Knowledge & data
Turn the cameras you already have into useful data.
People counting, attendance, number plates, safety and quality checks, analyzed on-site so video never has to leave your premises.
At a glance
Computer vision & CCTV analytics
- Uses your existing IP cameras where possible
- Video analyzed on-site, not uploaded
- Accuracy measured on your own footage
- Privacy reviewed before installation
Delivered from New Baneshwor, Kathmandu, on-site across the valley and remotely across Nepal.
In short
What is computer vision?
Computer vision is AI that interprets images and video. It can count people, read vehicle number plates, check whether safety gear is being worn, spot defects on a production line, or read meters and labels. In most cases it runs on the CCTV cameras you already have, with a small computer on site doing the analysis. Bit Microsystems designs and installs computer vision systems for businesses in Nepal.
What is included
Everything needed to make it work in practice
One team covers the full job, so there are no gaps between suppliers.
Site and camera survey
We check camera positions, resolution and lighting, and tell you what will and will not work before you spend anything.
Model selection and training
An existing model where one fits, or one trained on your own footage where your situation is specific.
On-site processing
A small edge computer analyzes the video locally, so footage does not need to be sent to the cloud.
Dashboard and alerts
Counts, trends and events in a simple dashboard, with alerts by message when something needs attention.
Integration
Results fed into your attendance, gate, ERP or reporting system.
Privacy review
Signage, consent, data retention and access rules agreed before the system is switched on.
Where it is used
Who this helps, and how
Typical situations we design for in Kathmandu and across Nepal.
Retail and showrooms
Footfall, busy hours, queue length and which areas customers actually visit.
Factories and workshops
Defect detection on the line and checks that helmets, gloves and other safety gear are worn.
Parking and gates
Number plate reading for entry logs, visitor management and parking billing.
Offices, schools and sites
Camera-based attendance with consent, and headcounts for safety and capacity.
How we work
From first conversation to a system your team uses
- 01
Survey
We visit the site, review sample footage and define exactly what needs to be detected.
- 02
Proof on your footage
We run the model on recordings from your cameras and report the measured accuracy.
- 03
Install
The edge device, dashboard and alerts are installed and connected to your systems.
- 04
Tune
We adjust for lighting, seasons and layout changes, and retrain if accuracy drifts.
Is it right for you?
An honest fit check
We would rather tell you now than after you have paid for the wrong thing.
A good fit if
- You already have cameras and someone watches or reviews footage manually.
- You need counts or checks that are too tedious for people to do consistently.
- A missed defect or safety breach is costly.
- You want facts about how your space is used, not impressions.
Probably not the right choice if
- Your cameras are low resolution or badly placed. We will say so at the survey, and better cameras may have to come first.
- You want to identify or track people without their knowledge. We do not build covert surveillance.
- You need perfect accuracy. Real-world vision systems are reliable but not flawless, and the process around them must allow for that.
Tools and technology we work with
Questions
Frequently asked
Straight answers to what people ask us most about this service.
Does computer vision work with our existing CCTV cameras?
Usually, if they are IP cameras that provide a standard video stream and the resolution and angle suit the task. Older analog systems are more limited. We confirm this at the site survey before quoting.
Does the system need internet?
No. The analysis runs on a small computer on your premises. Internet is only needed if you want to view the dashboard or receive alerts remotely.
Can it read Nepali number plates?
Yes, though accuracy varies. Nepal has both older plates in Devanagari and newer embossed plates, and results depend on camera angle, lighting, speed and the condition of the plate. We test on footage from your own gate and report the measured accuracy before installation.
Is face recognition attendance appropriate for our office?
It can be, but a face is biometric data and should be treated carefully: tell people clearly, get their consent, store the data locally, limit who can access it, and delete it when someone leaves. Where a card or fingerprint system would do the job, we will say so.
How accurate is computer vision?
It depends on the task and the conditions. Counting people through a doorway is highly reliable; recognizing small defects in poor light is harder. We always measure accuracy on your own footage first and state it in writing.
Related
Often combined with
Start a conversation
Thinking about computer vision & CCTV analytics?
Tell us what you want to achieve and what you have today. We will reply with a clear plan, a realistic timeline and an estimate.