ForestWatchers Lets Anyone Monitor A Patch of Forest

teleyinex writes "ForestWatchers.net is a citizen project with the goal of making it possible for anyone (locals, volunteers, NGOs, governments, etc), anywhere in the world, to monitor selected patches of forest across the globe, almost in real-time, using a computer connected to the Internet. The project has recently released a first alpha web application (built using the open source crowdsourcing PyBossa framework) where volunteers can participate by classifying satellite images of one area of the Amazon basin."

Comments

Rather simple task

[kiehlster]

Poking around at the contributions that users are expected to submit, this seems like a job that Watson or even a simpler machine could be taught to do without much human intervention. One could easily write an algorithm that chooses the most colorful image in the list of candidates. All I found myself doing was looking at a bunch of white cloudy images next to a couple greener images and an occasional black image. It's quite easy to use some color analysis to eliminate ultra-white and ultra-black images from the groups. Unless there's actual white structures on the surface, I don't see how this job should be tasked to a mechanical turk system. This thing also needs to learn how to prioritize image choices (best/top to worst/bottom) based on nearby approved images, not by date. It seemed like I was always scrolling down for images from a specific zone because the clear satellite image for that zone happened on a later date.