MCC GIS Summer Internship
Weeks 5 & 6 of studying at University of Maine - Orono with CAFS at the Wheatland Geospatial Lab
Week 5
7/11 Monday
Drone take off
Today we met up with Dave Sandilands for another drone data collection demonstration. We walked from Nutting Hall to the Woodland lot, and gave us a more detailed explanation of how drone flights work. We utilized a smaller DJI Phantom drone for a 5-min flight over the Woodland lot.
Drone Landing
This demo was very similar to the one Peter gave us in Fort Kent, however this time we would be able to work with the data collected from the drone. Despite our 5-min flight time, the amount of data collected took a few hours to process.
Later, Bryon took the drone data and imported it into Global Mapper, where we could visualize the images taken and stitched together.
Casmir taking photos of a tree, photos which will be used to create a 3-D image.
After the drone flight, Tony and Dave had us work on a 3-D mapping of a tree using images taken from a phone. Placing reference targets at five points on the tree at DBH and 1m above, 70 photos were taken from 15ft away around the permiter of the trunk. Casmir took these images and uploaded them into Agisoft, which is a software that processes photogrammetric data and converts it to spatial 3-D images.
After uploading the images, targets were georeferenced to stitch together a 3-D photo, which allowed matching data points to be converted to a point cloud. From this, only the tree trunk being analyzed was selected, and all other points removed. Next, a mesh cloud was made from the trunk points, and all holes in the mesh were filled to create a 3D image of just the trunk. From this, the average biomass and volume could be calculated.
Initial photos uploaded to Agisoft. In the main window, the target points can be seen and digitally identified. Some had to be manually identified, but the process was not tedious. All images taken of the trees can be seen at the bottom.
Here, we can see the progression from point cloud data being refined and meshed to create 3D imagery of the tree trunk. Excess data was filtered out to leave the trunk by itself. Once isolated, the holes in the mesh were filled and the mass and volume were digitally calculated within Agisoft.
Drone Flight
7/12 Tuesday
Categories used to define elements within the image for interpretation.
Today we worked on remote sensing labs created by Tony for his GIS classes.. The first was an image interpretation workshop, where we had to identify tree species, height class, and define crown closure from Sentinel satellite imagery. Images were visualized in false color, which helped to highlight different characteristics of each type of tree grouping.
Images elements such as tone/color, and the spatial arrangements (including size, shape, and pattern) of the defined tone were used to identify distinct species.
The false color imagery was compared to the base imagery, which looked different due to temporal factors, but could easily define which trees were hardwood or softwood (based off of their cover during the fall/winter imagery). Because we were using images of an active forest, there was no guarantee that two images taken at different times would look the same, due to harvesting of certain areas.
The satellitae image was split up into groupings of similar species. In the second image, we can clearly see grouping of white pines in C57, as their crown shape is more star-like compared to other tree types. Height class was defined by studying the pattern of their shadow; longer more pronounced shadows indicate higher tree tops. Crown closure for C57 would be defined as moderate, as there is mostly coverage with some spots of visibility to the forest floor. Plot C55 to the left has a less dense crown class, while C79b to its left has a crown class that would be closer to full.
7/13 Wednesday
We used Avenza Maps to track our progress on the town streets. The blue line is areas covered in previous work days, while the green line is what Bryon and Casmir covered on Wednesday.
Today was a field work day, originally scheduled for the Penobscot Experimental Forest, but due to some team members falling ill, Bryon and Casmir got to work with Dave Ludwig on his Masters project.
His project had us working with the Town of Orono to identify and classify trees on the public streets. The main goal is to find trees that are in poor health, having issues with pests/insects, or ones that could interfere with power lines. Parameters included measuring all trees within 15 feet of a roadway.
Our experience in the PEF with CFI plots helped us, as we had to find the DBH of trees at least 1 inch in diameter, and ones that were growing in a tree-like habitat (not unmaintained brushy areas with shrubs and bushes). Most trees were easily accessible, but some of the trees were much larger than any tree we had previously measured. For trees in brushy areas, they had to have a DBH of at least 5 inches.
All data was recorded digitally within the iTree application. Here is some of the categories for data collection:
- Tree species
- GPS location
- Tree status (alive/dead)
- Crown condition & light exposure
- Tree height
- Stem count
- Management info
- Sidewalk condition (heave or no-heave)
- Maintanence recommendation
- Priority
Bryon demostrates taking a DBH measurement, while Casmir measures the tree distance from the roadway.
7/14 Thursday
7/15 Friday
On Friday morning, we met with Rajeev for a demonstration of a different type of LI-cor machine. This particular instrument is able to provide analysis on tree canopy cover and leaf area index.
Rajeev set up two separate wands that had special fisheye lenses at the end: one that would be placed in an open field to view the open sky, which we set as our "control" . The other wand was held by us and used to stand under a tree at 5 different points and then compare those images to the control, and therefore able to calculate the predicted tree canopy cover.
Example of the images that would be seen from the LIcor fisheye lens. Image 1 would be considered the control image, and all subsequent images would be compared to the control to define its canopy cover and leaf area index.
Mon-Wednesday 7/18-7/20
Week 6
7/18 Monday
Monday was an early morning for us as we got a cab at 6:00am so that we would be able to get to Acadia in time for Peter to pick us up and take us to the work site. Peter picked us up from the Hull's Cove visitors center and we ascended Cadillac Mountain to the work site at the summit. Before getting into the work Peter walked us and the crew that was already there around the immediate area showing us larger versions of the things that we would be lookig for in the regeneration plots. The Schoodic team was monitoring these regen plots that had been placed/planted 3-5 years prior. The job was to identify what was in the plot and see how it has been developing over that period of time.
We used square shapes of PVC that had string fed through them to make a grid (seen above). These are called Quadrats and they are the tools used to identify the edges of the plot areas. Within the larger plot there were nails inserted at the four corners of where each quadrat would be placed to measure the plot. These were often times challenging to fine but finding 2 or 3 of them provided enough bering to be able to place the quadrat. The area within the quadrat was what was being observed. The goal was to identify each type of plant in the quadrat and count the number of them that were inside the quadrat and then estimate the percent area that each plant species covered. We had a list of species identified during the previous time the monitoring took place and we used that as a basis for what should be there this time. So as to keep the team on pace Casmir and Bryon took the role of recorders. They had to enter the number of each species within the plot area as well as the estimated coverage of each species. The counting and estimating was done by the members of the Schoodic team as they were much more familiar with identifying the small plants.
The first day ended with Peter taking Bryon and Casmir to the Motel where they had rooms reserved for the night. Bryon and Casmir had their first seafood experience of the trip that evening when they went to the restaurant across the street from their motel. Both ordered Maine Lobster Rolls and thoroughly enjoyed them.
7/19 Tuesday
Day 2 at Acadia started in the morning when Jon picked the two of them up from the motel. After getting a park pass from the Hull's Cove Visitors Center it was off to Bar Harbor to get some sandwiches for lunch before going back up Cadillac Mountain to do more work with the Schoodic team. This time the monitoring was being done closer to where many of the tourists like to go in an area the Schoodic team likes to call "The Fishbowl." As the name would suggest the work being done in this area tended to draw more attention from visitors and the team was asked questions by visitors often during the time they worked here.
USGS marker denoting the summit of Cadillac Mountain.
Jon, Tessa, Casmir, and Bryon decided to take a short hike to the top of Cadillac Mountain and take in the sights. After searching for a couple minutes, Jon found the USGS marker denoting the highest point of Cadillac Mountain.
After going to the true summit of the mountain everyone got back in the car to go back down the mountain. Bryon, Casmir, Jon, and Tessa briefly went back to Bar Harbor to look around for a short time before taking the hour and 15 minute drive from Bar Harbor to the Schoodic Institute where they would stay the night. Jon took Bryon, Casmir, and Tessa out to eat at a local restaurant. Jon got a Lobster roll and shared it with Tessa. Bryon and Casmir both got Scallops and both agreed that they were some of the best that they had ever eaten.
That evening Bryon and Casmir decided to go see the sunset from Schoodic Point, the very tip of the Schoodic Peninsula. The entire point was made up of giant slabs of Granite which people were allowed to walk out on. There was a little time before the show was to begin, so Bryon and Casmir explored the rocky coast for a bit finding a massive Crevase in the stone that saw the seawater rushing in and splashing up the walls. The sunset was magnificent even though the sun disappeared behind the clouds part-way through. When the sun went behind the clouds the light went upward turning the clouds and sky above beautiful shades of orange and yellow.
Before returning to Orono Bryon, Casmir, Jon, and Tessa decided to take advantage of the low tide and walk to Little Moose Island right off the shore of Schoodic Peninsula. At low tide the island is accessable by foot. They explored a little bit before returning to the mainland, but not before Bryon was able to catch a few crabs.
All returned to Orono an had a meeting with Kasey at 2pm.
7/20 Wednesday
In the morning Bryon decided to get up before the crack of dawn in order to see the sunrise. He ventured down to the beach right outside the entrance to the Schoodic Institute and waited patiently, despite the mosquitos, for the sun to rise. The sun crested over a hill right between the peninsula and Little Moose Island and was just as beautiful as the sunset from the evening prior.
7/21 Thursday
On Thursday Bryon, Casmir, and Jon met with Aaron Weiskittel the one who runs CFRU and helped Jon secure the grant that made the internship possible.
Summer Research Presentations
Bryon and Casmir presented their summer research to a group of faculty and graduate students. You can view their presentations (Aug 4) here: https://ensemble.itec.suny.edu/Watch/a6C4BoSd - Bryon Carroll: MCC GIS Internship - Casmir Brown (starts at 22:00): Cloud Masking with a Machine Learning Algorithm Advisor: Kasey Legaard
Casmir Brown's final poster.
Bryon Carrol's final poster.
A huge thanks to everyone for giving your time and expertise to support the 2022 NSF Skills Training in Advanced Research & Technology (START) Supplemental Funding Request for ATE at Monroe Community College (Award #1955256) with IUCRC Phase 3 at University of Maine - Center for Advanced Forestry Systems (CAFS). Additional project details are summarized by Meg below.