Although fragmentation is very much worth avoiding, first, usually you don't need to go as small as a 1300 MTU. 1400 usually is the typical Cisco recommendation.
Second, where and how you adjust MTU matters. For video, ideally you would set the reduced MTU on the source (camera) host. Assuming the router is the VPN endpoint (is it?), you adjust the MTU on the tunnel and it's also suggested to adjust TCP's MSS to MTU less 40. It's also suggested to enable PMTUD, on router.
Similar changes should be made on HQ side too.
Again, MTU adjustments are very worthwhile but packet loss is most often simply due to insufficient bandwidth, especially for video traffic. (BTW, fragmentation increases bandwidth need.)
If possible, if CCTV camera offers different quality or bandwidth usage rates, try the lowest setting, and see if you still have packet loss. If you don't, try next better setting. Repeat until packet loss again.
Whenever there's packet loss, you first need to identify where it's occurring. Once you do, you next identify exactly why it's occurring. Then you can work on mitigation.
Mitigation might be possible various ways, possibly not mutually exclusive. I've already mentioned the need for sufficient bandwidth, but if it's a transient need (common with video), it might be satisfied with prioritization of video traffic and/or additional buffering of non real time video.
Lastly, you note this is not an Internet issue. You know this how?
Keep in mind Internet is best effort. Unfortunately for some traffic kinds that require better than best effort, you need a WAN service provider, not an ISP (often same company can be both, but very different service offerings, and different costs).
