• State
  • Locked Locked
  • Replies 1 reply
  • Subscribers 30 subscribers
  • Views 239 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

IWR6843: Overhead VOD demo configuration for small room

Jukka Pajunen
Prodigy 10 points
Part Number: IWR6843

Hi,

we are using overhead VOD demo for experimenting how well we can track people inside a room (tracking is done by custom code operating on generated point clouds). People do not move a lot and hence are most of the time static. Room is small ~2x2x2m.

Problems start to arise when there are multiple people inside the room and point cloud generated by the demo focuses usually on one or two persons for multiple consecutive frames. I have tried researching documentation and code for possible ways to combat this but I have not had success yet. We have tried other demos but they seem to perform worse. 

I was wondering if there would be pointers for good resources or other ideas. I would be especially interested if it is possible to increase number of point cloud points the overhead VOD demo generates as I have not been able to accomplish this. 

  • 0
    TI__Mastermind 40025 points

    Hi, Jukka:

    For people counting, the seating is not stationary as in the car.  People can move their seats easily so the zone mapping method may not suitable for this demo.  Can you try the people counting demo instead.   

    Regarding your question that point clouds are focus on one or two person, it could be due to 1) the people are all seating close by and falls into the same range bin.  2) some are moving much more than the other.   To improve the separation, you can 1) mount the sensor a little bit lower so that the angle separation is larger.  2) mount the sensor so that some seats are closer and some seats are further i.e., separate them in range.     This suggestions are general to any demos, you can apply this tuning idea to people counting demo as well. 

    Best,

    Zigang