Camera placement is the single most important factor in LPR accuracy. A well-placed, correctly configured camera produces clean reads. A poorly placed one produces call-backs. This guide covers what you need to know for barrier gate and urban roadside installs.
You control two of them. GateIQ handles the third.
The software that reads plate characters. GateIQ uses a professional-grade OCR engine tuned for North American and international plates, with AI-assisted disambiguation for partial reads. You don't configure this — it works out of the box.
Frame rate, shutter speed, resolution, and IR illumination. A blurred, overexposed, or low-contrast image cannot be corrected in software. Getting these settings right at the camera is the only way to ensure clean input to the OCR engine.
Where you mount the camera and what angle it points at determines everything downstream. The best camera in the world, improperly placed, will underperform a budget camera that is correctly positioned.
The minimum camera settings you need for reliable plate reads in each situation.
|
Barrier / Gate
HOA, parking, car wash
|
Urban / Street
Lot entry, enforcement
|
|
|---|---|---|
|
Vehicle speed
How fast cars are moving when the camera reads them
|
Stopped or near-stopped | Moving at street speed |
|
Photos per second
More photos = more chances to get a clean read as the car passes
|
At least 5 per second | At least 15 per second |
|
Freeze speed
How quickly each photo is taken — too slow and moving plates blur like a long-exposure photo
|
1/250 sec minimum | 1/500 sec minimum |
|
How far away the camera reads
Distance from camera to where it reads the plate
|
~13 ft (4 m) | ~40 ft (12 m) |
|
Lanes covered
|
One lane | One or two lanes |
|
Image quality (one lane, 10 ft wide)
Think of it like TV resolution — HD is enough for one lane
|
1280 × 960 (standard HD) | |
|
Image quality (two lanes, 20 ft wide)
Full HD covers a wider area without losing plate detail
|
1920 × 1080 (full HD) | |
Two angles matter. Here's what they mean in plain English.
Think of tilting your head to look at something on the floor. Tilt too far and you're looking straight down — a license plate photographed at a steep angle looks squished and the letters become hard to read.
Imagine trying to read a name tag on someone's chest. Face-on is easy. If they're turned sideways, the tag looks narrower and harder to read. Same principle with license plates — the more the camera is off to the side, the harder the letters are to read.
Four settings to configure before any GateIQ install.
Frame rate determines how many chances the OCR engine gets to read the plate as the vehicle passes through the field of view. More frames = more opportunities = higher confidence reads.
Higher resolution increases CPU load — don't set frame rate higher than your scenario requires.
A slow shutter speed causes motion blur on moving plates — the number one source of misreads in the field. The faster the vehicle, the faster the shutter needs to be.
Resolution and lens focal length work together to produce the right plate character size in the image. The OCR engine needs plate characters to be 25–35 pixels tall on USA plates. Too small and it misses detail; too large and you waste processing cycles.
Always select the correct lens focal length for your read distance. Don't use excess resolution — it increases CPU load without improving accuracy.
License plate reads at night or in direct sunlight require IR illumination to produce clear, evenly lit images. Most dedicated LPR cameras include built-in IR. Standard IP cameras often require an external IR illuminator.
Short read distance (~4 m). Vehicles stopping or near-stopped. Single lane. Most HOA, parking, and car wash gates fall here.
Medium read distance (~12 m). Moving vehicles at street speed. Parking enforcement, lot entries, and access lane monitoring.
Same rules whether you're doing a gate or a street-side install.
When a camera looks steeply down at a plate, the letters appear squished and the plate recognition engine starts making mistakes. The steeper the angle, the worse it gets. Keep the camera positioned so it's looking at the plate fairly level — not staring straight down at it like you're looking at your feet.
A camera that's looking from the side at a passing car will see a plate that looks narrower and more distorted — similar to trying to read a book that's angled away from you. Mount the camera so it's pointing mostly toward approaching vehicles, not across the lane.
Site conditions and vehicle factors beyond camera settings.
High-contrast objects in the frame behind vehicles can fool the OCR into trying to read them. Clear the read zone background of:
Never aim the camera directly at a light source. These will overexpose the image and wash out the plate:
Note: IR illumination directly at the camera (from another source) can also overexpose IR-sensitive cameras.
Some misreads are not a camera or software issue — they are physical:
Front plates are generally cleaner than rear plates — rear plates accumulate road spray. Use front-read where possible.
Run through this before mounting anything.
This checklist is interactive — click items to mark them complete.
GateIQ works with any RTSP-compatible IP camera. Starter is $3,500 one-time — hardware, software, and the OCR engine included.