We have been waiting for a while to post and have found time now, only after a hectic semester up in New England. We wanted to go over what the Ecosynth and the Harvard University - Urban Wild Lab Penny White Project covered while we were down in Baltimore during the Spring of 2014.
First, let me say this. The team at Ecosynth is doing some cutting edge stuff; and by that I don't mean the flying a copter and putting into Photoscan. There are many other groups and private companies that are doing that. It’s more of the approach and atmosphere that accompanies Erl Ellis's lab. Ecosynth dont only have dominion over the logistics of flying but are pushing the edge of data processing landscaoe ecology. In this post I will cover the data processing since I am the guy in our lab doing the data processing. My time in Baltimore was mostly spent learning precisely that. I’ll let my colleagues, Manuel and Michalis cover the flying and mission planning part, which Ecosynth did an amazing job at training us also.
We arrived to Baltimore on a Thursday night and started early in the morning on Friday. Dana, Steven, William, Jonathan and Erl greeted us that morning and were ready to go. We covered simultaneously the logics of flying the copters and setting up to efficiently process the information we gathered. We had requested Ecosynth to have a special track of data processing because we wanted to have a full understanding of what it would take to process the data and to be able to manipulate it in post-processing workflows. Ecosynth responded by having their top guys, Jonathan and William, come and explain to us how to setup both the open source and proprietary programs to process the data, and presented us the theory behind the programs. Jonathan in particular did a great job of the elements of photo-telemetry and how they form the backbone to get a 3D point cloud through Ecosynther and Photoscan. We had tons of question which relate to our research interest and the whole team engaged these questions. Questions like what kind of geospatial information can we further extract from what the copter and the camera are already gathers, what are other method that information could be recorded without the copter, creating Digital Terrain Models, color spectrum analysis and selection and so on.
These guys have quite some knowledge and are ready to think out the box when it comes to data processing and that is because they know their material. This is why we were very excited to hear how their open source program, Ecosynther, worked. Jonathan explained the basics by going over Noah Snavely’s Computer vision paper. Later we worked with William, Dana, and Jonathan to process some test flight we did with them. Our workshop ended with having in hand a full set of point clouds that we processed with the flights we did that weekend. Overall, the whole series of workshops have setup us up to do our research work in Harvard University’s Arnold Arbortum in collaboration with Prof. Peter del Tredici.
Since the training, we have processed some of our information gathered, which we will post in the future. And again Econsynth has been there to help us troubleshoot both our flight missions and processing the data. We ran into some problems using Ecosynther and into some bugs but William and Jonathan were very responsive and open to help take our troubleshooting into improving the Ecosynther program.
Overall, I would highly recommend getting trained with these guys. They are top notched, friendly, and dedicated. We would not think twice to train again with them again. So to anyone out there considering training with Ecosynth, DO IT!
Keep up the good work guys.
Harvard University - Urban Wild Lab (Penny White Grant Project)
Héctor, Michalis & Manuel