I am looking for real datasets to validate and compare mutliple target tracking methods, for example radar positions of multiple airplanes models, positions & velocities of vehicles.
If multiple models are used to target tracking, it can represent an advantage (for example in airplane application : bombers, fighters, liners....
Radar is not the only interesting sensor. Datasets provided by other kind of applications would be very useful.
Hello, dear friend to visit the site now
https://www.ncdc.noaa.gov/oa/radar/radardata.html
Good Luck
Hello dear Janhangir,
thanks for the shared link. However, I'm looking for radar datasets concerning target tracking (e.g. boats, planes, or others) and their trajectories, which does not provide the link you shared me. But thanks for your answer !
https://www.researchgate.net/publication/271206929_1999_Brest_Joint_Target_Tracking_and_Classification
Data 1999 Brest Joint Target Tracking and Classification
Hello
I don't have any radar data sets to share nor any real data, but there are some challenging synthetic passive sonar data sets that can be tried. My colleagues and I will soon be publishing full details of a previously published realistic passive sonar tracking scenario that has numerous targets and a high false alarm density (and also a high probability of target detection for most of the scenario). There are actually two data sets: one for frequency - time and one for bearing - time. I will follow up with a post once this data has been made available. BTW this data is based on Brinkmann K., and J Hurka, “Narrowband Passive Sonar Tracking”, Proc. UDT Europe 2010, Hamburg, June 2010. I encourage people interested in multi-target tracking to try their hand on 1-D sonar data sets as this type of data is often considerably harder to track than 2-D (frequency bearing, range bearing or cartesian x, y) due to the higher effective clutter density, lower resolvability (e.g. during crossings) and rich dynamical behaviours (that look like "target manoeuvres"). This fact has largely been ignored in the mainstream published literature.
Dear Hugh and Graham,
thanks for your answers and the provided link to the interesting paper concerning a joint tracking and classification approach. It can be very helpful to compare your results with mine.
Graham, concerning your remarks, I will study your papers and a sonar application could also be imagined to validate my methods. Indeed, we build them in a general way and the application areas are multiple. Sonar tracking is an interesting problematic and I will study it in details and I will try to implement my method in this application.
I will contact you if I have some questions about this kind of data.
Thank you,
Benoît
Bonjour,
With a colleague, we have now uploaded some data files for a benchmark 4-target passive sonar tracking scenario. There is a full description of the scenario in the link below and the data sets including ground truth files have been published as supplementary resources. The experimental procedure should be entirely repeatable if further tests (e.g. Monte Carlo) need to be performed.
The data sets are freely available and can be cited according to the instructions in the description document. The title is "Benchmark 4-target passive sonar scenario description for 1-D tracking" by K. Tyson and G. Pulford.
Please let us know if you find this data useful or have any questions about the data generation process. Tell your colleagues that they are welcome to download this data and try out their new tracking techniques.
Cordialement,
Graham Pulford
Technical Report Benchmark 4-target passive sonar scenario description for 1-D tracking
Dear Graham,
thank you very much for these data files.
We will quickly evaluate our tracking methods on this set an we'll come back to you if we have additional questions.
Benoît
- Badges
- Science topic
- Similar topics
- Books