Making More Markers: Preparing for Summer Research

Here at Bee Lab, the last couple of weeks we have been planning and implementing projects for this next summer. More specifically, over the last couple years two of the major projects at bee lab are aerial drone mapping and pollinator visitation studies. To quickly summarize both, aerial drone mapping has involved taking aerial photos of the field station, stitching these images into larger maps, and running these maps through various programs to determine what flower species are present. The pollinator visitation project involves taking detailed observations of the number of bees and other pollinators that visit a plant, as well as information on the characteristics of each observed plant. As part of annual projects, this semester I have been designing new markers for both of these projects using feedback from the last year.

Problems and Goals

While these two projects may seem like significantly different ventures, because they happen over the same area, it was useful for use to design markers that would fit the needs of both. Although last year’s experiment used markers, what was discovered was that they were extremely difficult to see in the drone mapping images. As such we created a new system to address the following goals:

1.    Create markers that are visible from drone images (at 15-meter height), and can be uniquely identified

2.    Have markers point to a specific plant

3.    Build markers that are sturdy, but can be flipped over, as to not confuse bees with fluorescent and bright colors.

Figure 1: Old marker (on the left) and new marker (on the right)

As a little bit of a tangent, you might be wondering why exactly we need these markers to be so visible in our drone images. As mentioned, our plan is to run these maps through various algorithms to species type. However, in order to do so we need training data, which are smaller images that generally show exactly one known plant. Having easily identifiable markers would greatly speed up the process of creating accurate training data. 

 Building Prototypes

Looking at the tiny foam markers from last year, we immediately decided we needed something significantly larger and sturdier. Out of plywood, we cut 11 x 11 in squares as a prototype, and decided on a painting scheme. In order to have the markers both point to specific plants and be unique, we decided on three fluorescent spray paint colors that would show up well in photos, and created an “arrow” pattern that allowed us to make 12 different markers.

However, considering the intersection with the pollinator project, these markers also needed to be visible for someone walking around in the station. After obtaining 2 ft. wood stakes our initial plan was to simply nail or glue the two together. While for mapping this might have worked, for the pollinator project, there was worry that the bright colors would confuse or distract pollinators, effectively modifying any data we collected. Instead we painted the other side of the markers a more neutral brown, and drilled through both the top marker and post to insert a dowel, so the marker could be flipped.

Figure 2: Several designs of both the top and bottom of markers

Finishing Up and Placing Marker

Our final plan now done, we recruited the rest of the first years and turned bee lab into a marker making factory for about a week. We also designed and made corner posts using the same sort of design, except with taller posts, triangular markers (from cutting the same 11 in squares in half), and single colored tops.  With only a few minor setbacks, our markers are now out around the field station, ready for our experiments to begin! 

Figure 3: A marker in its natural habitat

Media Credits:

Photos taken by Michelle Lilly