Today we are happy to announce another feature requested by our faithful Mapillary users - traffic sign recognition. We have processed over 6 million images and detected all the traffic signs in the United States and Europe. From now on every image added in these regions will go through our system and will be categorized with traffic sign data.
Over the last year we have received a lot of feedback on our app. Some of the most asked for features related to how to improve photo collection, better reviewing before uploading on device, better browsing on device, more information on processing progress and what happens around you. After looking the same since we started, it is now time to update the app.
The purpose of this blog post is to welcome beginners to the Mapillary community and guide you on getting started with using Mapillary.
This is a guest post from Mapillary user Chris McNally on using video files for contributing to Mapillary. Until there is video upload support this workflow will help you process video files to create photos that you can submit to Mapillary. Thanks a lot to Chris for putting this together and posting it here!
New ways of following your progress and interacting are coming! We have just added the possibility to see interactions from you and other users - e.g. uploads, image processing status, comments, mentions, blur requests, edits, etc.
One of the most exciting things with working on Mapillary is that everything we do is global. Our apps are available everywhere and we make no restrictions on who can contribute from where. Being available globally from day one also presents some challenges.
One of the downsides of using simple devices for mapping the world is that the GPS accuracy is not always great, especially in cities with tall buildings. Since the start we have always wanted to correct this using image matching and we are now making progress in that area.
Mapillary is getting social! Today we’ve added a way to easily share images you like. Introducing: the share modal. Below you'll find a quick run through of the most important aspects.
Mapillary is increasingly used for many different scenarios, and the question of how to best cover a certain track or trail comes up a lot. In order to display it on a web page and provide the best view of a track all useful directions should be covered as much as possible.
There are a number of very capable, cheap Bluetooth GPS devices out there like this Garmin GLO which gets much better accuracy than a normal smartphone. Most of them work out-of-the-box with iOS, but on Android some setup steps are required to get them working.
Six months after our launch it is now time for an update of our policies. While our previous terms regulated what we can do with photos contributed to Mapillary we decided that more clarity and detail would be helpful on things like how you are allowed to use the service, your rights to photos, how we handle your personal data, copyright infringement and more.
Do you find yourself sometimes writing small scripts just to introspect some JSON? Say we want to get some image from the Mapillary API, the URL for getting GeoJSON from an image search
Filtering & Interface Improvements
At Mapillary, our main task is connecting the many photos uploaded to the service. We do this all the time, in the background, using computer vision and reasoners.
At Mapillary we do a lot of things with the photos you upload. One thing we do is to automatically blur license plates and peoples faces for privacy purposes. No license plate detector or face detector is perfect and sometimes we will incorrectly blur areas, sometimes we will miss important areas that needed blurring.
Today Mapillary - the project to provide a street level view of the world - arrived at the Windows Phone platform with the first release of the app for Windows Phone 8 / 8.1.
This past weekend we captured some data for the upcoming Traffic Jam Session hackathon here in Malmö. Since the event is about hacking public transportation data, we mapped some of the local bus lines. We used six Android phones running the Mapillary app that captured a total of 50GB of photos.
It has been six weeks since we announced that we were licensing our photos with Creative Commons. The license we chose was the Creative Commons Attribution-NonCommercial 4.0 License (CC BY-NC). Our reasoning was explained in the accompanying blog post.
Today, we finally got around to implement one of the most requested editing operations for Mapillary - the bulk-straightening for all images in a sequence.
When you are a startup, you don't have much time. In our case, we want to track some of our KPIs, as the numbers of users, images and power users over time (lots of other historic KPIs like user retention too, but that's another blog). We also want have a nice dashboard but not build a system for that.
Hi there, you probably have seen the disqus footers under the Mapillary pictures. While this has been working for a while (we are talking days here :), we now have integrated a polling on new blog posts that will mail the photographer of the image with any new updates on the discussions.
We got a complaint from a user who flagged an image by accident, see this issue.
As we are now entering a phase of Complaint Driven Development, this morning I found a
great little fix to this.
Angular.js which we use for the web client, does not come with a built-in confirmation dialogue. However, here is
a great little generic confirm-button directive that avoids the
onClick method and can be reused.
One thing we have been struggling with since starting Mapillary is the issue of licensing Mapillary photos.
The right stack for the job
One question we get a lot is if you can upload photos that you have taken with another camera, outside of our apps. Maybe you have some great street sequences taken with your DSLR or older photos that you would like to add.
We have been working with the OpenStreetMap community lately and we wanted to investigate how Mapillary can be used as a tool for some serious mapping.
Starting out Mapillary with the idea of crowdsourcing street view photos using smartphones we were fully aware of the limitations of the devices used. Typically, your phone will at best give position within a few meters accuracy, sometimes a lot more than a few meters.
The number one thing people have been asking us during our short history is "when will you have an Android app"?