Re: Point Grey Flea3 areaDetector ImageJ Viewer Problems

[Zachary Brown <zab8@cornell.edu>]

Re: Flycap:

We are using the USB3.0 interface on the FL3-U3-13YM-C camera.  In looking for the version number, I realized I installed the viewer, not the SDK, so I'm not sure if all of the functions in the SDK are included in the libraries for the viewer.  I will test first thing on Monday.

Thank you,

-ZB

On Sat, May 3, 2014 at 12:30 AM, Mark Rivers <rivers@cars.uchicago.edu> wrote:

Hi John,

OK so the camera claims to support embedded Frame counter data.

I am using the higher-level ImageMetadata SDK object to access the metadata, I am not extracting it directly from the Image object.  That way their SDK takes care of unpacking the metadata from the Image object and I don't need to worry about exactly how it is stored.

My code for the frame counter is working fine on the Flea2 (Firewire), BlackFly (GigE), and Grasshopper3 (USB 3.0) cameras.

What version of FlyCap2 have you installed on your Windows machine?  What interface are you using (USB 3.0, etc.)?

Mark



________________________________________
From: John Dobbins [john.dobbins@cornell.edu]
Sent: Friday, May 02, 2014 6:35 PM
To: Mark Rivers; Zachary A. Brown; tech-talk@aps.anl.gov

Subject: RE: Point Grey Flea3 areaDetector ImageJ Viewer Problems

Mark,

The Technical Reference Manual for the Flea3 (I'm looking at the manual for the USB3 version)  reads as follows:

>>>

The following frame-specific information can be provided:
    Timestamp
    Gain
    Shutter
    Brightness
    White Balance
    Frame counter
    Strobe Pattern counter
    GPIO pin state
    ROI position

If you turned on all possible options the first 40 bytes of image data would contain camera information in the following format, when accessed using the FlyCapture 2 API:

(assuming unsigned char* data = "" and an Image object rawImage):

n data[0] = first byte of Timestamp data
n data[4] = first byte of Gain data
n data[24] = first byte of Frame Counter data

If only Shutter embedding were enabled, then the first 4 bytes of the image would contain Shutter information for that image. Similarly, if only Brightness embedding were enabled, the first 4 bytes would contain Brightness information.

For monochrome cameras, white balance is still included, but no valid data is provided.

<<<

John


________________________________________
From: Mark Rivers <rivers@cars.uchicago.edu>
Sent: Friday, May 02, 2014 5:54 PM
To: John Dobbins; Zachary A. Brown; tech-talk@aps.anl.gov
Subject: RE: Point Grey Flea3 areaDetector ImageJ Viewer Problems

John and Zachary,

Both the ImageJ and IDL visualization tools provided with areaDetector put a CA monitor on the UniqueID PV.  When that PV changes the code does a caget from the ArrayData PV.  This is simpler than putting a monitor on the ArrayData itself, in which case the monitor would need to be cancelled and reissued each time the data type or size of the image changed.

The Point Grey driver is new.  I have tested it on a Flea2 camera, but not on a Flea3 camera.

In the constructor of the Point Grey driver I tell their SDK to embed a frame counter in each image:

    embeddedInfo.frameCounter._onOff_ = true;

In the function that reads the image and constructs the NDArray I copy that frame counter to the NDArray->uniqueID.

    pRaw_->uniqueId = metaData.embeddedFrameCounter;

That is working fine for Flea2, BlackFly, and Grasshopper3 cameras.  For some reason it does not seem to be working for your Flea3 camera.  Can you look at the hardware-specific documentation for your camera and see if it discusses what embedded image information that camera supports?

Mark


________________________________
From: tech-talk-bounces@aps.anl.gov [tech-talk-bounces@aps.anl.gov] on behalf of John Dobbins [john.dobbins@cornell.edu]
Sent: Friday, May 02, 2014 1:57 PM
To: Zachary A. Brown; tech-talk@aps.anl.gov
Subject: RE: Point Grey Flea3 areaDetector ImageJ Viewer Problems


​A quick look at the source code for ImageJ shows that it is looking for a UniqueID change, so I don't think ImageJ will work without the Unique ID updating.


John Dobbins


________________________________
From: tech-talk-bounces@aps.anl.gov <tech-talk-bounces@aps.anl.gov> on behalf of Zachary Brown <zab8@cornell.edu>
Sent: Friday, May 02, 2014 2:00 PM
To: tech-talk@aps.anl.gov
Subject: Re: Point Grey Flea3 areaDetector ImageJ Viewer Problems

From more testing:

A camonitor on PG1:image1:ArrayData shows that the values in the array are changing as the camera is moved around, so the array is getting updated.

ImageJ still shows a zero-value array.  The image shown from the plugin is 1280x1024 (8-bit) as it should be, but no pixel values.

The Array Counter in for the StdArray increases with the image count, but I did notice that the Unique ID for each array does not change (not sure if this is relevent, have not paid attention to this parameter before).

-Zachary Brown, John Dobbins


On Fri, May 2, 2014 at 1:02 PM, Zachary Brown <zab8@cornell.edu<mailto:zab8@cornell.edu>> wrote:
Hello,

I am using the new Point Grey pre-built in windows 7 x64.  I am having issues streaming the images to the ImageJ areaDetector plugin.

I am able to acquire images in all image modes (single, multiple, continuous) and save the images using the TIFF plugin, but the NDStdArrays plugin does not seem to be getting any data, so the callbacks do not update, so the ImageJ plugin does not stream.

PG1:cam1:ArrayCallbacks is set to Enable
PG1:image1:EnableCallbacks is set to Enable

If any more information would be helpful, please contact me, thank you,

-ZB