OpenStreetMap logo OpenStreetMap

kartonage's Diary

Recent diary entries

A quick reflection on my Mapillary contributions this year

Posted by kartonage on 6 December 2019 in English. Last updated on 23 September 2022.

Since the days are getting less than optimal for me to go out to record sequences, I took the time and reflected about what I had achieved this year so far. This diary might not be very informative or technically detailed but I just wanted to put down my thoughts on it in writing.

overview My coverage this year as per Mapillary. Since my activities spread all over the city of Berlin I tend to cross the inner city a little bit more.

Gear

This year I managed to intertwine my cycling habits and my contributions to OSM even more by purchasing an action cam that was finally satisfying my needs. While I already had a GoPro Hero5 Session in 2018 I was never really satisfied with the lack of internal GPS tracking and the additional steps to add my own gps tracks to the images. Especially with a lot of stop and go traffic inside the city I had to spend too much time processing the images and adding GPS data. It was actually somewhat of a mispurchase.

So with spring coming around I decided it could only be beneficial to look for a better cam. After some pondering, I decided to get a GoPro Hero 5 for a good price. I knew that the Hero5 Session had a quality I could live with and I wasn’t too annoyed by the GoPro antics. Additionally other action cam’s didn’t convince me or weren’t the size I wanted them to be.

And after about 4.000 km captured and 500.000 pictures taken, I’m generally satisfied with the quality of the Hero 5. I can usually make out anything on the images in the vicinity and beyond that I would argue that a survey might be more appropiate. The batteries last long enought for me so I don’t have to worry too much (usually between 1-2 hours). I bought some more cheaper rechargeable batteries for the GoPro which would probably last me about 8-10 hours. The 128 GB SD card I use has been plenty for now with a capacity for roughly 48.000 images.

innercity While most major roads are covered there are still quite a lot of side streets untouched. A lot of them are surfaced with sett or worse, which is why I don’t naturally ride through them.

Setup

Since 2013 I have been riding a stripped-down Cube bicycle for tens of thousands of kilometres and after a somewhat rough year where I had to replace a lot of parts I hope I can still go a few more years to make the 100k km with this bicycle complete. It’s pretty light-weight and is basically down to the bone, no-frills.

So my setup is minimal and the GoPro mount is usually not even recognised. I found a mount that’s basically nothing more than the mount to screw in the GoPro case and it is mounted only via zip ties. It’s not optimal since I could do without a screw to steady the GoPro to the mount. I’m used to simple inserts for my tracker or lights but I haven’t really found something I liked for the GoPro yet.

The position can be bit low to the ground at times, especially when bigger cars block the sight to the sidewalks, but in general it’s good enough. Usually my front tire will appear at the bottom but I don’t think this matters much.

General thoughts and outlook

The last few months I have been building routes through streets I haven’t captured yet, loading them into OsmAnd and following the constant callouts. I noticed that having a very tightly knit route will become very tedious quickly with a lot of turning and forced circles. After about 3 hours I would usually decide to just go a different more straight-forward route instead. I haven’t really found a way to motivate me to keep to my planned route yet. This year it has been more or less been a product of moving around naturally

Next year I have already planned a few routes to bigger towns outside of Berlin trying to cover them as much as I can in one go. Especially because there haven’t been any images at all in some of these towns. Depending on my personal/work situation I hope I can fill in the gaps and contribute even more images than this year.

southeastI try to go out of my way but there is still much more to cover.

I’m probably not going to aim for 100% coverage of Berlin or even close to that. Time is an issue and while some streets never change and would need to be covered only once, others change monthly and I still like chasing notes and other issues in Berlin. There also a couple of other contributors who have been very active. So I rather try to capture new long-term developments or major changes on roads with higher importance than to go into each and every little by-road. And in the end cycling is always my primary focus which is why only about 40% of my rides were recorded this year.

A side note: Although I get asked for directions or if I need help a lot when on foot, I have not yet encountered anyone who asked about the action cam. Not that I really mind, but I thought it might gather more (unwanted) attention than usual.


Naturally if you have any recommendations or suggestions feel free to write me.

Location: 52.521, 13.405

Beim herumstöbern in Berlin sind mir vereinzelt Kindertagesstätten aufgefallen, die sich direkt auf oder neben Straßenabschnitten befinden und nicht eindeutig einem Gelände oder Gebäude zuzuordnen sind.

Viele von diesen nodes scheinen nur aus name, note und amenity zu bestehen. Dabei sind die Namen öfter auch aus den Betreibern oder anderen abgekürzten Beschreibungen/Zugehörigkeiten zusammengesetzt. Sind durchweg schon so einige Jahre alt und bestimmt teilweise nicht mehr dort aufzufinden, wo auch immer sie dann in der Umgebung hingehören. Vielleicht ist es deswegen auch ein bekanntes Thema und war in der Vergangenheit noch zahlreicher aufzufinden.

Nachdem ich mich versuchsweise an einer Overpass-Abfrage dafür versucht habe, bin ich auf 274 Fälle gekommen. Vielleicht ein paar zu viel (mehr als 20 haben eindeutige Hausnummern), vielleicht ein paar zu wenig (5m Entfernung vielleicht nicht ausreichend). Über 200 sind es zumindest. Könnte man sich ja mal als Aufräumprojekt zurechtlegen. Gibt zumindest ein paar cluster, die sich gut überprüfen lassen könnten.

img