PixelFormat.Format32bppRgb, new IntPtr(ptr)))
                {
                    image.Save(@"c:\\testmap\" + nextpicture + ".jpg", ImageFormat.Jpeg);
                    if (nextpicture >= 10)
                    {
                        pbCameraPreview.BeginInvoke(new InvokeDelegate(InvokeMethod));
                    }
                    nextpicture++;
                }
            }
        }
    }
    catch (Exception ex){
        MessageBox.Show(ex.ToString());
    }

}

When i run this code it will work, but only if all values are the same for example: White (255,255,255) or Black (0,0,0). it is able to deviate about 5 up and down in the RGB(A) values until it will stop working. But as soon as the color of the image changes it will stop working without giving me an Exception of anything.

The only error/Exception i will get it if i leave it on for a minute and the VS will recognize that the code being executed is not doing anything and will give me a warning. > ContextSwitchDeadlock

What did i do wrong for it to crash? and what is the solution for it? for some reason it wont let me put on the using and namespace name... (updated)Complete code:

    public Form1()
    {
        InitializeComponent();
        pbCameraPreview.Image = defImg;
    }

    #region Global
    int ii;
    object __p1;
    EventArgs __p2;
    string path;
    Image defImg = Image.FromFile(@"c:\\testimg\def.jpg");
    UInt32 width;
    UInt32 height;
    int nextpicture = 0;
    FGNodeInfoContainer InfoContainer = new FGNodeInfoContainer();
    FGNodeInfo NodeInfo = new FGNodeInfo();
    FireWrap_CtrlCenter CtrlCenter;
    enFireWrapResult Result;
    UInt32 XSize = new UInt32();
    UInt32 YSize = new UInt32();
    UInt32 NodeCnt;
    public delegate void InvokeDelegate();
    Bitmap image;
    CameraCode Cam = new CameraCode();
    FGFrame Frame = new FGFrame();
    FGUIntHL Guid = new FGUIntHL();
    #endregion

    public void CheckDirectory()
    {
        path = @"c:\\testmap\" + ii + "\\";
        if (Directory.Exists(@"c:\\testmap\") == false)
        {
            Directory.CreateDirectory(@"c:\\testmap\");
        }

        if (Directory.Exists(path))
        {
            if (File.Exists(path + "0.Jpeg"))
            {
                ii++;
                CheckDirectory();
            }
        }
        else
        {
            Directory.CreateDirectory(path);
        }
    }

    //Haal de images op
    /// <param name="__p1"></param>
    /// <param name="__p2"></param>
    public void btStart_Click(object sender, EventArgs e)
    {
        Debug.WriteLine("btStart_Click is clicked");
        // Init module
        CtrlCenter = FireWrap_CtrlCenter.GetInstance();
        Result = CtrlCenter.FGInitModule();

        // Register frame start event
        CtrlCenter.OnFrameReady += new FireWrap_CtrlCenter.FireWrapEvent(OnFrameReady);

        // Get list of connected nodes
        if (Result == enFireWrapResult.E_NOERROR)
        {
            Result = InfoContainer.FGGetNodeList();
            NodeCnt = InfoContainer.Size();

            // Print Nodecnt
            Console.WriteLine(NodeCnt.ToString() + " camera found");

            // Connect with first node
            InfoContainer.GetAt(NodeInfo, 0);
            Result = Cam.Connect(NodeInfo.Guid);
            if (Result == enFireWrapResult.E_NOERROR)
            {
                Cam.m_Guid = NodeInfo.Guid;
            }

            // Set Format7 Mode0 Y8
            if (Result == enFireWrapResult.E_NOERROR)
            {
                Result = Cam.SetParameter(enFGParameter.E_IMAGEFORMAT,
                        (uint)(((uint)enFGResolution.E_RES_SCALABLE << 16) |
                            ((uint)enColorMode.E_CCOLORMODE_Y8 << 8) |
                                0));
            }

            if (Result != enFireWrapResult.E_NOERROR)
            {
                Result = Cam.SetParameter(enFGParameter.E_IMAGEFORMAT,
                            (uint)(((uint)enFGResolution.E_RES_SCALABLE << 16) |
                                ((uint)enColorMode.E_CCOLORMODE_Y8 << 8) |
                                    1));
            }

            // Start DMA logic
            if (Result == enFireWrapResult.E_NOERROR)
                Result = Cam.OpenCapture();

            // Print device settings
            Result = Cam.GetParameter(enFGParameter.E_XSIZE, ref XSize);
            Result = Cam.GetParameter(enFGParameter.E_YSIZE, ref YSize);
            Debug.WriteLine(Cam.DeviceAll + " [" + Cam.m_Guid.Low.ToString() + "] " + XSize + "x" + YSize);
            width = XSize;
            height = YSize;

            // Start camera
            if (Result == enFireWrapResult.E_NOERROR)
            {
                Result = Cam.StartDevice();
            }          
        }
    }

    public void btStop_Click(object sender, EventArgs e)
    {
        // Stop the device
        Cam.StopDevice();
        // Close capture
        Cam.CloseCapture();
        // Disconnect before ExitModule
        Cam.Disconnect();
        // Exit module
        CtrlCenter.FGExitModule();
    }

    /// <param name="__p1"></param>
    /// <param name="__p2"></param>
    public void OnFrameReady(object __p1,  EventArgs __p2)
    {
        Debug.WriteLine("OnFrameReady is called");
        FGEventArgs args = (FGEventArgs)__p2;
        Guid.High = args.High;
        Guid.Low = args.Low;

        if (Guid.Low == Cam.m_Guid.Low)
        {
            Result = Cam.GetFrame(Frame, 0);
            // Process frame, skip FrameStart notification
            if (Result == enFireWrapResult.E_NOERROR & Frame.Length > 0)
            {
                byte[] data = new byte[Frame.Length];

                // Access to frame data
                if (Frame.CloneData(data))
                {
                    DisplayAndSaveImageFromByteArray(data);
                    // Here you can start your image processsing logic on data
                    string debug = String.Format("[{6}] Frame #{0} length:{1}byte [ {2} {3} {4} {5} ... ]",
                        Frame.Id, Frame.Length, data[0], data[1], data[2], data[3], Cam.m_Guid.Low);
                    Debug.WriteLine(debug);
                }
                // Return frame to module as fast as posible after this the Frame is not valid 
                Result = Cam.PutFrame(Frame);
            }
        }

    }

    public void DisplayAndSaveImageFromByteArray(byte[] byteArray)
    {
        try{
            byte[] data = new byte[width * height * 4];
            int o = 0;

            for (int io = 0; io < width * height; io++){
                byte value = byteArray[io];

                data[o++] = value;
                data[o++] = value;
                data[o++] = value;
                data[o++] = 0;
            }
            unsafe
            {
                fixed (byte* ptr = data)
                {
                    using (image = new Bitmap((int)width, (int)height, (int)width * 4,System.Drawing.Imaging.PixelFormat.Format32bppRgb, new IntPtr(ptr)))
                    {
                        image.Save(@"c:\\testmap\" + nextpicture + ".jpg", ImageFormat.Jpeg);
                        if (nextpicture >= 10)
                        {
                            pbCameraPreview.BeginInvoke(new InvokeDelegate(InvokeMethod));
                        }
                        nextpicture++;
                    }
                }
            }
        }
        catch (Exception ex){
            MessageBox.Show(ex.ToString());
        }

    }
    public void InvokeMethod()
    {
        pbCameraPreview.Image = Image.FromFile(@"c:\\testmap\" + (nextpicture -10) + ".jpg");
    }

}
public class CameraCode : FireWrap_Camera
{
    public FGUIntHL m_Guid;
}}

Threads running: I recorded it for extra information: https://www.youtube.com/watch?v=i3TxWRyZaIU

Answer 1

I'm not 100% sure that I have understood your problem, since it is not very clear the format of the input array and how do you have to format it before parsing it into the Bitmap variable... But here we go, I hope these tips can help you. If they don't, please try to provide some extra details on what you are trying to do.

First of all, if I have understood well, you should increase "io" and update the variable "value" each time you assign it to data[o++] in the main loop, otherwise you are assigning the same value to R, G and B pixels, which will always result in a shade of gray.

Secondly, I see a couple things in your code that are not very .net-ly... .Net provides already ways to load an image from a byte array, using Memory Streams and stuff. Take a look at How to create bitmap from byte array?

And be sure to indicate the proper format of your byte array image when instantiating the Bitmap or Image --> https://msdn.microsoft.com/en-us/library/system.drawing.imaging.pixelformat(v=vs.110).aspx

Regards.

Answer 2

• If you have trouble posting your code in StackOverflow, make sure each line is indented by at least 4 spaces (mark the code in Visual Studio, press Tab, then copy it) and separated by any other paragraphs by at least one empty line. Alternatively add a line of three backticks (`) at the beginning and and of the code.

• When looking at your code, it seems to me that you are using a third-party control inside a Windows Form and try to bind the event of it to one of your Windows Forms event handlers. In that case you have to be aware that Windows Forms expect that all events are handled single-threaded (in the Event Dispatch Thread), so you should check (Debug.WriteLine) in your OnFrameReady method if InvokeRequired is true and if yes you have to take a few precautions:

  • Never access any of the Form's internal members (like pbCameraPreview) without wrapping the call into Invoke or BeginInvoke. Keep in mind that every Invoke call will effectively block until the single event-dispatch thread is available, so it will cost you a lot of performance to do invoke synchronously.

  • When accessing your own members (like width, height or nextpicture), make sure you use appropriate locking to avoid situations where one thread/callback changes the value in a situation where you don't expect it. In particular, since you have multiple cams but only a single width/height variable, if the resolutions differ, one camera callback could change the width while the other camera callback has just allocated the byte array but before passing the width to the Bitmap constructor, most likely resulting in a hang, crash, or memory access violation. Avoid this by passing width and heigth as variables into your method instead of using the global ones.

• Your form's event handler (btStart_Click) contains a busy-wait loop. This has multiple problems:

  • As stated before, every Form event handler runs in the same thread (and every Invoke as well). So as long as your loop is busy-waiting, no other events can be handled and the UI will be completely frozen. If code uses Invoke, that code will also eventually have to wait and be blocked, too.
  • Busy wait loops without any sleep code in it will cause your processor to run at the full cpu speed and it will eat 100% of that core, causing battery drain on notebooks and high power consumption and probably loud fans on other machines. Just don't do that.

In your concrete example, just remove the loop and move everything after the loop into the stop button instead. To avoid the problems with Windows Forms and invoking, I'd suggest to put all the image handling logic into a separate class that does its own locking (each instance of the class handling one camera only), and just call into it to start the process. You might even prefer to do your app as a Console application first to avoid any event-dispatching issues and later convert it to a Windows Forms application once it is working as you desire.