Project Diary

This is just an informal way to keep the team appraised of your projects and progress. We're growing, we're distributed, and we're collaborating. So, tell us what you're working on, finished, or starting!

Older items can be found in:

  • ScottLawrence - 2007 March 16
    • Wow... it's been a while
    • I just created the serial logger application; in CVS in the ser/ directory.
      • it assumes 9600 8N1, which should be fine
      • It currently saves to either a file specified on the command line or to a file with the current date/time as its name
      • It only prints out an "R" for each line of text received. It is an easy change to have it display the received text too.

  • ScottLawrence - 2007 Jan 17
    • The currently checked in version of WaspTracker? now supports:
      • ground-based GPS on or off (if "off", there's a pre-defined string to set your location)
      • antenna rotator support (again, on or off in the settings file.)
        • there's a "track" button on the top, and an infobox if it's "on"
        • it updates the rotator once a second or so, and if the amount to move is less than 2 degrees, it doesn't, to save wear on the motor and gears.
        • seems to work well.

  • ScottLawrence - 2006 Dec 13
    • The audio from Don Light's talk this morning is up on the SeminarSeries page
    • Wasp Tracker
      • I tracked down the bug in which the "home gps" was displaying in the incorrect location.
      • ENVI-HDR color boxes now are leg-color coded. Will test with the next flight.
      • due to memory issues, I've turned off GEO image display - just colored boxes for now * I can change this fairly easily later, but being that >20 images caused it to crash, I gave up on it.
      • I added an extensive readme file to the project which explains the .txt file, configurations, etc.
      • As long as the above changes work when we test, I'm fine locking this where it is

  • ScottLawrence - 2006 Dec 12
    • Leg color toggle button is working now. I tried it out at the test today, but I messed up something then.
    • I think I'm going to figure out some way to coordinate the leg color with the ENVI box drawing, so that it is easy to tell which images belong to which leg.

  • ScottLawrence - 2006 Dec 11
    • I just realized that I haven't updated in here for a little while, so I'd get caught up.
    • LIAS web and mailing lists
      • With help (and thanks) from Jim Bodie, we have a "lias" account on the CIS systems now.
      • This is where our mail lists will be managed from, as well as possibly the LIAS web pages.
        • email lists are set up now.
      • The Wiki is more of an internal thing
      • I'm not sure how this and lias.cis.rit.edu will be balanced out at this point
      • I'm also ready to go with updating the CIS CMS webpages wrt Lias, once Don gives me the content
    • Wasp Tracker
      • It now tiles ENVI files as they are received, but this is VERY memory intensive (ENVI Image data)
      • It also now can display just a rectangle (tinted or clear) to show basic coverage (ENVI Header data)
      • box outline thickness scales with zoom level
      • In progress: button to toggle leg and ENVI box color.

  • ScottLawrence - 2006 Nov 30
    • Wasp Tracker now does something really quite nifty. I'll demo it for people tomorrow. big grin

  • ScottLawrence - 2006 Nov 28
    • The Zoom and all of that is working perfectly now in WaspTracker?. I even added a little triangle to show the current heading on the aircraft position. It seems to work really well.
    • There's one minor bug, which might be due to my testing environment. I'll figure that out within the first minute or so of the flight tomorrow.

  • BobKrzaczek - 2006 Nov 27
    • JasonFaulring and I got the RIPCom straightened out, we think, and had a successful test flight.
    • We're ready for a Wednesday run-through!
    • Notes on the day's tests and whatnot at WaspNotes2006Nov27.

  • ScottLawrence - 2006 Nov 25
    • Or perhaps I'll put in a real formula for determining distance to the aircraft.

  • ScottLawrence - 2006 Nov 24
    • On Tuesday, I added a bunch of useful stuff to WaspTracker?: Image zoom, manual scrolling, position and history memory... so you only need to get a GPS lock once for the ground station, and restarts don't wipe the position history. I also added a widget to show where the plane is with relation to the ground station (direction).
    • In the shower today, I figured out a good, simple, easy way to determine how far away the aircraft is for the WaspTracker?. I'll try it out with the next flight.

  • ScottLawrence - 2006 Nov 18
    • I just implemented a simple history into WaspTracker?. It currently is set to save 1000 entries, but that is changeable. Here's a snapshot of it:

  • ScottLawrence - 2006 Nov 17 (part 2)
    • WaspTracker? now works great! I have the ground image that Don gave me for Utica, which works well for Bob's simulator data, and I also created one for the Rochester area that goes from roughly Parma to Canandaigua.
    • This movie shows a brief segment from my drive home on Nov 6. It shows the mechanism working. The backing imagery was swpied from google maps, by copy-and-pasting screenshots and hand-mosaicing them all in Photoshop. It took about an hour, but it looks pretty good. This movie has that imagery at 50%, since I was having a memory issue with the image sizes, which i've since fixed.
    • and here's another, longer clip, from Jefferson up to my house. This is with the proper imagery size:
    • If we drive out on Monday, I'd like to hook up the GPS to my laptop and watch the position update live... Although I don't have it with me, so i'd need for someone to snag it from my office. (Gps, black serial cable, USB serial device, USB cable)

  • ScottLawrence - 2006 Nov 17
    • The NMEA library I created, in combination with Processing means that it is super easy to make a simple applet/webpage that displays the current aircraft position via recentering a view with google maps. I should have this by the end of the day, easily. big grin (Note: this will require any machine capable of running Processing (just about any machine running OS X 10.2+ or Windows 98+) with a network connection.
    • I just made it much easier to configure Map type and zoom level for generating the google maps urls

  • ScottLawrence - 2006 Nov 16
    • Just putting the finishing touches on the NMEA java library. You feed it in a NMEA packet, and it maintains all of the values internally. You can get the Lat and Lon, for example out in the raw NMEA format, reformatted to be human readable, or in a Google Maps link URL.
    • It's got some other things. It's basically the code I wrote for a few recent Processing sketches, arranged nicely, and with a few added useful bits. I need to integrate this back into both the WaspFinder? as well as the AdpGui?, which I'll do tomorrow.

  • BobKrzaczek - 2006 Nov 15
    • Applanix -> GPS conversion is done and lightly tested. So far, it all looks good.
    • I've left a simulation running: about once a second, chapman broadcasts a GGA packet onto the LIAS network.
      • I'm using $PW as the station ID, per Scott's recommendation. This will make it easy to distinguish between GPS info from the aircraft and GPS info on the ground.
      • If you've got a UDP client or a packet sniffer handy, look for broadcast datagrams on port 5610 on our local network.
    • This, combined with the TCP repeater that's already in place, pretty much finishes the ADP-side of things for our GPS support. We can support antenna pointers, map trails ("Where's WASP?"), or whatever other GPS apps you want.
    • Example record being sent, that more-or-less means: At 12:49:49 EST, WASP was over 43.1311N 75.2495W at 2403 feet.
      $PWGGA,174949.78,4307.8688,N,07514.9677,W,1,,,732.356525,M,,,,,*04
    • Potential lurking errors:
      • wrong timezone (I only think I've got the conversion into UTC correct, I haven't proved it)
      • wrong altitude (I don't know if the Applanix is returning meters above sealevel, meters above the geoid, or meters above what, precisely)
      • wrong checksum (Scott reports a glitch when the checksums should be 0x0F that I need to check into)

  • BobKrzaczek - 2006 Nov 14
    • Big thanks to ScottLawrence for helping me track down the bug that was causing geo processing not to match up during a live flight, but to work successfully when simulated from the ground.
    • I can fix the bug, I'll need about 15 or so minutes on vulcan when we're next at the airport. I can upload the changes from my laptop and recompile on the plane.
    • You should see the mosaic scaling tweak to the GUI. It's very cool.

  • ScottLawrence - 2006 Nov 14
    • The GUI now has an easy way to increment directory names for saving image data ./Leg_# becomes ./Leg_002 and so on. It also creates the directories as needed, and warns when appropriate.
    • Also helped BobKrzaczek track down the weird geo'd VNIRS data stretch issue.
    • Finally added the ability to scale the mosaic during runtime as well (v2.70)
    • We should be good to go for a test tomorrow, assuming it happens.

  • BobKrzaczek - 2006 Nov 08
    • Thanks, ScottLawrence, for the text below. I'll get you the results of our tests tomorrow.
    • I'm done with the Applanix -> GPS converter and broadcaster. However, it's not tested yet; I don't have any Applanix group 1 data handy to test it out with. I'm sure it's not bugfree, but I'm optimistic that it won't take much to clean up any problems. I'll test it tomorrow in Utica, as well.
    • If you want to see the documentation system I've been loving lately, check out http://lias.cis.rit.edu/krz/aplxgps/main_8cc.html or just http://lias.cis.rit.edu/krz/aplxgps/ These docs are generated directly from the source code with no extra commands or special processing or anything. My Doxyfile is in CVS right now if you want to play on your own.

  • ScottLawrence - 2006-Nov-08
    • I checked in the gui version 2.67, to prepare for testing tomorrow. (It adds remote file saving, and other minor fixes)
      • the "Set" tab has the text box for setting file save paths, you probably want something like /Volumes/hercules/legs/01 (after you make a new "legs" directory, and toss a bunch of subdirs in there.)
    • you should be okay with typing make groundtest, and have it "just work"(tm)
    • oh, and yeah, you'll want to mount up smb://hercules and select the "adp" share.
    • In order to run the processing stuff: (for testing broadcast reception - if you want to use this, you can use netcat instead, obviously)
      • Install the latest version of Processing http://processing.org/download/ (0121)
      • Run it, so that it creates the Processing directory in your Documents directory
      • copy the Datagram_test folder from lias/rit/wasp/adp/gui/gpsim into that Processing directory
      • load it from the Processing gui (one of the buttons is "load", it has a dropdown that should list it.
      • click on run (the "play" icon). This should open a window for ya.
    • When a leg has completed, from orthos, run the gdal_process.pl script (i think that's what it's called) like so: ../gdal_process.pl . The only command line parameter is the directory to start in. This will create the Imagine subdirectory with those files in it for Shari to process (if you're going to do that too)

  • ScottLawrence - 2006-Nov-07
    • I created a simple shell script (thanks to BobKrzaczek for the starting point) which is in CVS in lias/rit/wasp/adp/gui/gpsim . The shell script "gpsim.sh" will read in the file "gpsdata.txt" one line at a time, and send it out like the broadcaster will do in the near future.
    • The gpsdata (unfiltered in "gpslogger/gpsdata.txt") is about 20 minutes of recording from RIT to my driveway at home. I figured this would be the best way to get some real data for testing use.
    • This should be enough for me to dive forward with the AntennaRotator?. I can produce data for the rotator software as if it was receiving it from the plane. Yay!

  • ScottLawrence - 2006-Nov-03
    • Gui not filing received data bug found and fixed - it now saves everything out great
    • Gui now has a --filer option to set where the images get saved to
    • Gui also has a while-running way to change where the data is saved, on the "Set" tab
    • Gui version number bumped to 2.65
    • running gdal_translate -of HFA envi_file.img imagine_file.img on 144 VNIRS (downsampled, georectified 512x512 terrapix) images on my laptop takes about 35 seconds (7.156u 6.549s 0:35.13 38.9%) according to 'time' This collection of data is larger than a usual flight leg. My laptop was not actively running anything else during thie time. It really is quite quick. I think we can be okay running this step from my laptop, or another laptop -- there's no reason to run this step specifically on orthos.

  • ScottLawrence - 2006-Nov-01
    • For what it's worth, gdal is available via darwinports on OS X. I can't test it out for speed since no data is online for me to test against right now. Once i can connact to 'orthos', and i've fixed the gui-save bug, I'll explore this.

  • ScottLawrence - 2006-Oct-20
    • Here's a video of the current ADP Gui running geoed Terrapix (VNIR), and basic operations too:
    • and here's the tail end of that, but with registered IR data instead:
    • I also made tweaks such that the three registered IR bands display with proper brightness in the gui. I added code so that three histograms are generated for each frame, then those are applied independantly to the three bands of data. I toned down the green to 2/3 brightness since the MWIR camera has the least dynamic range, and it washed out the images.
    • for the color IR above, I made the necessary tweaks to eliminate the "red banding" caused by the older alpha-channel formula.

  • BobKrzaczek - 06-07 Oct 2006
    • Trying to remember everything that's changed in the ADP in the last two weeks, and reapplying it to the source tree on vulcan that we lost when the drives were swapped.
    • Changed 0, 0, CenterX, CenterY to 1.0, 1.0, UpperLeftX, UpperLeftY in the ortho data portion of the ENVI writer.
    • Reformatted the way headers are printed to better match the ENVI format (after some random googling) in the ENVI writer.
    • Changed units from meters to Meters in the ENVI writer; believe it or not, it looks like that might matter to Imagine. Found this, along with header reformatting, after some creative Googling...
    • Removed the rotation=90.0 parameter of the ENVI writer.
    • Renamed the gimbal driver options in the ADP.
      • Nadir flights, supported by the "Move Once" driver: n
      • Manually gimballed flights, supported by the "Manual" driver: m
      • These options were backwards in the past, though probably no one dealt with it since it was covered by the adp startup script.
    • Updated SWIR and VNIR boresight data.
    • Worked around truncated text in controller's usage message to operator.
    • Rewrote the initialization of the triangle registration process. It now handles "no rotation", "image rotation", "table rotation", and "pre-rotated tables" at runtime. Also defaults now to 512 instead of 510 rows. Interestingly, somewhere along the way the bug with table rotation was fixed, so that works for us now.
      • ALERT! The new "pre-rotated" tables aren't valid. For example, if you examine the MWIR->SWIR mapping (TMSR.dat), the file starts at a lower offset than I'd expect, and ends with negative index offsets. "Whoops!" smile
      • Not a problem, though; we can run in flight with the "rotate the tables themselves" hack and it's all good for the meantime.
    • Updated adp startup script to understand new option structure to reg
    • Rebuilt entire ADP from scratch to propagate all changes through the controller, filer, and so on.
    • Tested registration and geo processing just to be sure.
      • Orthos of a single channel works. This proves that there's no further file corruption that might have crept into the leica s/w tree.
      • Ortho of registered IR imagery works. Combined with Scott's changes to the GUI, this should be that we can view a mosaic of IR imagery instead of VNIR live, while the plane is flying around.
      • Ortho of downsampled Terrapix imagery works.
    • ALERT! Timestamps are utterly screwed up. Looks like we've got GPS weeks instead of UTC again. Doesn't hold up the flight; we can fix it later.

  • ScottLawrence - 05 Oct 2006
    • JasonFaulring gave me an account on 'orthos' yesterday, and I tried out the gdal stuff, which seems to work really quickly. I started to write a perl script that finds the files and converts them in a directory heirarchy. here's an example command line for it: gdal_translate -of HFA envi_file.img imagine_file.img
    • BobKrzaczek got the simulator running with the register module, and I updated the ADPGui's "Image Data Manager" to deal with 3-band, 2bpp images, and display them as false color. Due to the design I have right now, the normalization can either base its ranges on all three frames, or just one. I set it to just one (SWIR) so that the color range and contrast look good. Right now, the current mapping is:
      • red -> SWIR
      • green -> MWIR
      • blue -> LWIR

  • ScottLawrence - 29 Sep 2006
    • I've re-re-fixed the geo data. it can now be found on the servers in /lias/wasp2/wasp_collects2/08-09-06/2006-09-08_WASP_NOROTATE_GEO.zip

  • ScottLawrence - 20 Sep 2006
    • I just wrote and ran a perl script /cis/staff/sdlpci/procgeo.pl on the last flight's GUI-saved data, and it can be found in /cis/staff/sdlpci/2006-09-08_WASP_FIXED.zip
      • can someone test that out for me? Don says it looks great!
    • I'll process through the acquired data tomorrow or later this evening.
      • Where is it? Only Registered IR data was saved, not Geo

  • BobKrzaczek - 14 Sep 2006
    • BuildingAceTaoOnOSX
    • We will have badger, the WASP gateway, ready for the flight today.
    • We will have an updated ADP, fixing the problems in the geo output files from the ADP described at yesterday's meeting.
    • We won't have aplxrep, the Applanix repeater, ready for the flight today.
      • Among other issues is dealing with its non-conforming broadcast packets in a way that doesn't break the ADP (which is a client of the Applanix just like this repeater needs to be).
      • Ugly Plan B hack: run the repeater as a client of the Applanix, then subscribe the ADP to that instead of the Applanix itself.
      • As designed, or plan B, will be ready for the next flight.

  • BobKrzaczek - 13 Sep 2006
    • DonMcKeown and I worked out why the SWIR GEO files produced by the ADP aren't tiling properly under Imagine or ENVI. We also found some other discrepancies between ENVI files written by the ADP, and those written by Imagine.
      • Image dimensions don't match. For example, the same SWIR image is marked as 813x814 pixels by the ADP, but Imagine generates it as 996x906.
        • Not a problem: Looking elsewhere, the ADP is marking the pixel scales as 1.228489m in the X direction, and 1.118940m in the Y direction. The Imagine data has 1.0m and 1.0m in the corresponding values. Guess what 813 times 1.228489 works out to be. It looks like Imagine is resampling the image to establish square pixels; the ADP doesn't. Both Imagine and ENVI seem to do "the right thing" with our non-square pixels, so this isn't a problem.
      • The ADP claims the upper left corner of each image starts at 0,0 on a local (relative) canvas. This needs to be 1,1 to match Imagine.
      • The ADP locates the X,Y of each ortho image by its center point (as generated by Imagine's geo-rect). ENVI files need the upper left corner, not the center point.

  • BobKrzaczek - 08 Sep 2006
    • As you've all heard and seen by now, WASP flew, did live geo-rectification of SWIR data, and sent this data to the ground, where it was mosaic'd in real time. Nice job, everyone!
    • The bug I was chasing down in geo rectification turned out to be a mere typo. Not Leica's problem at all this time; mea culpa.
      • Now I understand what happened and why.
      • I made a typo back in January 2006 during some geo-rect development. It went unnoticed.
      • I made the exact same typo (!) about two months ago during my updates to the system startup tool.
      • A test I wrote to catch exactly this kind of problem passed, when it should have failed, because of the original typo from January. Unbelievable.
    • Next three things on the list for WASP (in no particular order):
      1. Relay Applanix data between the two networks in the air.
      2. Create a simgeo utility for streaming captured data (like SWIRGEO or REGGEO) so ScottLawrence can add registered data support to the GUI and improve the canvas/mosaic code.
      3. Tinker with badger a little bit more, to make the RIPCom bridge for our ground based network.

  • ScottLawrence - 08 Sep 2006
    • WASP flew, but the 900mhz link was down, so I was unable to give WaspFinder2000 a test. We did get lots of geo'ed data downlinked though, it all came together nicely.
    • I just got the commanding done in the Antenna Commander. You can click an azimuth and elevation in the screen graphics, and it will rotate to the correct position. Whoo!
    • I am going to clean up the code a bit, and drop it back into WaspFinder2000, which needs a bit restructuring.
    • I added a third serial source for the WaspFinder2000 - ground station GPS position. I need to have another way to enter the data though -type it in or somesuch, since we might want to run without that gps. (perhaps plug in the gps, let it get the position, then plug in the aircraft positioner or something. This might become moot if we get aircraft positioning information via the RipCom? instead of the serial link.)

  • ScottLawrence - 07 Sep 2006
    • After trying to get things working in the WaspFinder2000 app, I decided to pull out the antenna code into a second app (antcmdr) to test it, and get it figured out. I should be shoving that code back into the main program tomorrow.

  • BobKrzaczek - 07 Sep 2006
    • Took way too long to get the registered imagery from the August 9 flight straightened out.
    • Banged out a quick conversion utility to convert BSQ data into BIP format for the registered images; also does an optional 14->8 bit conversion along the way. Pretty specific to WASP RegIR data at this point.
    • Wrote a tool that helps in processing multiple flight legs in a single session. Lets us quickly build up a single RegIR directory that is the sum of all previous flight legs, for example.
    • Results, details, and so on, available at WaspResults20060809.

  • ScottLawrence - 06 Sep 2006 (Wed) - WaspFinder2000
    • JasonFaulring received the antenna rotator, wired it all up, and debugged the unit, and got it working
    • I figured out how to talk to it from WaspFinder2000, and now have it updating the gauges based on the current position. It polls the rotator about once a second to ask the azimuth and elevation, which is displayed in the UI.
    • I need to think a bit about how to track the plane - the rotator is slower than I had anticipated, and I'm not sure how to get it to react when the plane is near overhead, when it'll be moving quickly with relation to the ground. Perhaps I need to lead the plane or somesuch... hrm.
    • I need a second USB-Serial adapter if we're going to have them both running.
      • Keyspan makes a 4 port serial adaptor for USB; it might save you from needing a USB hub and multiple adaptors and whatnot. It's marked as model USA-49W. -- BobKrzaczek

  • ScottLawrence - 01 Sep 2006 (Fri) - WaspFinder2000
    • Finished up the re-working of the code. The positional stuff is all heirarchical classes. I might need to revisit that; there's something about the current layout that doesn't sit right in my gut.
    • JasonFaulring gave me a pointer to NY State GIS data, and I've been working a bit on getting my head around it.
    • I added the abstraction for the AntennaRotator?, but we don't have any hardware yet.
    • Question: Am I giving too much detail in here?

  • ScottLawrence - 31 Aug 2006 (Thu) - WaspFinder2000
    • I just got Processing ported to SunOS?. It was remarkably simple, once I had the right pieces in place.
    • I reviewed the documents about the antenna rotator that JasonFaulring is picking up. I asked him to pick up another KeySpan? USB-RS232 adapter, so that I'll be able to control that from my laptop. If we need to control anything else serial (ground station GPS?) we'll need to get a third USB-RS232 adapter, as well as a USB hub. heh.
    • Re-working a lot of the code as well, and getting used to Processing 0115... They changed a lot of stuff a little while ago, and I'm getting used to the changes.

  • ScottLawrence - 30 Aug 2006 (Wed) - WaspFinder2000
    • Just signed up for a google maps API account thingy, and was reading up on it. I think that might be a good, quick way to get positional visualization on the ground. The GIS library I found was not as useful as I originally thought, but I would like to still pursue that. Any work I do on that will be directly filtered back to the ADPGui as well, so big bonus there.
    • I added a thing so that when you hit 'g' on the keyboard, it opens the plane's current position, centered in google maps. Nothing fancy, but good enough for now.

  • ScottLawrence - 29 Aug 2006 (Tue) - WaspFinder2000
    • I just finished creating and entering all of the vertices for the simple Aztec model. 85 vertices! whoo!
    • I also hooked it into the WaspFinder2000, so it displays the plane with proper orientation as per the data.
    • I'll be hopping over to some more serious stuff now.
    • Once Jason gets in the motor drive unit for the antenna, I'll figure out how to control it from WaspFinder2000

  • ScottLawrence - 28 Aug 2006 - WaspFinder2000
    • After showing the current software to JasonFaulring, I added GPRMC support. I'm not quite sure how I overlooked that before. Just prior to that, I added a simple Lat/Long to heading converter thanks to PatrickStein. So, now we have support for both, although I doubt there will ever be a need to do so.
    • I'm putting in a simple model for an Aztec and a 172 next.

  • ScottLawrence - 18 Aug 2006 - WaspFinder2000
    • I now have a running Processing sketch (program) that displays heading, and lat/long which it gets from a GPS hooked up via a serial port. I'll be trying it out with real data, rather than "demo mode" data tonight as I drive home. In its current state, it is usable for informational purposes, but I'd like to add a lot more to it, which I'll be doing when I get back from my vacation next week.
    • It currently supports the following GPS NMEA packets: (tested)
      • GPGGA, GPGLL, GPHDT
    • And the following Applanix NMEA packets: (untested)
      • INGGA, INHDT, INVTG, PASHR, PRDID, INZDA

  • ScottLawrence - 16 Aug 2006
    • JasonFaulring gave me the interface specifications for the Applanix serial output, as well as the GPS NEMA spec. I think the first version will be GPS based, so I can test it in my car.

  • ScottLawrence - 16 Aug 2006 - WaspFinder2000
    • Soon, the WaspFinder2000 will start. It seems like it'll be a very simple thing to make. Later today, I'll meet with Don to find out what he expects from it, and I'll be getting the communications protocol from Jason later today as well.
    • My first impression: Since the communications will be over the 900mhz line, it might as well be a standalone java app - no CORBA or anything, just a serial port.
    • One step further, I might as well just write it completely in Processing, and bling the hell out of it. Make it nice to look at, and at the same time, very informative. (Perhaps recycle lots of Processing code I already have and show a 3d model of the aircraft in its current orientation. Should be quick and fun to write.

  • BobKrzaczek - 9 Aug 2006
    • Okay, Microsoft Certified Professional Magazine Online leads me to think
      • either route add 129.21.0.0 mask 255.255.0.0 192.168.0.1 -p
      • or netsh routing ip add persistentroute 129.21.0.0 255.255.0.0 "Local Area Connection 2" nhop=192.168.0.1
      • But in both cases the route will need to be deleted when we come back to the lab.

  • ScottLawrence - 8 Aug 2006 (Tue)
    • ENVI file saving should be complete now, as per the logic/structure that BobKrzaczek has sent to me. I doublechecked the code multiple times... I think it's good.
    • Just to be sure, I md5sum'ed the IR data file saved by these routines, and compared it with the original files, and they were the same. (I was nigh positive they were, but I just wanted to be sure.)
    • The data sources I had were all LSR non-geo'ed, so I couldn't check Utm or Geo .hdr output, however, I'm confident they should be good.
    • Tomorrow after the demo, I'll be poking at Java-GeoTIFF with a pointy stick.

  • BobKrzaczek - 8 Aug 2006
    • I think these commands will be equivalent for our purposes
      • unix route add -net 129.21. 192.168.0.1 255.255.0.0
      • winxp netsh interface ip add route 129.21.0.0/16 interface="Local Area Connection 2" publish=yes
      • or maybe netsh interface ip add route 129.21.0.0/16 interface="Local Area Connection 2" nexthop=192.168.0.1 publish=yes
    • badger has been configured with the new network params from JimBodie; it's ready for the roof of Bldg 76 right now.
    • Per my request, JimBodie has also changed the samba share from pegasus so that we can access it from both wired and wireless networks anywhere on campus. This will support our use of it from the Romper Room. He'll switch it back next week.

  • ScottLawrence - 8 Aug 2006 (Tue) Java ADP
    • I'm going to be finishing up the ENVI support for file saving
    • I need to look into geotiff-jai and xtiff-jai for Java-GeoTIFF file saving support.
    • Hrm... I might want to pull the image saving out of the Data Catcher, and into its own class... i need to think about the best way to do that...

  • ScottLawrence - 7 Aug 2006 (Mon) ADP Gui / Java ADP
    • At BobKrzaczek's request, the ADP Gui now has basic filer-like functionality (via the Data Catcher object)
    • It saves out ENVI data files (.img and .hdr files)
      • The .img files are the same as what was sent (VNIR is processed/smallened before sending, so I wasn't able to test it). I have not loaded these into ENVI to confirm their accuracy, but I'm pretty confident they will be just fine.
      • The .hdr files are a different story. They currently accurately portray the sent data, although not the GEO data for the images yet. (Also, for some reason, the FPA temperatures sent with the image are zeroes, rather than the expected values. I need to look into this more deeply.
    • I should be able to crank the GEO data into the .hdr files tomorrow no problem.

  • BobKrzaczek - 3 Aug 2006
    • Following our successful deployment on the aircraft last week, which I should have logged here about... wink
    • Built up a little network bridge out of my old Sun LX and some spare network cards. It's a system we can place on the roof of building 76 with the Ripcom, and connect to the plane remotely from the "Romper Room" in 17.
    • Tested it. Works well. Will run without a console, so you can just "plug it in".
    • Bandwidth is limited by one of the interfaces being only 10bT. I was able to reliably push 700 kB/sec through it, though. This is almost twice our original assumption of 384 kB/sec, though, so it'll be fine. We won't be able to connect to everything like we did at Landcare last week, but that wasn't in the script anyway. wink
    • Need to figure out how to manually add a route to Windows XP. I'm sure "netsh" will do the trick, as soon as I learn enough to translate route add -net 192.168. badger76.cis.rit.edu 255.255.0.0 into Windows-ese.
    • Eventually, we'll want a newer system (my LX is over ten years old); but in the meantime, this works fine.

  • BobKrzaczek - 24 Jul 2006
    • Successful test, including use of Ripcom, on live hardware with a laptop running as a ground station. Looking good!
    • Near complete rewrite of system startup; reflects a lot of what we learned. New startup is more intelligent about saving flights to disk, about starting up registration and rectification processes.
    • Need to incorporate a new network setting. Windows gets confused easily, doesn't route between interfaces well.
      • Main ethernet: 10.0.0.4, connects to Applanix.
      • Secondary ethernet: 192.168.0.4, connects to Ripcom.
    • Take advantage of extra day (Tue) to test better startup and network hacks before WASP heads into the plane on Wed.

  • ScottLawrence - 20 Jul 2006 (Thu) ADP Gui
    • implemented NAMESERVICE environment variable support (ie: context:chewie or netmap:hansolo )
    • if there's no FOOTLOCKER specified, it assumes "." (handy for testing)
    • Application now exits on close button
    • added a little more output to the log, so that you know what variables are set, which netmap file is loaded, etc.
    • added a few targets to the makefile to make things a little easier (air/ground/etc...)

  • ScottLawrence - 19 Jul 2006 (Wed) ADP Gui
    • After a test this morning with Geo-LWIR, I have fixed the mosaicer to use properly normalized imagery
    • created the 2005/2004 archives of this project diary to make this page smaller. smile
    • proper command line arguments are done
    • netmap is loaded from the FOOTLOCKER path, rather than cwd
    • fixed a minor issue where the GUI log wasn't updating to stdout anymore. oops.
    • Tweaks to do next:
      • figure out the manual normalization slider strangeness
      • QUESTION? Where should my log files go? Is there an equivalent for FOOTLOCKER for acquired data?
        • Not specifically, no. I'd suggest c:/adp/products for this, but there's no hard and fast rule for log files, yet. -- BobKrzaczek

  • ScottLawrence - 18 Jul 2006 (Tues) ADP Gui
    • The reconnect issue has been worked out, and seems to work well.
    • there's a new command line option for the GUI: --GroundStation It sets things up to not connect on start up, so you can connect individual items at will. It also disables active updates and autoreconnects.
    • --help, -h, -? also exist to display usage (which is currently incorrect while i rework the argv parser)

  • ScottLawrence - 17 Jul 2006 (Mon) ADP Gui
    • I'm flooding. sorry.
    • Initial testing in the lab went really well. It brought a few minor issues to my attention which I'll be fixing tonight and tomorrow.

  • ScottLawrence - 17 Jul 2006 (Mon) ADP Gui
    • netmap connections work! I had to mangle the strings from something like iiop://vulcan.cis.rit.edu:31313/swir into something like corbaloc::vulcan.cis.rit.edu:31313/swir but it works!
    • normalization is mostly working; manual adjustment doesn't seem to be behaving properly, but otherwise it's good
    • I changed the EO Params display to only update when you press the huge "update" button on that tab, to reduce update bandwidth. I might do more of this, just to try to keep the app from bogging down.
    • active/passive updates work great. I thought i was having an issue where the ControllerProxy? was forcing an update when 'active' was set to false, but it turns out that it was autoreconnecting. heh. oops.
    • i changed the net/startup tab to allow you to autoreconnect to the controller or image sources independantly...
    • that panel also allows you to selectively use the netmap and/or nameservice (or neither)
    • To Do Next:
      • adjust the command line parser to work the same way as BobKrzaczek designed the other apps
      • connection/reconnection tests (Simulate bank turnes, etc)
      • fix manual normalization sliders

  • ScottLawrence - 16 Jul 2006 (Sun) ADP Gui
    • I'm working from home, so I can't test these things, but they build and should work.
    • just finished:
      • Normalization hooks in the gui
      • Active/passive GUI updates (Controller was the only one that was being polled for status)
      • Name Service / Netmap toggle from gui filters down to the CorbaHelper? class
      • band selection on an item that doesn't exist doesn't throw exeptions - draws filler image
    • Tomorrow morning I get to test this all, including the netmap stuff. Whee!

  • ScottLawrence - 16 Jul 2006 (Sun) ADP Gui
    • I finally checked in the code I've been working on for the past few weeks. I wanted it to be at least close to the functionality as the old code before I checked it in.
    • The entire app has pretty much been redesigned, recoded, and refactored. I don't think there's a single source module that hasn't been touched. (Minus the ImageDataManager?, which I'm using for now, so that I can hit the Monday deadline... It doesn't compromise the new design, and it "just works".)
    • There are a few aspects to the new version of the app that need tweaking before tomorrow:
      • Normalization hooks need to go back into the GUI
      • image source reconnections need testing
      • active/passive gui updates need to be implemented (a simple if block here and there)
      • band selection on an item that doesn't exist shouldn't throw exceptions, and should clear the image view
      • Name Service / Netmap toggle needs to be passed down to the CorbaHelper? class
    • these should be massaged into place tonight, ready for testing tomorrow.
    • eventually, I'd like to drop in the new mosaicer, as well as the new Image manipulation class, but I'll stick with the ImageDataManager? for the immediate future, since it works good enough.

  • BobKrzaczek - 15 Jul 2006 (Sat)
    • Took the breather in WASP activities to get some SOFIA work done.
    • Made, documented, tested, and installed all necessary changes to support design reference missions in the Sofia Proposal Tool. Look forward to feedback on this in the coming week.
    • Nice to be out from under that rock.

  • BobKrzaczek - 14 Jul 2006 (Fri)
    • New TaoHelp networking is present in the system controller, mon, filer, and even the old fire detection prototype we have lying around.
    • Tested it with the Ripcom (thanks, Jason): simulated flight on vulcan, data monitor on chapman simulating the ground station. We can turn the Ripcom on and off at will, everything seems to work.
    • Monday: integration with Scott's GUI, testing the system as a whole, finding out why the geo rectification process doesn't build any more.

  • ScottLawrence - 13 Jul 2006 (Thurs) ADP Gui
    • pulled out a bunch of hacked code leftover from when the net tab was its own window
    • cleaned up the GUI layout a bit via tweaking tab names and such
    • Currently, I'm migrating all of the new network code from the test app into the Gui application:
      • Done:
        • Convert the Controller Proxy over to use the CorbaHelper? class
      • Next:
        • Convert the Gui engine to refer to the new Proxy for vitality updates, connect/reconnect, etc
        • Switch the image catchers to use the new DataCatcher?
        • Switch the image processors to be the new ImageMogrify?, rather than ImageDataManager?
        • Update the Gui engine to refer to these new items as above

  • BobKrzaczek - 11 Jul 2006 (Tue)
    • Ported filer and triangle registration over to new TaoHelp. The geo rect process won't build currently, something's hosed with the build environment, we're missing Lggm header files. Where did they go?
    • Added CORBA policy manager (polman) support to TaoHelp.
    • Added a default polmanhook function that sets an ORB-wide timeout value. Time is set on the command line, or the TIMEOUT environment variable, or a default of ten seconds.
    • Limited testing at present... but it works!
      • system controller running on chapman, running part of the DHS demo
      • monitor also running on chapman, connected to SWIR camera
      • pause the monitor process (simulate a network outage)
      • controller correctly notices the timeout, but doesn't drop the subscription
      • controller doesn't block, continues to service simulated cameras
      • release the monitor process (simulate the network's back)
      • system picks up where it was, no data lost, no data backlogged, no reconnects, monitor just starts receiving new SWIR frame
    • Next tests:
      • Run monitor on vulcan, plug and unplug the network cable.
      • May need to capture other exceptions "down low" in the system; CORBA::TIMEOUT is expected, but CORBA::TRANSIENT might appear as well, now that I think about it.
      • Figure out why I can't build geo process?!?!

  • ScottLawrence - 11 Jul 2006 (Tuesday)
    • Would have had this finished last night, but the building decided to catch on fire, it seems.
    • Just finished up the first cut of the new network interface design (it's in my test svn repository... once it's stable and working, i'll move it into CVS.)
    • It now will look at (in order) Netmap, IOR reference, Name service, when it has a string identifier to connect to an object. When one fails, it falls back on the other ones.
    • I'm in the process of restructuring everything else around this now... the current AdpGUI2?.java file will soon be much smaller, much more of a skeleton, and less of a skeleton with a bunch of muscles and cartilage messily stuck to it. I'm really happy with the new design. Gone are the arrays of various objects (net objects, strings, ints, etc) and now we have a real object-oriented design. go figure.
    • Oh, and I also tweaked the command line options to conform to what BobKrzaczek has mentioned in WaspCommandLineChanges
    • Note for future-self: switch this all over to use the latest version Java

  • BobKrzaczek - 11 Jul 2006 (Tuesday)
    • Moved the controller over to the new TaoHelp. Working with it and mon revealed a few more bugs. Now fixed.
    • We now have seamless support of both CORBA Name Service and our new static Netmap file. Objects are on the net. Binding, lookup, and unbinding all work correctly.
    • Let me repeat: Objects are on the net!
    • Two sets of tasks remain:
      • Move remaining objects (filer, reg, wasp-geo) over; this is pretty easy now, especially since we can just look at the before and after versions of mon to illustrate the changes. (One of the many benefits of source control!)
      • Add POA manager policies for network timeouts. Remove automatic delisting of remote objects on network failures, instead make them "keep trying" until some total number of errors. With aggressively small timeouts, this should be the final hack to support the RF network demo.

  • BobKrzaczek - 10 Jul 2006 (Monday)
    • Minor bugfixes to the TaoHelp changes. Reworked some usage messages, command line scanning, etc.
    • Ported the mon utility to the new TaoHelp. Works great. Support for CORBA Name Services vs. Netmap seems to be fine.
    • I'll move the system controller over to it next; that will let me test the netmap layer further, as well as give me the barebones of an end-to-end test system.
    • Then we move filer, reg, wasp-geo over to it as well, and we merge with ScottLawrence's changes to the Java side of the house.
    • Feeling muuuuuuuch better about things right now.

  • BobKrzaczek - 10 Jul 2006 4am (Monday)
    • Finished what turned out to be a rewrite of the TaoHelp library; the change from CORBA to Netmap resulted in a number of minor, but essential, changes to the TaoHelp class. As the changes kept piling up, we ran into distinct initialization differences between the CORBA name service and the Netmap name service. Eventually, I threw in the towel and rewrote it.
    • Hey, PatrickStein! I know you've referenced (if not outright used) TaoHelp in other software you've written. You might want to look over newtaohelp.pdf and see what I've changed. It's not quite 100% compatible anymore.
    • I'll move the controller and a few other objects over to it tomorrow morning, Monday. We should have the objects on the net that I promised last week ready, then.
    • This change is becoming a real hassle; CORBA does not want to play nice with intermittent networks. Hindsight is 20/20, of course, but if we knew the network would come and go, we might have sacrificed some of our network flexibility for a simpler middleware layer, back when the ADP was designed.
    • We'll be able to make the demo; but this isn't going to be a long term strategy for the system.
    • Going home, sleep now.

  • BobKrzaczek - 28 Jun 2006 (Wednesday)
    • WASP Networking
      • Got the Netmap and related classes into the aid library.
      • Next: move the mon and controller executables to netmap as a first integration; we'll worry about filer and wasp-geo and such later.
      • POA timeouts will be a piece of cake. Object reconnections, on the other hand, will be "interesting" in the more intricate things like the controller.
    • InfoSciTech SBIR
      • Had our tech telecon with them. They will be here on July 6 for a more detailed review. Can't wait.

  • ScottLawrence - 27 Jun 2006 (Tuesday)
    • Just finished the java version of the Netmap class. It should be fully functional, as per BobKrzaczek's design on WaspNetmapFile.
    • While working on this, I realized everything I've got to do to clean up the GUI networking code, and in the process, get reconnect working.

  • ScottLawrence - 26 Jun 2006 (Monday)
    • No current milestones have been reached, but I thought i'd throw this out here so that people know where I am and what I'm working on.
    • Currently on my plate:
      • Re-working the Mosaicer - this seems like it's a non-critical thing to work on, but it's helping me hammer out the updated network interface, which will allow automatic reconnects.
      • Writing the java equivalent of Bob's networking/name lookup stuff, and integrate this into the GUI code. I'll probably write up a simple version of this for subscribing, to be tested once the final versions of the server side are pushed into use. Then, time permitting, I'll add the server side of this to the java library. (This is not used currently, but can be for testing, etc.)
    • Up shortly:
      • look through the NUC selection code to see if the odd behavior Bob found was in the gui code, and fix it if it is.
    • Hrm, I really should take a step back and consider making a Java Aid Jar, putting all of the common stuff into it, rather than having some image stuff being in the GUI app, some in the CORBA Jar...

  • BobKrzaczek - 25 Jun 2006 (Sunday)
    • Tested WASP hardware with updated RTIE logic in the controller.
    • Couldn't open the roof door. My keys don't work there. sigh I won't even ask why the exterior door has a lock in the first place.
    • Not letting the lack of images hold us up, I proceeded with testing the FPA and NUC table, since that's the focus of the work anyway. But it sure would be nice to test with images...
    • Found and squashed a few small bugs in the changes to the code.
    • Verified that both the current NUC, and list of loaded NUCs, are being reported to and decoded by the ADP controller correctly.
    • Verified that new NUCs can be selected via the GUI, live, on a running system.
    • Verified that the new timing loops to the serial RTIE boxes run at a "good rate" without overrunning the system, or going so slow that FPA temps fall out of sync with collected imagery.
    • Reported minor display glitches to ScottLawrence but overall, the GUI was "ready to go" and implemented the interfaces properly. Kudos!
    • Bug The SWIR camera's FPA temperature is bogus. It should be around 200,000 mK. Instead, the temperature reported is roughly four million degrees, putting it right up there with shells of gas surrounding supernova remnants...
      • Kidding aside, the RTIE returns a four byte quantity of millikelvins. The most significant byte of this number, which should always be zero, is actually 0xff --- clearly, something's come unconnected or undone.

  • BobKrzaczek - 23 Jun 2006 (Friday)
    • Had a nice chat with Mike Richmond about LIAS possibly joining SNAP. Basically, RIT is considering becoming a member organization of SNAP. In order to be a member, RIT needs to name three individuals willing to join the SNAP collaborative. They have two, and are looking for a third. Stefi put our name out there. Quoting no one in particular smile it seems that there are things SNAP will need addressed (e.g., daily downlink from the spacecraft, management and reduction of data on the ground, onboard (limited) reduction of data) that no one is currently focused on. Further discussion will ensue when Don Figer approaches HarveyRhody (heads up!) later this week.

  • BobKrzaczek - 22 Jun 2006 (Thursday)
    • WASP Networking - Create a new syntax for the WaspNetmapFile (better comments and text formatting), implemented a parser for it, implemented the actual Netmap class, wrote test code to exercise valid syntax and errors, and tested the lookup(), iorbinding(), and endpoints() functions. It's all good.
    • Sat in on telecon regarding SBIR proposal for image conflation workflow thing. Would up having to spend a half hour next week talking to InfoSciTech engineers about their involvement in the project, if we accept it.
    • Got Oracle information collected and sent to Guy Johnson.

  • BobKrzaczek - 21 Jun 2006
    • WASP Networking - not as far as I wanted. I have a parser for the netmap file, but it still needs to be integrated into the ADP aid library. Also, Jason thinks he'll have a test environment for WASP (full hardware) to test RTIE changes in ADP tomorrow morning.
    • Proposal Work - Further "noodling" on ideas for an SBIR proposal we're considering.
    • SOFIA - Reviewed and edited Pat's writeup on their wiki on Archive Cache interactions. Worked out a plan of changes with Pat to the Archive Cache, End-Of-Flight processing, and Pipeline Pedigree processing to deal with all the recent issues.

  • ScottLawrence - 21 Jun 2006 - WASP ADP Mosaic issue
    • I tracked down the issue with the ADP Gui mosaicer cutting off the easternmost bits of the Pittsford-Henrietta flight. (DHS Demo) It turns out that if the georect module is run without the dem (roc-new-sub35.img), it doesn't clip. I asked Shari to help show me where the edges of that DEM lie, and it goes exactly where the clipping is occurring. I asked her to generate a new DEM for this dataset at her next convenience. There's no pressing need for it right now, it's not causing anything to crash, and we don't have any demos immediately.

(I added the horizontal rule to help emphasize when we had our last meeting... wednesdays.. if this is a bad idea, remove it. :] -- ScottLawrence )

  • ScottLawrence - 19 Jun 2006 - WASP ADP Gui
    • I'm experimenting with rendering the graphics for the GUI in a different way, so that we don't have to have such a high memory requirement for the GUI... also at the same time, this should lead to a solution that renders without the clipping issue we've been seeing in the GUI recently, and also, it should provide for a more flexible in-flight mosaic rendering.

  • BobKrzaczek - 19 Jun 2006 - WASP Networking
    • Started FixingObjectAddresses as a collection of notes on fixing CORBA objects to a specific network address.
    • Working! I've got the implementation of NameServer-less objects working in some test applications. I'll give it a spin on vulcan tomorrow, just to make sure there aren't some "Unix only" isms in the code, but I think we're set with it.
    • Next step will be to define the network map that goes in the electronic footlocker, write a parser for it, and start to move the ADP over to this new approach.

  • BobKrzaczek - 18 Jun 2006 - WASP Networking
    • Developing a "standalone" test implementation of our "fixed network" approach to object location on the net. PatrickStein suggested that we implement our changes in a test environment, before making sweeping changes to the WASP code.
    • Been trying just that on Friday and today (Sunday). It's not hard so much as it's just tedious. Every single niggling little detail has to be set just right, or it just doesn't cooperate. Estimated completion: Monday afternoon.

  • BobKrzaczek - 14 Jun 2006 - WASP OSU code
    • BillHoagland and I continue to review their source code.
    • Pushed the OSU AIMS software through Doxygen in an attempt to get a handle on this codeset. http://lias.cis.rit.edu/osu/aims/index.html No amount of callgraphs and source code analyzers, though, can beat solid documentation and commenting... oh well.
    • Still looking like we'll extract the algorithms and other content from their sources, but wrap them into our own class hierarchy and library scheme.

  • BobKrzaczek - 13 Jun 2006
    • WASP Networking and Ripcom
      • With suggestions from ScottLawrence, and a review with PatrickStein, I think we have the network topology settled on WASP between the air and the ground.
      • We're going to need to introduce some new object-finding and and object-binding code in the ADP aid library to support our new Name Server-less tree.
      • ScottLawrence and HarveyRhody both suggested that we cover this at the 11am seminar on 2006 Jun 21.
    • Technology Licensing, Commercialization, etc.
      • Had a very promising conversation with Varda Main about ways that we could market prototypes and proofs-of-concepts like WASP and WASP-LT. This might lead to a path to commercialization and licensing we hadn't considered. Will continue to pursue this.

  • BobKrzaczek - 13 Jun 2006 - WASP ADP
    • Added formal timestamps to the ImgWriter generation of ENVI header files. Some months ago, I updated all files in the wasp2 hierarchy to include timestamps based on the associated Applanix data. Now, this data will be automatically stored in all future images collected by WASP.
      • ALERT! Currently this is based on the operating system's time, and not the Applanix. The reason is that the Applanix time is in the GPS week, and I haven't written the conversion from GPS week to UTC time. As long as vulcan's time is kept up to date, this skew should only be on the order of seconds.
    • Added support for getting the active NUC table, getting the loaded NUC tables, and selecting a new NUC table.
    • ALERT! In order to support a richer NUC table interface, the RTIE looper had to be changed. This needs to be stress tested in the lab.
    • Other tests we need to run soon:
      1. Run the full system, on live hardware (no simulator) with orthorectification of registered IR data. Let's make sure we have enough idle cycles on the existing system to play with the extra RipCom functionality we're planning.
      2. Run the Leica mosaic module without using GeoTIFF as the output format. See if we can't whack that four hour compute time down to 45 minutes or so, as JasonFaulring reports for the interactive Imagine 9 time to mosaic fullsize Terrapix data at 0.5 GSD for the Henrietta Pittsford flight.

  • BobKrzaczek - 2006 Jun 09 - Cell
    • Summary of recent work:
      • Some weeks ago, we selected Bayer decoding as our "test algorithm".
        • Prove we understand CBEA.
        • Give us some exposure to the running system.
      • New host system is being ordered (is already ordered?) by JasonFaulring.
      • Have different approaches to implementing the algorithm on multiple SPU. We don't want to go with all 8 processors, since we want processing power left over for other apps.
      • Simple subsetting is too easy for this system, but it forced me to think about how to scatter/gather the data into Local Storage.
      • Want to see implement something that would be a payoff for Cell: something like AHD or VNG techniques, which take a minute or more realtime to run.

  • BobKrzaczek - 2006 Jun 09 - WASP
    • Got to get back in the habit of writing status entries, everyone.
    • Follow this example.

This is the old project diary, from when it was the WASP Project Diary. There is some LARCH status in here as well. From now on, be sure to sign your status with the project or task, as you see above...

  • ScottLawrence - 28 Feb 2006
    • Fast Normalization removed, since it wasn't working anyway
    • Image Data Manager class' generateProducts() function rewritten - better to maintain now
    • "Hardcoded Gain" adjuster added
    • default normalization is changed to something that looks better now.

  • ScottLawrence - 14 Feb 2006
    • Histogram Equalization implemented. It seems to hang the idm occasionally though. need to track that down.

  • ScottLawrence - 13 Feb 2006
    • Manual/Auto/Smart normalization complete. (I had to clean up some code in the IDM as well)
    • live visual feedback for normalization settings is implemented too.
    • imo, it looks really nice. I'm very happy with it.
    • HarveyRhody explained Histogram Equalization to me, so I think I'm gonna implement that next.

  • ScottLawrence - 10 Feb 2006
    • Manual Histogram generation complete
    • histogram only is generated on demand
    • as the sliders are moved, it updates the image/histogram
    • The min/max positions are also shown in the histogram
    • (Min is the lower of the two sliders, Max is the higher of the two)

  • ScottLawrence - 09 Feb 2006
    • Previously (between 06 and 09 Feb)
      • put in the new terrapix boresights
      • cleaned up vnir geo data to look much better in the mosaic
    • Just finished the Manual normalization.
      • the user can select manua/fast auto/slower auto
      • the sliders for min/max as well as the number displays for them update when 'automatic'
    • histogram generation still seems to bog the thing down. That's next to fix.

  • ScottLawrence - 06 Feb 2006
    • I must remember to update this more frequently
    • AdpGui?
      • updated to show north upwards for vnir mosaicing
      • transparency for vnir/color mosaicing cleaned up substantially
    • To Do next:
      • finish up the Manual Normalization
      • fix the histogram rendering
      • use proper terrapix boresights to clean up vnir mosaic

  • BobKrzaczek - 05 Jan 2006
    • Making a note here so I don't forget (this is as good a place as anywhere).
    • During Jeff Naus' visit this week, we've made a lot of changes so that geo-rectification temporarily runs under their debugger. This means, at present, that FOOTLOCKER and NameServiceIOR are being hardcoded in the geo rectification application. This will not work when WASP has to fly! But it got the debugging "off the ground" while we're working together on straightening out Leica's Lggm libraries.

Older items can be found in ProjectDiary2005

toggleopenShow attachmentstogglecloseHide attachments
Topic attachments
I Attachment Action Size Date Who Comment
pdfpdf newtaohelp.pdf manage 276.8 K 01 Feb 2007 - 11:57 BobKrzaczek rewritten TaoHelp? class
pdfpdf fireTest.pdf manage 33.0 K 01 Feb 2007 - 11:57 HarveyRhody Illustrations of fire detection results
This topic: LIAS > WebHome > ProjectDiary
History: r190 - 29 Mar 2007 - 07:15:00 - BobKrzaczek
 
This site is powered by the TWiki collaboration platformCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback