|
Size: 3618
Comment:
|
← Revision 15 as of 2012-01-03 23:02:40 ⇥
Size: 1693
Comment: converted to 1.6 markup
|
| Deletions are marked like this. | Additions are marked like this. |
| Line 1: | Line 1: |
| 1. Make GigE camera(s) work: full resolution, high dynamic range (all the bits), lower frame rate. | 1. Make GigE camera(s) work: full resolution, high dynamic range (all the bits), lower frame rate. |
| Line 4: | Line 5: |
| 1. 2-D Fit determines beam centroid, width, and intensity | |
| Line 6: | Line 8: |
| 1. TIFF files are roughly 0.5 MB per image, plain text is roughly 3.5 MB per image 1. Normal file compression (zip compression in OS 10.5) compressed files to roughly 1/4 their original size 1. At above rate of archiving and ~15 cameras per site, this is ~90000 images for one month and another ~67000 images for each following month. To fit a year's worth of data (~827385 images) on a 1 TB drive, this requires that each image be compressed to ~1.2 MB. Thus, the basic TIFF format is reasonable but could likely be improved upon. |
|
| Line 7: | Line 14: |
| 1. 2-D Fitting is too slow, center of mass method easily fast enough but may not be accurate enough 1. Need to look into EPICS control methods |
|
| Line 10: | Line 19: |
| ---- Current Status: 1. Prototype executable code "Snap" located in /cvs/cds/caltech/target/Prosilica/40mCode/ a. Has many options including Exposure Time, Gain, Region of Interest, Packet Size, file names, loop option to save images a. Only saves to .tiff format and no streaming a. Works on any of the linux machines (statically compiled code) a. C code lives in /cvs/cds/caltech/target/Prosilica/examples/Snap/ 1. SampleViewer code streams images a. Current bugs with controlling the camera - use Snap to work around for the moment a. Works only on mafalda at the moment (dynamically compiled currently, need to setup a statically compiled version) a. C code lives in /cvs/cds/caltech/target/Prosilica/examples/SampleViewer/ 1. Partly Gigabit Ethernet network a. One computer (Mafalda) with a GigE network card with packet sizes up to 1500 (Prosilica suggests jumbo packets of ~9000) a. Mafalda connected to GigE network hub in control room a. Have 1 portable GigE network switch along with several cat 6 cables (max length 30m at the moment, need some longer ones) |
|
| Line 26: | Line 20: |
| Things to be done this summer: 1. Develop code to manipulate the saved images a. Calculate beam size and location on ETMs or ITMs and be able to hand off this information to other code/scripts a. Use beam information in an auto-alignment script a. Image beam profiles and calculate Gaussian modes (Josh's beam scan stuff integrated with image taking) a. Be able to tell what mode a cavity is locked in (distinguish between TEM00 and TEM01 for instance) a. Integrate mode identification with appropriate scripts so as to use the information to get into the correct mode 1. Develop a robust camera network and associated code capable of handling many cameras, streaming, and the storage of images a. Setup up a robust Gigabit Ethernet network with appropriate hubs, cat 6 cabling, and GiGE network cards along with the cameras a. Ability to route the same image to multiple locations (multicasting to a live stream as well as the normal storage routines) a. Ability to select which cameras are streamed to the operators (to be able to replace analog cameras and monitors) Plus more to be added as with think of it. |
Subpages |
| Line 40: | Line 22: |
| -Joe | GigeCameraProgramRequirements |
| Line 42: | Line 24: |
| GigeCameraCurrentStatus | |
| Line 43: | Line 26: |
| ---- === Cameras === Web page with a bunch of GigE camera vendors (not all support Linux): http://www.pleora.com/camera_partners/index.php |
GigeCameraHardware |
| Line 47: | Line 28: |
| ---- I tried to capture frames with Snap from both cameras. It seems that the images from the 650 work less well than those from the 750 (image from the 650 is a bar across the top with some data, and then all black). While the -Exposure setting works well, there might be times in which a -Gain option is also nice. |
GigeCameraSurf |
| Line 50: | Line 30: |
| I also spent a little time with the cameras on the GC network. I was able to get about 10fps from both cameras at the simultaneously. The 750 alone maxed out at 33fps, even with 200x200 image size. The 650 was able to produce 188fps with a 200x200 image size. All of this was done with the packet size set to 1024b. | Presentation for 6/27/08 ISC meeting: [[attachment:DigitalCameraStatus062708.pdf]] |
| Line 52: | Line 33: |
| -Matt | Outline for July Progress Report (6/30/08): [[attachment:Eric_Mintun_LIGOSURF_ProgressOutline_063008.pdf]] Presentation for 9/5/08 ISC meeting: [[attachment:CameraEvaluation.pdf]] |
1. Make GigE camera(s) work: full resolution, high dynamic range (all the bits), lower frame rate.
- Explore required network bandwidth: how many GigE cameras can we have on single hub?
- Some image processing: beam sizes, beam centroids, total power, diffs,...
- 2-D Fit determines beam centroid, width, and intensity
- Image archiving: 1/sec for the past 5 minutes, 1/minute for the past day, 6/hour forever
- Image compression? Just record the diffs?
- TIFF files are roughly 0.5 MB per image, plain text is roughly 3.5 MB per image
- Normal file compression (zip compression in OS 10.5) compressed files to roughly 1/4 their original size
- At above rate of archiving and ~15 cameras per site, this is ~90000 images for one month and another ~67000 images for each following month. To fit a year's worth of data (~827385 images) on a 1 TB drive, this requires that each image be compressed to ~1.2 MB. Thus, the basic TIFF format is reasonable but could likely be improved upon.
- Use TM spots to do auto-alignment.
- 2-D Fitting is too slow, center of mass method easily fast enough but may not be accurate enough
- Need to look into EPICS control methods
- Port Josh's beam scan stuff to use these cameras.
- Buy more cameras.
Subpages
Presentation for 6/27/08 ISC meeting: DigitalCameraStatus062708.pdf
Outline for July Progress Report (6/30/08): Eric_Mintun_LIGOSURF_ProgressOutline_063008.pdf
Presentation for 9/5/08 ISC meeting: CameraEvaluation.pdf