Is camera-based attendance legally valid as a CLRA attendance record?

Camera-based attendance — whether from a face recognition terminal or CCTV-linked AI — is legally valid as an attendance record under CLRA provided it meets three conditions: (1) the record is identity-linked, meaning each event is tied to a specific named worker (not just a headcount); (2) the record is timestamped and tamper-evident; and (3) the data can be exported in a format that populates Form XVI (the CLRA muster roll). A standalone CCTV system that counts people but cannot attribute each entry to a named worker does not meet condition 1. A face recognition terminal linked to a CLMS meets all three conditions and produces a more defensible record than a manual register.

Can cameras alone verify overtime — or does a CLMS handle that?

Camera attendance records the punch-out time. Whether that punch-out time constitutes overtime — and whether that overtime is approved and payable — requires the CLMS. The CLMS knows the worker's contracted shift end time, the OT approval threshold, and whether a supervisor pre-approved overtime for that worker on that date. Without the CLMS, a late punch-out is just data: it cannot be used to validate or dispute a contractor's OT invoice. With the CLMS, the late punch-out is compared against the approved shift record, and any OT above the approved amount is flagged before it enters payroll.

What is the difference between a face recognition terminal and CCTV-based AI attendance?

A face recognition terminal requires workers to present themselves individually to a reader — one worker, one authentication, one attendance event. It works best at controlled entry points with moderate throughput. CCTV-based AI attendance uses cameras positioned at gates or production floor entries to detect and identify multiple workers simultaneously, without individual presentation. It handles higher throughput and is less disruptive at busy shift-start periods. The tradeoff: face recognition terminals typically have higher identity accuracy per event; CCTV AI handles throughput better but may require higher image quality and lighting consistency. Both can integrate with a CLMS; both produce identity-linked timestamped records.

Which cameras qualify as STQC-compliant for post-ban procurement in India?

From April 1, 2026, cameras used in government-linked projects must carry STQC (Standardisation Testing and Quality Certification) certification. Hikvision, Dahua, and most Chinese-origin brands are not STQC-certified and cannot be procured for compliant deployments. Approved alternatives include cameras from Indian manufacturers (CP Plus with STQC certification, Godrej Security Solutions) and non-Chinese international brands that have obtained STQC certification. Verify the specific model's certification on the STQC portal before procurement — brand-level certification does not extend to all models.

Can a factory replace its existing biometric readers with cameras for attendance?

Cameras and biometric readers serve overlapping but distinct functions. Face recognition terminals are purpose-built for attendance: they enrol workers individually, authenticate against a stored template, and produce a clean identity-linked event per worker. CCTV-based AI attendance handles throughput better but typically requires higher infrastructure investment (camera quality, lighting, AI analytics software). For most manufacturing deployments, the practical upgrade path is to add CCTV validation on top of existing biometric readers — using cameras as a cross-check layer — rather than replacing readers outright. Full camera-only deployments make more sense for new campuses where gate infrastructure is being designed from scratch.

How does InOps handle workers who are not yet enrolled in the biometric system?

Workers who are not enrolled in the biometric system — typically on their first day before induction — cannot authenticate at the gate. InOps CLMS holds the gate until the enrolment is complete: the worker is enrolled during the onboarding process (face or fingerprint capture, identity document scan, contractor assignment) and only then receives access. This prevents a common compliance gap: a contractor deploying a worker before their Form V and enrolment are complete. The gate enforces the onboarding sequence, not just the time of day.

← Back to blog

Biometrics

Camera-Based Attendance and CLRA Compliance in Manufacturing: What It Can and Can't Do

February 202611 min read

AI camera system in a manufacturing facility representing camera-based attendance and CLRA compliance

Camera-based attendance systems — face recognition terminals, CCTV-linked AI attendance, and biometric entry gates — solve the proxy attendance problem: they prove that a specific person was physically present at a specific time. For manufacturing plants managing contract workers, that is necessary but not sufficient. What a camera cannot do on its own is tell you whether the worker who just walked in has a valid Form V on file, whether their continuous attendance is approaching the 9-day CLRA threshold, or whether the hours they just accumulated will match the invoice their contractor agency sends at month-end.

That compliance intelligence layer is what a CLMS adds to camera attendance. InOps is the only platform that connects face recognition and CCTV-based attendance directly to the CLRA compliance stack — making every camera event an auditable compliance record, not just a headcount log.

What camera attendance systems do well

Face recognition terminals and AI camera attendance solve the identity verification problem reliably. A worker whose face matches an enrolled biometric record was physically present. Liveness detection prevents photo spoofing. CCTV-based AI can process multiple workers simultaneously at high-throughput gates without individual presentation. For a 1,000-worker plant with three entry gates and staggered shift starts, this matters.

The attendance record produced is timestamped, identity-linked, and tamper-proof. These are the properties that make biometric attendance superior to manual registers and badge-swipe systems for compliance purposes.

What camera attendance systems cannot do alone

A face recognition event records that Rajesh Kumar, Badge ID 4471, entered Gate 3 at 06:47 on Tuesday. The camera system does not know whether Rajesh Kumar's contractor — Shree Labour Contractors Pvt. Ltd. — has a current CLRA licence. It does not know whether Rajesh has worked 8 consecutive days without a weekly off. It does not know that the shift he just started will push him into overtime at hour 9, and that his contractor will invoice for that OT at double rate.

Standalone camera attendance systems (Truein, eSSL, SalaryBox, generic facial recognition tools) capture the identity event but stop there. The compliance questions that define a principal employer's legal exposure under the Contract Labour (Regulation & Abolition) Act require data that lives outside the attendance system — unless the attendance system is the CLMS.

The CLRA compliance questions camera attendance must answer

For camera attendance to serve as a CLRA compliance record, it must answer these questions at the moment of each gate event — or within the same payroll cycle: Is this worker's contractor licensed under CLRA for this establishment? Has Form V been issued for this contractor? Is this worker's continuous attendance approaching the weekly off threshold? What is this worker's minimum wage category and is their current shift accruing at the correct rate? Is this shift's OT pre-approved, or is it unauthorised overtime that the contractor will invoice?

None of these questions can be answered from an attendance event log alone. They require the attendance data to be joined with contractor records, licence databases, compliance rule engines, and payroll calculations. That is what InOps CLMS does.

How InOps connects camera attendance to the compliance stack

Every face recognition or CCTV-based attendance event captured by InOps-integrated hardware triggers a compliance check in the CLMS. The check runs in real time: contractor licence status, Form V/XIII currency, continuous attendance counter, shift rules and OT thresholds, and minimum wage applicability. Exceptions are surfaced immediately — a worker from an unlicensed contractor triggers a flag before their shift generates any payable hours.

The attendance record produced is not just a timestamp — it carries the worker's contractor ID, licence status at time of entry, site and zone assignment, shift type, and any compliance flags. At month-end, this record is the source of truth for invoice reconciliation, statutory reporting, and CLRA audit evidence.

OT verification: the case cameras alone cannot close

Overtime fraud in Indian manufacturing almost always follows a predictable pattern: a worker stays late, a supervisor approves it informally, and the contractor invoices for the hours at the OT premium rate. The camera attendance record shows the worker's late punch-out. But without a CLMS that knows the worker's contracted shift hours and has an OT approval workflow, the late punch-out is just data — it cannot be used to validate or dispute the invoice.

InOps CLMS calculates payable OT from the biometric punch-out time against the worker's registered shift end time. OT above the configured threshold requires digital approval in the system before it becomes payable. When the contractor submits an OT invoice, it is compared line-by-line against the CLMS-approved OT record — discrepancies are flagged before payment. This is OT verification that camera attendance enables but cannot perform alone.

Existing CCTV infrastructure and the post-Chinese camera ban landscape

Following India's STQC certification mandate effective April 1, 2026, Hikvision and Dahua cameras — which dominate the existing installed base in Indian manufacturing — are no longer compliant for new procurement. Many plants are now evaluating replacement camera infrastructure and, in parallel, asking whether their new cameras can serve double duty for attendance.

InOps CCTV attendance integrates with STQC-compliant IP camera hardware from approved Indian and non-Chinese manufacturers. Plants replacing their Chinese camera fleet can simultaneously solve the compliance gap and deploy camera-based attendance — with InOps providing the AI analytics layer that turns a security camera into a CLRA-linked attendance record.

What this means for your camera attendance strategy

If you are evaluating camera-based attendance for a manufacturing site with contract workers, the question is not which face recognition terminal has the best accuracy. Accuracy matters, but it is a commodity — every enterprise-grade terminal meets the accuracy bar for industrial use. The question is what happens after the face is recognised.

Truein records the attendance. InOps records the attendance and routes the event through a CLRA compliance engine that tells you whether that worker is compliant, whether their shift is generating liability, and whether their contractor's invoice at month-end will match the record. That is the difference between attendance data and compliance data.

Frequently asked questions

Is camera-based attendance legally valid as a CLRA attendance record?: Camera-based attendance — whether from a face recognition terminal or CCTV-linked AI — is legally valid as an attendance record under CLRA provided it meets three conditions: (1) the record is identity-linked, meaning each event is tied to a specific named worker (not just a headcount); (2) the record is timestamped and tamper-evident; and (3) the data can be exported in a format that populates Form XVI (the CLRA muster roll). A standalone CCTV system that counts people but cannot attribute each entry to a named worker does not meet condition 1. A face recognition terminal linked to a CLMS meets all three conditions and produces a more defensible record than a manual register.
Can cameras alone verify overtime — or does a CLMS handle that?: Camera attendance records the punch-out time. Whether that punch-out time constitutes overtime — and whether that overtime is approved and payable — requires the CLMS. The CLMS knows the worker's contracted shift end time, the OT approval threshold, and whether a supervisor pre-approved overtime for that worker on that date. Without the CLMS, a late punch-out is just data: it cannot be used to validate or dispute a contractor's OT invoice. With the CLMS, the late punch-out is compared against the approved shift record, and any OT above the approved amount is flagged before it enters payroll.
What is the difference between a face recognition terminal and CCTV-based AI attendance?: A face recognition terminal requires workers to present themselves individually to a reader — one worker, one authentication, one attendance event. It works best at controlled entry points with moderate throughput. CCTV-based AI attendance uses cameras positioned at gates or production floor entries to detect and identify multiple workers simultaneously, without individual presentation. It handles higher throughput and is less disruptive at busy shift-start periods. The tradeoff: face recognition terminals typically have higher identity accuracy per event; CCTV AI handles throughput better but may require higher image quality and lighting consistency. Both can integrate with a CLMS; both produce identity-linked timestamped records.
Which cameras qualify as STQC-compliant for post-ban procurement in India?: From April 1, 2026, cameras used in government-linked projects must carry STQC (Standardisation Testing and Quality Certification) certification. Hikvision, Dahua, and most Chinese-origin brands are not STQC-certified and cannot be procured for compliant deployments. Approved alternatives include cameras from Indian manufacturers (CP Plus with STQC certification, Godrej Security Solutions) and non-Chinese international brands that have obtained STQC certification. Verify the specific model's certification on the STQC portal before procurement — brand-level certification does not extend to all models.
Can a factory replace its existing biometric readers with cameras for attendance?: Cameras and biometric readers serve overlapping but distinct functions. Face recognition terminals are purpose-built for attendance: they enrol workers individually, authenticate against a stored template, and produce a clean identity-linked event per worker. CCTV-based AI attendance handles throughput better but typically requires higher infrastructure investment (camera quality, lighting, AI analytics software). For most manufacturing deployments, the practical upgrade path is to add CCTV validation on top of existing biometric readers — using cameras as a cross-check layer — rather than replacing readers outright. Full camera-only deployments make more sense for new campuses where gate infrastructure is being designed from scratch.
How does InOps handle workers who are not yet enrolled in the biometric system?: Workers who are not enrolled in the biometric system — typically on their first day before induction — cannot authenticate at the gate. InOps CLMS holds the gate until the enrolment is complete: the worker is enrolled during the onboarding process (face or fingerprint capture, identity document scan, contractor assignment) and only then receives access. This prevents a common compliance gap: a contractor deploying a worker before their Form V and enrolment are complete. The gate enforces the onboarding sequence, not just the time of day.