Kalman Filter with GIS data

[SANJAYSGOWDA]

Hello everybody, I am trying to solve the tracking error in the Kalman Filter. The Kalman Filter is good and have almost nicer prediction in the straight road, but when the car(for ref, I am doing tracking on the Pano image, My data is obtained by the Ladybug Camera and I only have Panoramic image, with this I have camera Extrinsic information like (Easting, Northing, Altitude, Yaw, Pitch and Roll of the camera). My Plan is to write a personalized Kalman Filter wherein I can give a correction step , since I know the YPR of the car , I need to corelate it to the shift in the Kalman Filter prediction. I have detection from another custom trained model (So detection part is solved). i need to just track it, somebody help