This is a guest post from Mapillary user Jorge López Fueyo @nokutu reporting on the state of the JOSM Mapillary plugin.
I'm Jorge, a 19 year old computer science student from Spain. I discovered OSM about one year ago. I really enjoyed making maps and when the opportunity to participate in Google's Summer of Code working with OSM appeared I was willing to do it. There were a couple of interesting projects for me but one of them caught my eye, the Mapillary plugin for JOSM. I started talking with the people in charge of the projects, who would be the mentors, and they answered all my questions. The more I knew about this particular project, the more interested I became.
When I was selected I felt kind of scared and excited. I didn't know how hard the programming could be, as my programming skills are quite limited. But as soon as I started, all this fear disappeared. There were many things I didn't know, but you can find all answers on the Internet if you ask the correct question. For example, I had never worked with threads or swing before, but it wasn't too difficult to learn.
Since I started, the plugin development has gone much further than I had expected. I've learned a lot of things really fast, as I needed them in order to progress with the plugin. A great thing about free and open source software, is that it allows one to stand on the shoulders of giants.
In the current version you can already see any image hosted at Mapillary as a new layer of JOSM and browse through sequences, or jump to nearby pictures of other sequences.
Image by @neogeografen
It’s also possible to import and export pictures.
There is also support to move pictures and improve the direction the camera was pointing. At the moment this only makes sense with imported pictures, for geocoding them and reexporting. If you do that with your new pictures before uploading them to Mapillary, they will have better positioning than what your GPS and phone compass was able to provide. The guys at Mapillary are frantically working on enabling support for feeding back this information through their API. We’re hoping that in the long run this will become possible for (un)blurring and traffic sign detection as well. At the moment I'm working on filtering of images based on time, photographer and presence of traffic signs. Later on I will improve working with traffic signs.
Be aware that this is still an early beta and things might crash or damage data - use the plugin at your own risk and back up your images first!Short into to the current state of the JOSM Mapillary plugin