Does the data look way off? Give your Telraam a bit of time to calibrate itself. A few days will be required.
Check the camera position. The camera may have dropped a little so the image is no longer good enough to count. For instructions, see: https://telraam.zendesk.com/hc/en-be/articles/360026468392-Camera-position-checking-images.
Do you still see very strange things? Mail us on firstname.lastname@example.org. Describe the problem as clearly as possible.
Telraam is under development, so it is possible that some of the data presented is not (yet) completely accurate. We are working on automatic solutions to filter out bad data, but it is not yet in production.
What kind of strange things generally come up?
On certain days and times, the distribution of cars/cyclists/trucks changes completely.
There is a range of possible causes and it can vary from one Telraam to another. For example, it can be caused by the position of the sun, which can mean that shadows make cars/cyclists seem bigger. Telraam detects the whole of a vehicle and its shadow as a single object which means that Telraam could assign it to the wrong category
But I do have a lot of trucks in my street, according to Telraam. Is that right?
Probably not. The dividing line between car and truck is somewhat approximate when it comes to delivery vans. It is however possible that large cars (SUVs, pick-ups, etc.) are counted as trucks while small, low vans might be counted as cars. Telraams which are newly active, use a generic (and therefore fairly inaccurate) way of making the distinction. Some Telraams will thus count far too many (or far too few) trucks. This improves once there is a few weeks’ data available for each Telraam. Such an update happens roughly monthly. The distinction between cars and trucks will continuously improve the more data is available, so keep on counting! Read more about the cars/large vehicle distinction
Cyclists, pedestrians, cargo bikes, groups of cyclists,… how does Telraam handle that?
The distinction between bicycles and pedestrians is not simple to make, particularly when pedestrians and/or cyclists are in groups. The recognition of the class (cycle/pedestrian/car/truck) is currently performed on the basis of size and speed. Cyclists on foot or pedestrians with buggies are generally counted as cycles, joggers are faster but will be counted as pedestrians. Some (fast) cargo bikes or bike + trailer combinations will also be counted as cars. Groups of cyclists form a single large object and are thus a special case which Telraam finds difficult to distinguish and will therefore generally be counted as cars.
Here again, the more data we collect, the better Telraam will eventually be able to make the distinction.
What about traffic jams?
If there is a (serious) traffic jam (if there is no movement for more than a few minutes) then Telraam does not usually work well. Cars are not suddenly counted as trucks. The Telraam just does not count very well. That will come out as data labelled as “poor quality” for that time period.