one failed but interesting approach to classifying images using hierarchical trees.

Multi image stacking that’s been driving me crazy these last weeks. Just completed the first mosaic, 2008, comprised of ~30 images, yes! Running a despeckling program in IDL, hope it looks good.

Here is an example of the core of the despeckler procedure:

I’ve been planning to optimize it to run without the use of for loops, however at the moment this is just not necessary given we only have 3 mosaics to run (but it would go from 30 minutes processing time to probably around 30 seconds!).

Sensors and climbing gear ready to load into the truck. Leaf dryer in LL corner.

Full climbing setup for LiCOR 6400 canopy work up to 100 ft. Sweet! The climbing gear for climber is a stored in a separate larger backpack.

02/01/11

System for canopy LiCOR measurements. Allows for ground based assistants to lift the LiCOR into the canopy, then fine adjustments via the climber.

 

System now including the extension for the cuvette head. This version will be cleaned prior to starting measurements on Wed or Thur. I’ll post a photo of the final cleaned setup before we head out.

01/27/11

Close up of fern fronds used for LiCOR 6400 analyses. These are scanned at 600 dpi and used for area calculation prior to drying and lab analysis.

01/26/11

Week of Jan 24th 2011

Rapid progress on understory LiCOR curve sets.

01/13/2011

Removed the sensors for calibration. Putting them up again early next week.

Angelica carrying a full geo-referencing kit.

LiCOR 6400 on the back, essential gear on the front. Ready to go!

Set up in the woods.

01/10-12/2011–Mon-Wed

LiCORing in the rain. Good progress on understory curves. Fixed a mystery leak, hard to track down, but ultimately solved by tightening a few screws around the cuvette head. A matter of 0.8 ppm C02 variation…

Saturday–01/08/11

Friday 01/07/11

• Job/Post-doc applications.
• Bolivia analysis.
• Sensors up and running (took 3 days to set up).
• Prepping for LiCOR work starting tomorrow through late next week.
• Road delineation in the Peruvian, Bolivian and Brazilian Amazon.

Have identified nearly 5000 km of roads within our study area in the year 1986. This will also be conducted for 2000 and 2008.

Friday 12/17/2010

New upper elevation location install of microclimate sensors.

One tired pup heading home from a long day in the field.

Thursday 12/16/2010

 

 

Success. The preliminary version of the image merging procedure is functional. This will allow for highly controlled stacking of SLC-OFF Landsat imagery (2003+).

Tuesday–12/15/2010

IDL programming with help from D. Fanning. He has a great website for IDL reference available at: http://www.dfanning.com/

Image pointers are a great way to define new variables within FOR loops. Crucial for reading in multiple Landsat images of varying ZYZ dimensions.

Tuesday–12/14/2010

Sweet! Made the cover!

Monday 12/13/2010

heading out to take down our sensors with a bag full of climbing gear.

the most important knot you can learn for tree climbing. the anchor knot.

Finally up and getting to work.

Friday 12/03/2010

 

Drying desiccant. We’ve managed to convert most of our house into lab space!

Monday-Wednesday–12/01/10

• Job applications
• Bolivia article: introduction, research questions, methods, statistical analyses.

Friday–11/26/2010

Collecting LiCOR 6400 in the rain in the Laupahoehoe Experimental Forest.

Sat-Sun: 11/21/10

• Completing reviews for Bio. Cons., For Eco., and JAG.
• Organized LiCOR 6400 data, verified curves, leaf drying.
• Rebuilt LiCOR 6400, new tubes, seals and chemicals.
• Finished Scan Line Corrections for Amazon land cover change project (70+ Landsat images).
• Job applications: organization and preparation.

Example linear correlations between SIB2 Modeled and field collected Net Carbon Assimilation Rates. This includes data from both the light and C02 response curves.

A dynamic light response over a 3 minute period following a replicated sunfleck (1500 ppm transition from steady state at 20 ppm).

Friday 11/19/10

 

car parked at one of our study sites

setup pre rain

setup post rain

home again, changing the filters for the c02 cartridges, the only good use for cigarettes, cheap filters…

Wednesday–11/18/10

Job applications, curve collection.

Improved LiCOR 6400 handle, reduces stress on attachment screws and allows for clipping a carabineer when collecting canopy measurements.

Second tripod used to collect location offsets from the differentially corrected stake. This particular offset allowed for curve collection in a forest gap 35 meters from the stake.

11/16/10–Tuesday

• Completed a job application.
• All Landsat images AutoMCU processed.
• Field work complete for the day, sensors up and running.

Interior forest sensors up and running

Sat-Sunday: 11/13-14/10

Field work setting up microclimate stations

Job applications

Hidden up there are two microclimate stations we spent 3 hours tree climbing to set up. Beautiful!

Thursday-Friday: 11/11-12/10

Thursday collecting LiCOR curves.

Notice abundant invasive species, including ginger, raspberry and guava, all within one small gap within one of the better native forests around. bummer!

At home the leaves are scanned at 600 dpi for area measurement.

Then oven dried at 55-65 C for 48 hours. Following this the samples are stored with desiccant until they are weighed, then analyzed for isotopes and nutrients.

Monday/Tuesday–11/10/10

1. LiCOR data collection
2. Checking on microclimate stations
3. Programming IDL procedure for writing large images (8-20 GB)
4. Job applications

An understory microclimate station with LiCOR 6400 collecting a C02 response curve in the background.

The Trupulse 3D laser system with LiCOR 6400 taking curve measurements in background.

My field assistant, Kai Turbo, taking a break.

Wednesday–11/10/10

• Article review
• Submitted job application
• Fixed computer issue resulting in frequent crashing
• Rebuilt LiCOR 6400, had leaks at tube seals and throughout the cuvette head. Ran full diagnostics. Runs perfect now! However, the tubes are a bit short so will look into ordering a replacement set.