Image Server
Yes, you will get time stamps. What you get is described in the GenericByteData format: https://doc.developer.milestonesys.com/html/index.html?base=mipgenericbytedata/main.html&tree=tree_3.html
Mobile
In the Mobile server protocol there is a field called TimeStampUtc (in ms).
It represents the actual timestamp of the Transcoded video frame.
If however Direct Streaming is used, the field represents the timestamp of the first video frame of the following mp4 fragment.
Other timestamps in the fragmented file should be reconstructed from the timestamp of the first frame and relative timestamp increase in the MP4 file.
In all the cases I would not recommend to be implemented LPR integration trough the Mobile server.
It has limited performance and could not handle large amount of video streams, especially in high resolution and transcoding.
DS in other hand introduces another layer of complexity forcing the client to understand structure of the MP4 files.
ONVIF Bridge
In the RTSP server (of the ONVIF Bridge) video frame timestamp could be reconstructed from the RTP packet timestamp and RTSP sender report correspondence.
Every RTP packet has a RTP timestamp, which is in 90 kHz units.
From time to time (for sure once after the first RTP video frame is sent) RTSP server sends RTCP sender report packet.
It contains correspondence between RTP time and NTP time.
This NTP time is the actual video frame timestamp for this particular RTP time.
So for every next RTP video frame the actual NTP timestamp is calculated from the RTP timestamp and the difference with the last RTP timestamp for which was received RTCP sender report with NTP timestamp.
This is also described in this question:
https://developer.milestonesys.com/s/question/0D53X00006tSt04SAC/when-getting-video-streaming-from-onvif-bridge-for-video-anlaytics-does-it-support-to-get-the-time-information-from-the-frame-the-lpr-system-is-getting-video-streaming-thru-onvif-bridge-and-they-system-needs-to-know-the-time-information-of-frame
If the you aim to integrate toward the ONVIF Bridge, it is better to be used TCP as underlying transport. Otherwise, in case of UDP, it could lose frames. ONVIF Bridge has better performance as it doesn’t do any transcoding and could handle easily more 100 video streams. It however works only with H.264 cameras.
If I remember well by reconstruction of the actual frame time from the NTP timestamp (low and high 32 bits) have to be used rounding in order to not be loosed accuracy.