Proper inertial navigation is hard to do, particularly on a lightweight platform such as a drone. The location or distance calculated by such a system is produced by integrating speed and acceleration over time, each signal having a significant noise component that needs to be filtered out. The net result is that the errors in the system gradually build over time, leading to steady state error. This is why any inertial platform needs periodically aligned.
Modern drones, like the DJI, use optical flow and visual odometry to determine position alongside GPS. Inertial navigation isn't used as such, the accelerometer and gyro sensors being used for the stability control. For optical flow, cameras capture images of the surrounding environment and motion within that image is estimated to detect global and individual motion vectors that can be used to calculate motion and cues about depth of field.