Sea Turtle Drone Detector

The Problem

In SE Asia local villagers have taken to stockpiling endangered sea turtles for the illegal wildlife trade markets. They keep them in floating fish pens until they have sufficient, and then secret phone calls are made to get the poaching vessel to sail over and pick them up. These turtles are the proverbial golden egg-laying turtles - they are reproductively mature and are the future of the turtle population. Losing them means losing an entire stock. Not good. If we can fly a low cost drone over a large area and video as it goes, and then download the video and process it through software to detect floating rectangular fish pens (probably hidden up estuaries where nobody will find them) we can provide the locations to authorities and stamp this out once and for all.

Our Proposal

We want to build a low-cost drone, under USD 500, which can fly 3 to 4 hours in an autonomous flight plan, collecting quality video imagery that can be analysed by a computer to detect floating rectangular fish pens which might be where sea turtles are being stockpiled. The software would need to be able to select for rectangular objects in the sea, and have some programmable feature allowing us to select a buffer from nearby inhabited areas (it is unlikely the poachers would stockpile the turtles in front of the village). There are suitable high density styrofoam UAVs on the market and good autopilot units and open-source software for the actual UAV. The software would be another issue altogether though.

We Assume that...

Sea turtles need to breathe, and need to be in water so they don't dehydrate. Fish pens are common, so this is the most likely place to keep them.

Sea turtles are legally protected. So you don't keep them in your village. You have to store them out of sight. The sea pens therefore are likely hidden up mangrove creeks away from the village.

Locals know it is illegal, so it is unlikely the practice will be common knowledge outside of the community. Thus we need to be able to detect the pens remotely.

Law enforcement agencies will become adept and will use the platform once developed.

Constraints to Overcome

Accessing remote locations and covering a massive geographical area, and detecting the floating fish pens remotely without the need to review hundreds of hours of footage manually. At present we can only use boats. These are slow, often can't reach the areas where turtles are stockpiled, can't cover large areas, and cost immense amounts of money in terms of fuel, manpower, and platform cost. Local law enforcement officers need low cost solutions that are easy to deploy, easy to fix, easy to build, easy to replace, easy to understand, and which give them the edge over illegal wildlife traffickers.

Current Work

Put a stop to illegal stockpiling of endangered turtle by local SE Asian villagers, destined for the China and Vietnam markets

Current Needs

Develop the software algorithms to detect the fish pens and a user interface to upload video and set parameters for filters and then set the software going - possibly with a video view so the user can determine that the software is rapidly scanning through the imagery - and with the ability to pause and rewind and double check something that might have been a fish pen but which was not picked up by the software. This is where we need the most help