Automated optical inspection (AOI)
Check visible parts against agreed inspection rules.
Manufacturing
Industrial video analysis for evaluated camera feeds.
Lumetras Vision analyzes selected industrial camera feeds against configured rules. Teams can evaluate possible uses for operations, visual quality review, or security. Camera compatibility and results depend on the site. People must review detections before acting on them.
This silent product demonstration shows a configured rule flagging a possible missing glove and sending a WhatsApp alert. A detection needs human review.
The footage and any people, records, or detections shown are illustrative. They are not measured results.
Demo walkthrough
How the example works
This is an illustrative product demonstration, not a measured result.
01
A supervisor describes a condition to review, such as a worker who may not be wearing gloves. The rule and its limits need to be checked for the site.
02
The team selects a camera feed and configures the rule. Camera access, stream quality, and processing rate depend on the equipment and site setup.
03
The system analyzes video for the configured condition. Results can vary with camera angle, lighting, motion, and other site conditions.
04
A possible event is recorded with available details for review. A detection is not a confirmed safety violation.
05
A person may review the event and decide whether it needs follow-up. Identification is a separate option and requires an appropriate site review.
06
If configured, an alert can be sent through a supported channel. The example shown is a product demonstration, not a measured response time.
How it works
Evaluate a camera, then review the results
01
Review the camera setup. Lumetras Vision may work with IP camera streams over RTSP or ONVIF. Compatibility, access, and video quality require evaluation at the site.
02
Describe a condition to check in plain language. The team reviews the rule and its limits before use. Examples include a possible missing hard hat or an item left in a marked area.
03
Configure and evaluate the rule on selected footage or camera feeds. Results depend on camera placement, lighting, motion, and other site conditions.
04
Review detections and configure notifications. Available details and delivery options depend on setup. Human review is needed before acting on a detection.
Factory AI automation / quality
Camera rules for part inspections
Describe the condition a camera should check. Test sample images under the lighting used on your production line. Agree on acceptance criteria before using a rule for automated optical inspection (AOI).
Computer vision quality control
Flag missing parts or incorrect assembly for an operator to review.
Surface defect detection
Test sample parts for scratches or finish defects. Validate the model before production use.
Missed-step alerts
Flag a visible step that may have been skipped. Operators review the footage.
Detection results depend on site conditions. Operators must review alerts and test performance on their own lines.
AI video analytics / industrial security
Review site activity with camera rules
Define rules for visible events in storage areas or site boundaries. Check camera coverage and any required inventory-system connections before setup. Security staff review the alerts and decide what action to take.
Inventory movement review
Compare unusual goods movement with inventory records.
Perimeter entry alerts
Define boundaries in a camera view. Flag entry for review.
Warehouse asset tracking
Log the last observed location of visible pallets or equipment. Camera coverage affects results.
Tool checkout review
Review camera activity at checkout and return areas. Compare it with tool records.
Restricted-area alerts
Flag entry into a defined area. Camera rules do not replace physical access controls.
Unusual activity alerts
Set rules for events such as after-hours activity. People review the alerts.
Scrap yard and dispatch review
Flag unusual movement for a security team to investigate. An alert does not prove theft.
Possible evaluation areas
Review specific camera use cases
Personal protective equipment. A site evaluation can assess whether a camera view is suitable for checking visible items such as hard hats, vests, gloves, or eye protection.
Restricted zones. A configured rule may flag possible entry into a marked area. Teams must review detections and determine the appropriate response.
Vehicle and pedestrian activity. A site evaluation can assess whether camera placement supports review of vehicle routes and nearby people.
Blocked exits and fire lanes. A configured rule may flag objects in a marked area. A person should verify the event before taking action.
Line activity. Video analysis may help teams review line stoppages or idle periods where camera views and operating conditions allow.
Spills and housekeeping. A site evaluation can determine whether the camera view supports review of visible floor conditions or material placement.
Access activity. Teams can evaluate camera views for specific access-related events. Lumetras Vision does not by itself confirm a person's intent or identity.
A site evaluation is needed to determine whether a camera view and rule suit a use case. Do not treat detections as a substitute for human review.
Where the alert goes
The alert reaches people where they already are
WhatsApp.
Send configured notifications to selected recipients through WhatsApp.
Microsoft Teams.
Send configured notifications to a selected Teams channel.
Email.
Send configured notifications by email.
Webhook.
Send event data to a configured endpoint. A connection to an MES, CMMS, ERP, or ticketing system needs setup and testing.
Event records.
Configured detections can be recorded for review. Available records depend on the system setup and retention settings.
Camera setup
Check camera support at your site
RTSP and ONVIF streams may be supported. Camera compatibility depends on stream access, network conditions, and video quality. Confirm support during a site evaluation.
Choose on-site or cloud processing based on your infrastructure, privacy requirements, and operating needs. Confirm which deployment options fit your site.
Configure and test each camera and rule. Review results under site conditions before using detections.
IP cameras
RTSP / ONVIF
On site
Cloud
Privacy and review
Set clear rules for video use
Video analysis can affect workers and visitors. Define its purpose, scope, access, and review process before use.
Identification is a separate option. Evaluate its need, limits, and impact with the people responsible for workplace privacy before enabling it.
Define which cameras and areas are in scope. Review those settings with the people responsible for the site.
Set retention and access rules for clips and event records based on your requirements and system configuration.
Decide who can view video and event records. Confirm what access records are available in your deployment.
Explain the purpose and scope of monitoring to workers. Your organization is responsible for its workplace policies and notices.
Frequently asked questions
Questions we get asked
Do we have to replace our cameras?
Not if their feeds are compatible. We check network access and stream format before setup.
How long does it take to get running on one camera?
Setup depends on stream access and the rules you need. We review camera compatibility and network permissions before giving a setup estimate.
Does it need training data from our plant?
Some rules use existing models for people or protective equipment. Rules for a particular part or machine state may need sample footage from your plant. We confirm this during the site review.
How accurate is it?
Each detection has a confidence score. You choose the alert threshold. Test it against your own footage and record missed events and false alerts. Lighting, steam, and camera placement can affect results.
What happens if the network goes down?
On-site processing can continue and queue alerts. Cloud processing needs a connection. Available storage limits the alert queue. Existing camera recording remains separate from this analysis.
Is this replacing our safety team?
No. It flags camera events for your team to review. People decide what action to take.
What is a vision-language model in factory AI automation?
It reads a written instruction alongside images. Lumetras Vision uses this to apply monitoring rules to compatible camera feeds. Test each rule under site conditions.
Can Lumetras Vision perform automated optical inspection and surface defect detection?
These use cases need a site trial. Test representative parts under the intended lighting. Agree on acceptance criteria before using the rule in production.
Can AI video analytics detect warehouse shrinkage or theft?
Rules can flag unusual goods movement. An alert does not prove theft. Your team must review the footage and compare relevant inventory or access records.
Can it detect perimeter breaches and restricted zone access?
Rules can flag entry into defined zones. They do not replace locks, access controls, or human investigation.
Can it track assets and manage a tool crib in real time?
Camera views can log visible pallet or tool movement. Tracking across cameras needs suitable coverage. Tool records may need a separate system connection.
Next step
Evaluate one camera feed
Contact us with the camera type and a use case. We can discuss site requirements, compatibility, and how to review results before deciding whether to proceed.