I get asked this a lot, usually by someone holding a phone photo of a monitor showing a blurry number plate. The short answer is that enhancement can bring out detail that is faintly present in the recording, but it cannot invent detail that was never captured. If the plate occupies twelve pixels, no software on earth will read it.
The first thing to ask for is the original export from the DVR or NVR, not a screenshot and not a WhatsApp forward. Every re-encoding throws away information. A native export keeps the full resolution and, importantly, gives you multiple consecutive frames. Stacking several frames of a slowly moving object can genuinely improve legibility, because noise is random but the real detail is consistent.
Once you have the original, a few techniques help in order of usefulness: contrast-limited adaptive histogram equalisation (CLAHE) to bring up dark regions, a mild deblur such as a Wiener filter if the motion is uniform, and frame averaging where possible. If the footage was photographed off a screen, a frequency-domain notch filter can remove the moiré pattern.
Set expectations early. Treat anything enhanced as an investigative lead, not proof. Police and courts will want the untouched original, and so should you.
Markdown for AI
A clean, structured version of this page for AI assistants and LLMs.
We use cookies to ensure you have the best browsing experience on our website. By using our site, you
acknowledge that you have read and understood our
Cookie Policy &
Privacy Policy.
I get asked this a lot, usually by someone holding a phone photo of a monitor showing a blurry number plate. The short answer is that enhancement can bring out detail that is faintly present in the recording, but it cannot invent detail that was never captured. If the plate occupies twelve pixels, no software on earth will read it.
The first thing to ask for is the original export from the DVR or NVR, not a screenshot and not a WhatsApp forward. Every re-encoding throws away information. A native export keeps the full resolution and, importantly, gives you multiple consecutive frames. Stacking several frames of a slowly moving object can genuinely improve legibility, because noise is random but the real detail is consistent.
Once you have the original, a few techniques help in order of usefulness: contrast-limited adaptive histogram equalisation (CLAHE) to bring up dark regions, a mild deblur such as a Wiener filter if the motion is uniform, and frame averaging where possible. If the footage was photographed off a screen, a frequency-domain notch filter can remove the moiré pattern.
Set expectations early. Treat anything enhanced as an investigative lead, not proof. Police and courts will want the untouched original, and so should you.