Object detection with radar
Edit1: I did now some research and so I realized that I have to describe the problem in more detail. I have a radar sensor which provides a pointCloud to me with the measured points. Objects that are more than 100m away are normaly represented by only one point in the point cloud. Now I am not sure if it is really a object or a measurement error. So the solution is to watch over for example 5 samples and if all 5 sample includes the point, with a small change because of the movement, than there is a object. This should also work if within the 5 samples is one point missing because of an measurement error.
I found during my research the kalman filter but i am not sure if it is the right one and how to use is. Thanks.
br Harald
Hi
I am using a radar sensor to detect moving objects. The speed of the moving objects (e.g. cars and persons) is up to 140kmh. I get a pointCloud from my radar sensor and if an object is moving with for example 50kmh I see the track of the object. Now I need a algorithm that is able to detect an moving object also when the object is only represented by one point. So my question is if there is any recommendation for a specific filter or any tip that could help me to do that! Thanks.
BR Harald
Hey Harald,
What kind of radar are you using? What's the hardware and did you get an open source to make it play in ROS? I'm interested in knowing that!
Hey Ish_Harie, I know this is an old topic, but did you find a suitable radar sensor? I couldn't find when I was looking so me and some colleagues decided to make one (with ROS integration) which we will be launching soon (radariq.io). We currently have pointcloud support enabled but are tinkering with doing full-object detection on-module (kalman filtering, speed, and acceleration data) what sort of information are/were you looking to get from the radar?