This collaboration is inspired by a deep interest in the way people perceive non-human animals and questioning the similarities that can be drawn between humans and all living creatures. Leslie Ruckman and I created SurveillAnts, as a means to provoke people’s curiosity about non-human life, emergent intelligence, and systemic patterns. We wish to evoke that by revealing the unseen.


In SurveillAnts, there are three distinct ways to see the unseen. First, through the structure we have built, we allow a view into the unseen, underground tunnels of ants. Second, through computer vision tracking and projection mapping, we visualize the chemical pheromone trails ants leave behind as they explore their space in real time. Third, we record and collect the paths of the ants in order visualize movements over time. Through the data collected, viewers are invited to explore the movement of the ants over time, revealing colorful drawings and potentially, unseen patterns in the daily activities of ants.


Tunnels that were dug by the ants.

Trails in real time.

Trail’s history.


As a neuroscientist and an artist my Goal is to combine these two words that define me – art and science. I am very interested in how we, as humans, use other animals as a tool in order to experiment and to accumulate more knowledge about our world. Recently I became interested in ants as lab animals. They are perceived as insignificant or as pests and it made me want to change the way we see them and to evoke curiosity regarding them.



The ants were ordered online and were put into test tubes in an envelope.

the ants are clenching to a granola bar.


Throughout this process, we had 4 different homes for the ants.

The ant-farm near an industrial box that enabled us to first track them from a top view.

Building the first prototype ant-house.

The second prototype.

From right to left: the first prototype box and the ant’s current house.

The last house is a 15-15 inch box made of acrylic plastic and sand that will be stand alone and viewable on all sides. Living inside is a colony of about 35-40 red harvester ants. We are visualizing the trails they leave behind through projection mapping using a projector and webcam mounted and hung above the enclosure from the ceiling.

Updated bill of materials:


How it Works:

We are using a Logitech webcam to feed into Processing where we found a great Open CV library by Julien Gachadoat and an image processing blur function by Mario Klingeman. Through modifications to the blob detection function, we are able to identify and track individual ants, give them id’s and have them draw in different random colors on the screen.

Since ants are so small, we needed to maximize the sensing capabilities of at the camera. We did this in two ways: one the design of the Ants’ environment, and secondly through image processing. In the code, we are using THRESHOLD filter in order to clean unnecessary noise. The following picture shows what data the code is processing:

The data the code processes, what seems as black dots is ants.

With trails



Below is what the sketch looks like that we are projecting.


We are projection mapping this sketch onto the Ants with a program called Madmapper. This is how we are able to produce the final result of a line following an Ant.


To allow users some control over how much time is being revealed, we built a custom controller using potentiometer knobs and a button.



Sand was added to the knob enclosure in order to have similar aesthetic to the ant’s house. We will hopefully add to each knob an ant head as depicted in the following image.


Finally, each time the program runs, we are automatically recording each Ants position over time and capturing this into a CSV. We’ve created a couple of files with set amounts of time or can access the history of the current file in order to conjure up all the past Ant trails.


Next Steps:

  • We would like to be able to draw all maps at the same time. Currently, only one map can be drawn.
  • We would like to add opacity element to characterise the time differences.
  • Get more Data: Ideally, we could collect ant paths from longer periods of time. The longest data-set we have now is 23 minutes, but exploring days of data would be super interesting.
  • Perfect the Interaction and Code: We would like to continue to develop the interaction of drawing over time and the use of knobs or other types of inputs to provide the audience with some means to explore the ants’ behaviors.
  • Put it on a Pedestal: Literally, we would like for people to walk around the piece and see if from all sides so we’ll need to build a ceiling mounted projector rig and a custom podium.
  • Take it to the Third Dimension: It would be cool to incorporate data about their tunneling habits and otherwise include their activity underground in a better way.
  • Make Physical Prints: If all goes well for the show, we may print some of the physical drawings to hang around the space.

Categories: Computational Media, Physical Computing

Leave a Reply

Your email address will not be published. Required fields are marked *