Perfect traffic data from video even in complex scenes with small objects and irregular movements!
Вставка
- Опубліковано 13 кві 2023
- Analyzing complex, large area scenes with many occlusions, small objects, and even irregular movements is possible with our TrafficSurvey! Our special 100% analysis provides you with research-grade data that can download to your computer and can analyze the way you need with a range of filtering, export, and visualization options. And if you need just basic traffic counts or OD matrix, our fully automated data extraction service has you covered too! Get a sample analysis of your video now at: ai.datafromsky.com/picker
Video courtesy of COWI: www.cowi.com/
We are DataFromSky and we do AI video analysis for traffic, parking, retail and safety. What exactly? We do offline traffic surveys that help redesign roads and cities. We can do analysis of traffic in real time to detect dangerous situations, improve traffic flow or detect driving violations. Want to detect the number of free and taken parking slots in an open area with a single camera? No problem! Count the number of people passing by your store?
What is special about us? Our framework uses AI and can turn any camera into a smart sensor. The cameras can then work together and create unique single trajectory data that can not only classify the object but also detect speed, stationary duration or distance between objects.
Give it a go! Get your video analysed: datafromsky.com/trafficsurvey/
Or download the demokit of the super easy to use FLOW framework and try to design your own live sensor in minutes here: datafromsky.com/download-flow/
#trafficanalysis #traffic #trafficdata #smartcities #analytics #trafficengineering #trafficsurvey #survey #DataFromSky #mobility #urbanmobility #innovation #machinelearning #artificialintelligence #video #research #ai #data #safety #roadsafety #trafficresearch #research #trafficcounts #vehiclecounter #drone #uav #dronetrafficdata - Наука та технологія
Get a free sample analysis at: ai.datafromsky.com/picker
Note that the output you see above is the 100% special analysis as mentioned in video description.
To make the image more visible, I would abbreviate the labels and put a legend. P - pedestrian, C - car, B - bicycle. For speed, rounding to maybe 1 decimal place and remove the units, and just indicate the unit in the corner. Also consider transparent labels.
You can add and remove parameters from the labels as needed. Legend might make sense in bigger scenes like this. Thank you for the suggestion.
Awesome! how did you get speed of all objects in frame??
By using georegistration intercom.help/datafromsky/en/articles/3526089-getting-speed-and-position-data-in-trafficsurvey-viewer-using-georegistration-how-to-guide
looks so smoooothh🤤
We are glad you like it!
How are pedestrians tracked when they are behind a building? There's a lot of this going on at the bottom of the picture. Are there other cameras?
Most likely these are predictions. If you look more closely, you can see how the algorithm makes mistakes in its predictions. Fantastic.
Hello we have some 1. Built-in prediction
2. User-managed prediction that can prolong the trajectories start and end points based on multiple parameters
3. 100% analysis that can further enhance the data. @athNick
It does Read out the memory of Our World that is why he can Track them😅😂
@@helloitsme1169 Note that the output you see above is the 100% special analysis as mentioned in video description.
How many people worked on this project and how long did it take?
Sorry, the marketing team does not have that information.
How is this tracking people behind objects?
Note that the output you see above is the 100% special analysis as mentioned in video description.
bro really put elevator music
Hey, why not!? XD Not the typical elevator music - "jazzy" style music but still chill.
i signed in just to comment "LOL"
@@NolenFelten Elevator music is the best.
This is sus. It sees objects behind obstacles without any problem.
Hello we have the following tools to help us achieve that:
1. Built-in trajectory prediction
2. User-managed trajectory prediction that can prolong the trajectories both start and end points based on multiple parameters
3. 100% analysis that can further enhance the data.