Most monitoring tells you a camera is online. Picture Watch tells you it can still see. It checks every camera's picture against its own history and, the moment one fades, hands you the proof.
Security said a vehicle-gate camera at a test site looked terrible at night. Every dashboard said it was healthy: online, recording, network clean. Only the picture had gone bad, and nothing watched the picture.
Picture Watch compares each camera with its own past, at the same hour on earlier days. When a picture clearly and lastingly drops, Signal Command opens an incident and attaches the evidence itself: Good, Degraded and Bad in one row.
Sep 29, 8 PM: a healthy night
Oct 3, 8 PM: where it starts to drop
Oct 23, 8 PM: the latest nightHaze, grain, a fogged dome: online and recording, but not seeing.
Three dated pictures in the incident, for the client or for the camera maker.
It uses the camera connection Signal Command already has.
A new site gets the answer on day one, in about three minutes.
For when your IT or security lead asks. Security first, then the rules.
Four times a day (2 AM, 8 AM, noon, 8 PM, site time) it reads one low-resolution still per camera from Verkada.
The picture is decoded in memory and measured: brightness, tonal range, edge sharpness. Then it is dropped.
Each hour is compared with the same camera at the same hour on earlier days, not with a fixed standard.
A drop must persist over several days and show at two hours. Dark scenes are not judged.
Only then: an incident with three labelled frames. It reopens, not duplicates, if it comes back.
Start with a conversation. If it's a fit, we run a free network assessment live with you and show you what Signal Command finds.