A Robo-Weeder for the Marsh

The Problem

Coastal salt marshes are threatened by invasive species, such as phragmites and pepperweed[1]. As the marsh is weakened by the loss of biological diversity it becomes more vulnerable to storm damage which in turns increases threats to surrounding environments and structures. 1. http://www.bioone.org/doi/abs/10.1614/IPSM-D-09-00015.1

Our Proposal

A “smart” autonomous solar powered device that uses robotics and computer vision to selectively removes phragmites or other invasive species. In the context of the Con X Tech Prize, we will submit a proper hardware design and proof of concept of the computer vision software with the initial prize, and then create the device itself with the grand prize.

We Assume that...

The removal of invasive weeds from the marsh will aid and benefit conservation efforts

There will be ample sunshine during the growth season of the weeds to facilitate a solar powered autonomous device

Invasive species such as pepperweed can be detected via computer vision

The robotics of the device can be programmed in an autonomous fashion

Constraints to Overcome

Practically speaking, the biggest constraint is the time required of the conservation organizations to detect and remove these invasive weeds. Removing this daunting, seemingly futile task will be the breakthrough. Technically, the barriers are programming the behavior of the autonomous device, and writing the computer vision code to visually detect the invasive weeds.

Current Work

For the initial Con X prize, we would develop proof-of-concept software, a hardware design, and a full budget and work plan. For the final prize, we would build and deploy a prototype of the marsh roomba, and publish our findings.

Current Needs

We would benefit from mentoring from both estuary botanists and ecologists, as well as robotics and computer vision experts to consult with.