Wild Farms
follow.email
  • Blog
  • About
  • 2015 Conferences

Wild Farms.

A blog about wildlife, agriculture, and everything in between.

Fieldworking out 2- Let there be data!

4/30/2014

0 Comments

 
PictureMayra harvesting alfalfa
Time is flying! Yesterday the temperature hit 95 and tomorrow is the first day of May.

We've been busy in the #wildag experiment lately. It was a mad dash to harvest all of our alfalfa samples from the 19 fields we have set up exclosures in before the tractors started cutting. Harvesting involves setting up a quadrat that is slightly smaller than the exclosure itself (to avoid harvesting stems that may have been inside the exclosure on some visits and then outside on others), and then hand-cutting all of the alfalfa within it and bagging them up. In the lab, we then shake the L out of the samples (creating Afafa?) to knock weevils off of the stems and then sort and count all weevils into size classes. It's a labor-intensive job, but gives us a good chance of finding all of the weevils in our haystacks. The alfalfa samples are then weighed and dried in an oven for at least 2 weeks so we can eventually compare the number of weevils per pound of dry weight of alfalfa.

We're also in the middle of setting up most of the tomato exclosures. So far there haven't been any pest insects on the young plants, but as the weather warms up we're expecting to start seeing some damage.


I'm lucky to have a team of amazing interns and mentors who kept the project going while I was in the Great Smoky Mountains for a week-long Smith Fellows retreat. All of the interns are going to be adding their thoughts to this blog in the next few weeks- so keep an eye out for their bios to go up, and for their posts. I'm excited to see what they have to say!

Click on the 'Read More' link below to see the latest storify of our fieldwork adventures


Read More
0 Comments

Tech Post: Estimating Vegetation Cover with Photoshop

4/1/2014

2 Comments

 
Picture
Selection process for estimating vegetation cover. Upper left: original image. Bottom: Image with shadows and highlights removed. Upper Right: Selected areas of alfalfa after selection process and manual erasing of weeds.
This post is a bit more technical than most of what you'll find on this blog, but I've been facing the predicament of having to estimate the initial vegetation cover in my alfalfa plots and have figured out a way to do it using photoshop. Here's my how to, for anyone who faces a similar task.

If you have suggestions for a better way to estimate the vegetation cover for photos like the one I've used as an example here, I'd love to hear about it. Email me, or add comments below.
PictureInitial alfalfa photograph: alfalfa, weeds, dirt.
Step 1- Take photos looking straight down at vegetation. Import into photoshop. I'm using Photoshop CS4, 11.0.2

PictureSame photo with highlights and shadows removed.
Step 2- Remove shadows and highlights. Using the Image--> Adjustments --> Shadows/ Highlights tool and setting both ‘Shadow’ and ‘Highlight’ settings to 100%.

PictureUse the magic want tool to select color ranges with a +/- 10 threshold
Step 3- Select a color range using the magic wand tool. I set the magic want tool to a tolerance of 10 (selecting the 10 colors above and below the selected pixels) and selected 10 points within the alfalfa around the photo (using additive sampling).

Step 4- Select and refine. Use the Select--> Similar tool to select all greens within the 10 shades above and below the colors selected with the magic wand tool. Then, use the Select-->Refine edge tool to expand the selection. I set the refine to the maximum radius, contrast and smoothing settings, a feather of 8.1px, and an expansion of 32% because alfalfa grows in clumps and the initial selection was only choosing the center of many of the clumps.
PictureUse the 'select similar' and 'refine edge' tools to expand the color selection across the whole image




Step 5- Select more. Use the quick selection tool to select any areas that the above steps missed.
Step 6-
Copy and paste the selected parts of the image into a new layer and turn off the background layer. You should only see the areas that were selected. Use the erase tool to erase any areas that should not be included in the pixel count (for me, these were areas in shadow that were picked up, as well as weeds that I didn't want to count).
PictureSelected area on transparent background, 'erase' tool to touch up, and expanded histogram tool showing pixel count for 'selected layer'.


Step 7- Open the histogram box and make sure you're looking at 'Expanded View'. You should see a box below the histogram with data including a pixel count. Toggle between Source: Entire Image (for the pixel count for the original image), and Source: Selected Layer (for the cutout of the selected vegetation). Divide the latter by the former and you'll get the proportion of the original image that was selected in steps 3-6. Add to your spreadsheet and you're done.

2 Comments

    Sara

    Conservationist, explorer, 2013 Smith Fellow, amateur birder and wine enthusiast

    Archives

    November 2014
    September 2014
    July 2014
    June 2014
    May 2014
    April 2014
    March 2014
    February 2014
    January 2014
    December 2013
    November 2013
    October 2013
    September 2013

    Categories

    All
    Alfalfa
    Almonds
    Art
    Birds
    Ecosystem Services
    Fieldwork
    Hummingbirds
    Pollination
    Science + Life
    Teaching
    Tech
    Tomatoes

    RSS Feed


Picture
Wild Farms
Email us
Proudly powered by Weebly