I'm not an expert yet in the field and I have some questions. I have some data of birds and drones taken from a radar. I want to create a classifier that differentiates them. At first I'm trying and approach over the differences of the trajectories of these tracks.
My first problem arises when it comes to normalising the data. It is well known that target trajectories can be of different lengths. Moreover, these trajectories can be closer or further away from the sensor, directly affecting how well the trajectory is filtered. Over the last few weeks I have been looking for ways to normalise the data. The most interesting one I have seen is "centre-max" normalization, but I think my model will be affected by the lengths of the trajectories.
I have also thought about the idea of transforming the trajectory of each of the data into images. Is there any reference about transforming time series into images and that this transformation does not take into account the length of the time series. Also, will these transformation find covered patterns in the trajectory that we cannot think of?
