This is a project I pursued as part of a team in the course “ENAS 115: Engineering, Innovation, and Design”, a freshman engineering survey seminar, in which we designed, implemented, and delivered a tailor-made product for our clients at Yale School of Forestry & Environmental Studies.

The goal of our clients was to acquire a system that could aid researchers in accurately estimate the size of hippopotamus populations in Mara River, Tanzania. A tedious, extremely time-consuming, and dangerous process traditionally performed by human researchers, obtaining hippo population estimates is crucial in estimating hippo excrement, which directly affects the ecosystem as well as the life of nearby village populations. Problems with manually counting hippos include: 1) there are a LOT of them, 2) they are often indistinguishable from each other in the river, or partially/fully immersed, 3) the waters are muddy and often completely dark (see pictures below), and 4) it is dangerous to count them on land.

Our product was a combination of hardware and software solutions, utilizing a drone to be flied over the Mara river; two cameras, a regular and a thermal one, to capture the view of hippos, attached to the drone; an integrated speaker on the drone (to drive hippos out of the water); and an image stitching and analysis process to facilitate combining the camera views and counting the observed hippos. Below, I present some images of the environment, the final product, and the process of testing (performed on campus).