The video foundation of a smart city: fusing data matters more than adding cameras
The bottleneck in city-scale projects has moved from camera count to data governance — getting video to corroborate access control, vehicle and sensor data.
PoE power and cabling standards: the decisive variable in mid-size project cost
One cable for both power and data greatly simplifies installation, but power budget, cable gauge and run length must be worked out in advance.
Hardening cameras: why default passwords and firmware updates keep being ignored
The risk to networked cameras rarely comes from sophisticated attacks, but from factory passwords, firmware left unpatched for years, and direct exposure to the public internet.
Video data compliance: from storing footage to auditing every access
Video images count as sensitive personal information. Retention limits, access control and operation logs are becoming acceptance requirements for new projects.
After H.265: how fast H.266 will actually arrive in surveillance
A new codec can save meaningful bitrate at equal quality, but the pace of adoption is set by the decoding ecosystem and front-end cost.
Low-light imaging evolves: choosing between starlight sensors and full-colour night vision
Night-time image quality is the most common complaint about surveillance systems. Sensor size, aperture, illumination and ISP tuning decide the result together.
Compute moves to the camera: edge processing becomes the new battleground
Running AI analysis at the front end cuts bandwidth and central compute load, but it raises the bar on thermal design, power budget and firmware maintenance.
GB/T 28181 plus ONVIF is becoming a hard requirement in device selection
As surveillance systems grow larger and platforms are consolidated, whether a device speaks both the Chinese national standard and open protocols decides future expansion and maintenance cost.
Large models enter video surveillance: from seeing to understanding
The value of video surveillance is shifting from recorded evidence toward real-time understanding, as multimodal models give cameras a semantic grasp of the scene.
