What Are Your Cameras Actually Seeing at 2am?
A hospital pharmacy in Abu Dhabi discovered a controlled substances discrepancy during a routine Thursday morning audit. The cabinet had been accessed properly — a valid staff credential, a logged door event, a camera pointing directly at the location. When the security manager pulled the footage, it showed: a corridor. Ceiling. Wall. The camera was angled six degrees too far upward, and four years of “surveillance” had been capturing the top of a door frame in sharp 2-megapixel detail.
The cabinet was visible in the left edge of the frame. The person accessing it was a silhouette.
This isn’t a story about inadequate camera specifications. It’s a story about the difference between a camera on a wall and a surveillance system that actually works. The camera was genuine, expensive hardware. The network was properly configured. Nobody had thought through what the camera was actually pointed at, what it could genuinely identify from that angle, and what would happen when the footage was needed for something that mattered.
Avigilon security cameras address this problem at three levels simultaneously. The imaging hardware is among the best available — wide dynamic range, low-light sensitivity, resolutions up to 4K Ultra HD. The onboard AI analyses what the camera sees in real time, classifying objects, tracking movement, and generating alerts without requiring a human to be watching at the right moment. And the platform connecting the cameras — whether cloud or on-premise — gives operators a way to investigate incidents in minutes rather than days.
This guide covers the complete camera lineup, the five innovations that set Avigilon apart from standard IP cameras, how four different industries use them in practice, and what a properly executed installation actually involves.
The Complete Camera Lineup — Built for Every Environment
Avigilon designs each camera series around a specific class of physical environment and operational need. Selecting the right product for a location isn’t just about megapixels — it’s about coverage geometry, mounting constraints, AI capability requirements, and the specific compliance obligations your sector carries. Select each series below to understand what it does and where it belongs.
The highest-performance series — with two distinct compliance paths
The H6A and H6X represent the top of Avigilon’s commercial camera range. Both deliver resolutions from 2MP through to 8MP (4K Ultra HD) in dome and bullet form factors, with Wide Dynamic Range processing and full AI object analytics. Where they diverge is in the secondary hardware they include — and that difference shapes which one belongs in your building.
H6A — AI Audio Analytics Model
The H6A includes built-in directional microphones and onboard audio processing. Beyond capturing sound, it classifies it — distinguishing breaking glass from ambient office noise, distinguishing raised voices from a loud phone call, identifying car alarms in a car park. When a classified audio event is detected, the camera generates an alert that names what it heard, not just that it heard something. This second sensory channel means the system can respond to incidents that haven’t yet become visually obvious — a smashed window triggers an alert before anyone appears on camera.
H6X — Privacy and Compliance Model
The H6X is designed without microphones or wireless radio chips. This hardware design choice is significant: it means the device is architecturally incapable of audio recording or wireless data transmission, satisfying compliance mandates where audio recording is restricted and wireless transmissions require sign-off. Government agencies, defence facilities, and high-security research environments frequently mandate equipment at this specification level.
Best suited to: Corporate headquarters, secure facilities, environments with specific audio threat profiles (schools, warehouses, car parks), and any regulated environment with data residency requirements.
One camera, four lenses — coverage geometry that would otherwise require a network of separate units
Open spaces create a specific surveillance design challenge. A large car park, a building corner, an outdoor atrium, an intersection inside a logistics facility — these areas have no natural chokepoints. Covering them with standard cameras requires multiple units, multiple runs of cabling, multiple network switch ports, and multiple software licences.
The H5A Multisensor camera contains up to four independently adjustable lens modules inside a single housing, connected to the network via one cable. The lenses can be configured to produce seamless 180°, 270°, or full 360° coverage from a single ceiling mount point. The AI analytics run across the combined field of view — a person moving from the coverage zone of lens one into the coverage zone of lens three is tracked continuously, not handed between separate camera feeds.
The cost saving comes not from the hardware unit price but from the installation: one cable run instead of four, one switch port instead of four, one software licence instead of four. For large sites with open coverage requirements, this arithmetic compounds quickly.
Best suited to: Parking structures, building corners, large open floorplates, logistics yards, intersections inside industrial facilities.
Built for long corridors and tight spaces where two cameras would otherwise face each other from opposite walls
Corridors, stairwells, and entry gate lanes present a specific mounting challenge. The space is long and narrow. Effective coverage requires looking in both directions. The standard approach — two cameras on opposite walls pointing at each other — doubles the hardware, doubles the cabling, and creates a natural blind spot in the middle of the space where neither camera has a useful angle.
The H5A Dual Head puts two independent, varifocal lens modules inside a single low-profile housing. Mounted in the centre of the corridor, one lens points left and one points right. Full bilateral coverage from a single mounting point.
The varifocal lenses are adjustable at installation, allowing the field of view for each head to be tuned to the specific length and width of the corridor without swapping lens units. A long hotel corridor and a short stairwell landing take the same hardware, configured differently.
Best suited to: Office building corridors, stairwells, hotel hallways, narrow entry passages, underground car park aisles, entry gate lanes.
Long-range optical zoom and AI-assisted subject tracking — for perimeters where static cameras can’t reach the detail you need
Pan-Tilt-Zoom cameras serve a different function from fixed cameras. They provide operator control and long-range reach that fixed-lens cameras cannot replicate — the ability to zoom in on a specific person or vehicle hundreds of metres away, track movement across a wide outdoor area, and deliver identification-quality imagery at ranges where a fixed camera would show a moving pixel.
Avigilon PTZ cameras feature up to 36x optical zoom combined with long-range infrared illuminators. At maximum zoom, they can capture licence plate characters and facial details in conditions of near-total darkness at distances that far exceed what any fixed camera can cover.
The AI layer adds autonomous tracking capability. When unusual motion is detected at the perimeter, the PTZ can lock onto the subject and smoothly follow them across the property boundary without an operator manually controlling the pan and tilt. This means one PTZ can maintain a continuous view of a single intruder across an area that would otherwise require coordinating between multiple fixed cameras.
Best suited to: Port terminals, large industrial sites, stadium perimeters, open outdoor areas where the distance to the subject makes fixed cameras impractical.
Engineered specifically for environments where the camera hardware itself is a physical vulnerability
Standard camera housings project from the wall surface. In a standard commercial environment, this is acceptable. In a psychiatric hospital ward, a detention facility, a remand centre, or a secure holding area, any protruding object with a gap between it and the wall surface is a ligature risk — a point where material can be attached and misused.
The H5A Corner Camera uses a flush-mount anti-ligature design. The housing sits completely flat against the wall surface at the corner of a room, with no gap between the unit and the wall on any side. The cast-aluminium exterior is designed to withstand deliberate impact — it doesn’t shatter or fracture under the kinds of forces common in environments where patients or detainees may attempt to destroy equipment. Despite this hardened design, the wide-angle lens delivers full room coverage from the corner position.
Best suited to: Psychiatric inpatient facilities, remand centres and prisons, holding cells, forensic mental health units — any environment where both room coverage and physical safety around the hardware itself are requirements.
Serverless, self-managing, and designed to scale across multiple locations without a central IT infrastructure
Every other camera series in this list feeds footage into a server somewhere — whether an on-premise NVR or a cloud platform. Alta cameras are different in architecture: they process AI analytics at the edge and stream directly to Avigilon’s cloud infrastructure without any intermediate server hardware on-site.
For a business with a single large headquarters, this architecture is an option. For a business with ten, twenty, or fifty locations — a retail chain, a restaurant group, a franchise operation — it changes the economics of surveillance entirely. Adding a new site means plugging cameras in and registering them in the cloud dashboard. No on-site server procurement, no IT infrastructure build-out at each location, no inter-site VPN configuration. The camera handles its own AI processing, stores footage locally if the internet connection drops, and syncs to the cloud when connectivity returns.
Management across all sites happens through a single web dashboard or mobile app, regardless of how many locations are in the estate.
Best suited to: Multi-site retail, hospitality chains, distributed logistics networks, businesses that want to add locations without an IT project at each one.
Five Innovations That Separate Avigilon From Standard IP Cameras
Most commercial-grade IP cameras produce reasonable video. The gap between “reasonable video” and “security intelligence” is where Avigilon invests. These five capabilities are what that investment looks like in practice.
🔊 Integrated Audio Analytics
A camera that only sees is limited to responding after something becomes visually obvious. The H6A cameras process sound using onboard machine learning — continuously listening and classifying audio events against known threat signatures. A smashed window generates a “glass break” alert before anyone appears on camera to investigate. A rapid escalation in voice volume in a corridor triggers a “screaming” classification. A car alarm in the car park flags immediately.
This second sensory layer means the system can flag incidents at the sound stage rather than the visual stage — seconds or minutes earlier than a purely visual system. In an environment where physical violence escalates quickly, that lead time is operationally significant.
🔐 FIPS 140-2 Level 3 Hardware Security
A camera mounted on an external wall is physically accessible to anyone who can reach it. Disconnecting it from the network cable and connecting an attack device is a standard method for attempting to penetrate a corporate network through an overlooked physical entry point.
Avigilon counters this at the hardware level. A FIPS 140-2 Level 3 compliant Trusted Platform Module chip sits inside the camera itself. When the camera starts up, Secure Boot technology verifies the integrity of the firmware before allowing any processes to run. If the firmware has been tampered with, or if an unauthorised device attempts to join the network by impersonating a camera, the unit locks itself down. The corporate network behind it stays clean.
🎭 Dynamic Privacy Masking
Businesses operating public-facing cameras face a tension between security coverage and data privacy obligations. Storing identifiable footage of members of the public who aren’t involved in any incident is a data protection liability in many jurisdictions.
Dynamic Privacy Masking automates the resolution. The AI software tracks faces and vehicle plates in the live feed and applies a blur in real time — the recording contains no identifiable personal data for routine viewing. When an investigation requires the raw footage, a security-authorised administrator enters credentials to lift the mask. The full-quality original is available for law enforcement purposes without the organisation having stored hours of identifiable public footage as a matter of course.
☀️ Extreme WDR Low-Light Clarity
The worst lighting conditions for a standard camera aren’t pitch darkness — they’re the combination of darkness and a bright light source in the same frame. A car driving into an underground car park at night faces directly toward the camera, flooding the lens with headlight glare and making the driver and registration plate completely invisible behind it.
Avigilon cameras combine high-sensitivity image sensors with Wide Dynamic Range processing up to 150 dB, merging multiple simultaneous exposures to render both the dark areas and the bright light source in the same frame without either washing out or going black. Under these conditions, the registration plate next to the headlight beam reads clearly. The driver’s face is visible, not a white overexposed silhouette.
📦 HDSM SmartCodec
4K video from a network of 50 cameras demands substantial storage and bandwidth. Standard compression algorithms treat every frame equally — compressing the empty car park at 3am at the same quality level as the busy lobby at 9am, consuming the same resources for footage with no security value.
HDSM SmartCodec analyses the content of each frame. Static elements — the floor, the walls, a parked vehicle that hasn’t moved in three hours — are compressed heavily. Moving subjects — a person, a vehicle entering the frame, an object being carried — are preserved at full resolution. The result is storage and bandwidth consumption up to 50% lower than standard codecs, with no degradation in the image quality of the subjects that matter.
Standard Codec vs. Avigilon HDSM SmartCodec
The difference in how these two compression approaches handle a real security scene becomes significant at scale. Here’s the comparison across the metrics that matter for a multi-camera commercial deployment.
| Standard H.264 / H.265 Compression | Avigilon HDSM SmartCodec |
|---|---|
| Compresses every element of every frame at the same rate regardless of content | Analyses each frame — distinguishes static background from moving foreground subjects |
| Consumes bandwidth continuously at a constant high rate | Compresses background heavily, preserves full resolution on moving objects |
| Storage drives fill at a predictable but high rate across quiet periods | Storage consumption drops by up to 50% without reducing detail where it matters |
| 4K streaming can create visible network lag for other applications on the same switch | Efficient bandwidth profile allows 4K camera networks to coexist with office traffic |
Four Industries, Four Different Security Problems — One Platform
Avigilon cameras appear across sectors with genuinely different security objectives. The platform’s range means the right product exists for each context — what changes is the specific model, placement logic, and AI configuration. Expand each sector below.
The problem: High-rise commercial buildings in business districts handle thousands of visitor and tenant movements daily across shared lobbies, lift banks, parking structures, and tenant-exclusive floors. The physical security challenge isn’t finding one specific bad actor — it’s managing the boundary between legitimate access and unauthorised presence across a building with dozens of points of entry.
The Avigilon deployment: H6A dome cameras cover lobby entrances, turnstiles, and lift banks — feeding AI object classification to the VMS in real time. When an employee badge-taps a turnstile, the camera verifies that the person crossing matches the credential using AI face matching against the ID database. When two people walk through on a single tap — a common behaviour that standard access control hardware can’t detect — the camera flags the tailgate event and alerts the security desk with a video clip rather than an audio-only alarm.
In the parking structure below, dedicated LPR cameras read plates as vehicles approach the barrier. Tenants are on an approved list — the barrier lifts as they pull up. Visitors not on the list reach an intercom. Vehicles on a watchlist trigger an alert to security rather than an automatic entry permit.
The problem: A logistics facility covering twenty acres of outdoor shipping yard presents a surveillance scale that can’t be solved with standard cameras at standard heights. Containers, vehicles, and workers move continuously 24 hours a day. The threat profile includes both external perimeter intrusion and internal stock discrepancy — someone walking cargo to the wrong truck, or someone who shouldn’t be in the yard appearing in a restricted area.
The Avigilon deployment: H5A Multisensor cameras mount on tall poles across open yard areas, delivering wide-area coverage with a fraction of the unit count of individual cameras. H6A bullet cameras aim down perimeter fence lines for long-range intrusion detection.
The self-learning AI models the normal activity at each camera position over its first week of operation. It learns that vehicles move through the main gate continuously during day shifts, that workers cross specific paths at predictable times, and that forklift activity clusters around specific bays. When a person moves along the perimeter fence line at 2am, the motion pattern is immediately anomalous — the system raises the alert, activates external strobes, and routes a push notification with a live video clip to the duty manager’s phone.
The problem: Educational environments present a unique tension. They need to remain open and welcoming — fenced perimeters and manned checkpoints create exactly the wrong atmosphere. But the safety of students and staff requires genuine real-time awareness of what’s happening across a large campus with many buildings, corridors, and outdoor spaces.
The Avigilon deployment: H5A Dual Head cameras in corridors and stairwells cover long linear spaces efficiently from central mounting points. H6A cameras with audio analytics cover cafeterias, gymnasiums, and common outdoor areas where physical confrontations are most likely to begin as noise before they become visible.
The occupancy counting capability from the AI layer gives school administrators a live view of crowd density — if a group of students gathers rapidly in an unusual location, the density change itself is flagged before any incident is visible. Audio alerts for screaming, glass break, or unusual noise volume route to the on-site resource officer’s radio, pinpointed to the exact camera location on a digital building map.
The problem: Healthcare environments carry two simultaneous and sometimes contradictory requirements: absolute visibility of clinical and medication areas for security and safety purposes, and strict data protection obligations for patient privacy. The same camera network needs to serve both objectives without compromising either.
The Avigilon deployment: H6X cameras cover pharmacy storage areas, controlled substance dispensing points, and emergency triage zones. The H6X’s hardware architecture — no microphones, no wireless — satisfies the compliance requirements that typically apply to these areas.
Privacy masking runs automatically on all patient-area feeds, blurring identifiable facial data in the live view and in recordings used for routine monitoring. When an audit reveals an access discrepancy, the masked footage can be unblurred by an authorised security administrator for official investigation purposes — the original, unmasked version is preserved in the encrypted audit store.
Corner cameras in patient rooms — the anti-ligature H5A model — provide fall detection alerts. If a patient attempts to stand unassisted, the positional change triggers an alert to the nursing station before any fall occurs.
Unity On-Premise vs. Alta Cloud — And the Hybrid Middle Ground
The platform feeding your cameras shapes how your security team operates on a daily basis. Neither architecture is universally right — the correct choice depends on data residency requirements, site geography, IT resource availability, and how your organisation plans to grow.
| Dimension | Avigilon Unity — On-Premise | Avigilon Alta — Cloud-Native |
|---|---|---|
| Data Location | On your servers, inside your building — never transmitted externally | Avigilon’s cloud infrastructure — accessible from anywhere |
| Server Hardware | Required — physical NVRs in your IT room | None — cameras connect directly to cloud |
| Internet Dependency | Continues recording fully if internet fails | Stores locally during outage, syncs to cloud on return |
| Scale Ceiling | Up to 10,000 cameras per site | Unlimited — add cameras and sites without infrastructure changes |
| Updates | Manual — managed by IT team or partner | Automatic overnight — always current |
| Best For | Government, defence, banking, hospitals — data localisation mandates | Multi-site retail, hospitality, logistics, growing commercial businesses |
🔄 The Hybrid Deployment — Best of Both
Many enterprise organisations don’t fit cleanly into either category. A manufacturing company with a primary production facility and twenty regional sales offices is a common example: the main site has the network infrastructure, the IT staff, and the data volume to justify on-premise Unity servers. The regional offices have minimal IT footprint and benefit from a cloud model that requires no local server at each location.
Avigilon supports this mixed deployment directly. The main site runs Unity on-premise, handling heavy local video traffic with no cloud dependency. The regional offices run Alta cloud cameras, managed through the same web dashboard. Security operations has a unified view across all locations — different underlying architectures, single management interface.
Five Tips That Determine Whether the Camera Does Its Job
The hospital scenario that opened this page wasn’t a hardware failure. It was a configuration and placement failure — and it’s far more common than camera failures. These five practices are what separate a camera network that performs reliably from one that produces recordings nobody can use when something matters. Expand each tip below.
Security camera tasks divide into three levels — Detection, Recognition, and Identification (DRI) — and each requires a different relationship between camera resolution, lens focal length, and the distance to the subject.
Detection means confirming that a human figure is present in a scene. A wide-angle camera covering a large car park can detect bodies at considerable distance. Recognition means determining whether you’ve seen this person before — enough clarity to match against a known face. Identification means capturing facial detail sufficient for a court to use — the quality that a prosecution depends on.
If you need identification-quality footage at the entry gate 15 metres from the camera, a wide-angle lens won’t deliver it. A telephoto lens on the same camera changes the outcome. Match the lens specification to the DRI task at each location, not just to the area you want in frame.
Mounting a camera too high creates the problem the pharmacy camera in this page’s opening scenario had — the angle becomes so steep that the camera captures hairlines and shoulders rather than faces. The person walks under the camera and the footage shows the top of a hat.
For indoor facial capture, mount dome cameras at 2.5 to 3 metres and angle them slightly forward rather than straight down. For outdoor perimeter AI analytics, mount high enough to prevent physical tampering, but angle the lens outward far enough to capture approaching subjects from a useful angle — closer to profile than to crown. The correct mounting configuration for each specific camera model and location should be confirmed by an engineer who has measured the actual scene dimensions.
Infrared illuminators project invisible light to extend a camera’s visibility in darkness — but they interact poorly with nearby objects. A tree branch intruding into the front of the camera’s field of view will reflect IR light back into the lens, creating a washed-out bright patch that blocks visibility. The same effect occurs with roof gutters, wall edges, and any surface that sits within about 30cm of the front of the camera lens.
Before a camera is commissioned, clear its immediate field of view of any overhanging foliage, nearby surfaces, or reflective objects. For underground car parks and dark alleys where the AI needs clean ambient light to classify objects reliably, install supplementary external LED illuminators rather than relying entirely on the camera’s built-in IR.
Security cameras are network devices. If they share a subnet with employee computers, a compromised laptop becomes a potential entry point to the camera network — and from there, to whatever the camera network connects to. The reverse is also true: a compromised camera (a risk that Avigilon’s FIPS hardware security addresses, but which exists as a theoretical path on generic brands) could reach devices on the same subnet.
The correct configuration is a dedicated VLAN for all surveillance hardware, isolated from both the general office network and guest Wi-Fi. Strict firewall rules control what traffic can pass between the camera VLAN and everything else. This is a standard network hygiene measure that costs nothing beyond the configuration time.
Outdoor cameras in a Gulf climate face conditions that degrade performance faster than in temperate environments — high ambient dust, UV exposure that breaks down cable sheathing, temperature cycles that work loose connections over time. A camera that’s technically functioning but has a dust-fogged lens produces footage the AI can’t classify reliably.
Schedule quarterly physical maintenance visits: clean lens faces with anti-static microfiber cloths and appropriate lens solution, inspect cable gland seals for moisture ingress, check bracket fasteners for vibration loosening, review error logs for packet drop events or hard drive health warnings. The problems maintenance visits catch are all cheap to fix. The problems they don’t catch tend to become expensive to discover.
Six Stages of Professional Installation — What Hutaib Infotech Delivers
The right camera in the wrong position is the pharmacy story over again. Professional installation isn’t box-mounting — it’s the process of translating a security requirement into a camera network that actually meets it. Hutaib Infotech Solutions (www.securitysolutionsdubai.com) executes every Avigilon deployment across six structured stages.
01
Site Survey & Risk Mapping
Engineers walk the physical site before any specification is drafted. They map current blind spots, test ambient lighting at night-time as well as daytime, measure network bandwidth, and document the specific risk vectors at each location — the threat is different at a loading bay versus a reception lobby.
02
Custom Network Architecture
A dedicated video network backbone is designed: PoE switch selection, VLAN configuration, bandwidth calculation across the total camera count and resolution mix, and firewall rule design. The network that feeds an 8-camera installation differs from the one feeding 80 cameras in 4K.
03
Precision Hardware Mounting
Cameras are mounted at engineered heights and angles, verified against the DRI requirements for each position. Industrial-grade brackets and weather-sealed cable connections — not cable ties and silicone. Every installation point documented photographically before the ceiling is closed.
04
AI Calibration & Tuning
Engineers configure every detection zone, perimeter tripwire, and audio classification threshold manually — not from factory defaults. They test each configuration in the actual physical environment and adjust based on observed false positive rates before signing off. LPR watchlists and access control mappings are built and verified end-to-end.
05
System Integration
Camera feeds are connected to access control, fire alarm panels, and gate barriers where specified. Unified management is tested — does a door access event correctly pull up the camera footage at that door? Do camera alerts route to the right person’s device? Integration testing covers the full workflow, not just the components in isolation.
06
Staff Training & Handover
Security operators are trained in live working sessions — Appearance Search, Focus of Attention, alert management, and footage export for investigations. IT administrators are trained on user management and system maintenance. Handover happens when every user group is confident, not when the hardware is powered on.
Frequently Asked Questions (FAQs)
Standard motion detection responds to any change in pixel values within its configured zone — which includes leaves blowing in wind, shadows moving across a wall, raindrops, and insects on the lens, as well as actual security events. Avigilon’s self-learning AI classifies objects by their shape, movement vector, and speed before deciding whether to raise an alert. It distinguishes a human walking from a bush moving in wind. You can configure it to alert only on specific confirmed object types — humans, vehicles, or specific combinations. False alarm rates drop significantly, which means the alerts that do fire get taken seriously rather than being habitual noise.
Yes on both platforms, though the connection method differs. Alta cloud cameras are directly accessible through the mobile app from any internet connection globally — the footage is already in the cloud. Unity on-premise footage requires a VPN connection to your internal network, after which the same mobile app accesses the feeds as if you were on-site. Remote access for both platforms includes live streaming, historical footage review, PTZ camera control, and push alert notifications.
Yes. Built-in adaptive IR illuminators adjust their output power based on subject distance — close subjects don’t get washed out with too much light, distant subjects aren’t lost in darkness. For extreme long-range applications, Avigilon’s PTZ cameras with extended IR throw and thermal imaging options provide detection and identification capability hundreds of metres away in total darkness. The thermal models detect heat signatures where even IR illumination can’t reach.
Alta cameras contain onboard solid-state storage that activates the moment cloud connectivity is lost. Recording continues uninterrupted, with footage written to local storage. When connectivity returns, the camera synchronises the locally stored footage back to the cloud automatically. From the operator’s perspective, there’s no gap — the timeline in the management interface is continuous regardless of when the connection failure occurred.
Usually yes. Avigilon platforms conform to ONVIF Profile S, T, M, and G standards, which means the VMS software will communicate with most ONVIF-compliant third-party cameras. Additionally, connecting existing cameras through an Avigilon AI NVR appliance unlocks AI analytics, Appearance Search, and smart compression on those older feeds — meaning the existing cameras gain significant capability without physical replacement. Your installation partner tests compatibility during the site survey.
Resolution directly determines how far from the camera you can reliably identify faces and read text. A 2MP (1080p) camera works well for identification at close range — a doorway or a small indoor space. A 5MP camera extends that identification range and provides more digital zoom headroom in recorded footage. An 8MP (4K) camera provides four times the pixel density of 1080p — the ability to crop a small section of a wide-angle frame in post-investigation and still read a face or a licence plate clearly. For large spaces where the subject distance from the camera is significant, the higher resolution is what determines whether the footage is usable as evidence.
Retention time is a function of storage capacity, camera count, resolution, frame rate, and compression efficiency. With HDSM SmartCodec reducing storage consumption by up to 50% compared to standard codecs, a given drive capacity stores roughly twice as much footage at equivalent quality. Avigilon’s management software allows differentiated retention policies by camera — the front entrance might retain 90 days, a back corridor 30 days — with automatic deletion of older footage when capacity is needed. Many organisations are also subject to minimum retention requirements from industry regulators; your installer configures the policies to meet those obligations.
Security is implemented across multiple layers. At the hardware level, a FIPS 140-2 Level 3 Trusted Platform Module chip verifies device integrity at startup — the camera won’t boot if its firmware has been tampered with. Communications use HTTPS encryption. Firmware updates are cryptographically signed before they’re applied, preventing installation of unsigned code. User authentication requires strong credentials and supports role-based access control. The combination means the cameras themselves are a very poor attack vector compared to many other network devices.
Appearance Search is a deep-learning AI engine built into the Avigilon VMS platform. It analyses recorded footage across all cameras to find instances of a specific person or vehicle based on visual characteristics. An operator can click on a person in a video clip and ask “find everyone who looks like this across all cameras in the last 24 hours” — the system builds a timestamped movement trail. Alternatively, physical descriptors — “orange jacket,” “dark SUV,” “man with large backpack” — can be entered as text search terms. Searches that would take hours of manual scrubbing return results in seconds.
The most useful starting point is a professional site survey, not a product selection. The survey maps your current camera positions, identifies blind spots, measures lighting conditions at the times of day that matter most for your security requirement, assesses network infrastructure, and produces a specification that matches camera models to specific locations based on what each position needs to achieve. Without this groundwork, camera selection and placement is educated guesswork. Contact Hutaib Infotech Solutions (www.securitysolutionsdubai.com) to arrange your site assessment.
Don't Discover What Your Cameras Were Missing After an Incident
The pharmacy camera problem is still sitting in buildings all over the region right now. Cameras mounted too high, angled too steeply, aimed at the wall instead of the subject, running in conditions where the image the AI receives is noise rather than video. They’re recording. Nobody would know they’re not working until the day someone needs the footage and discovers what it actually contains.
Avigilon hardware is the right starting point — the imaging quality, the AI capability, the cybersecurity protection, the HDSM compression efficiency. But hardware correctly specified for a location, mounted at the right angle, calibrated with detection zones that match the actual physical environment, and integrated with the access control and alarm systems it needs to talk to — that’s what the pharmacy needed, and what it didn’t have.
That’s the gap that professional deployment closes. Hutaib Infotech Solutions (www.securitysolutionsdubai.com) are certified Avigilon integration specialists. They survey your site, design the specification, mount the hardware correctly, tune the AI, integrate the systems, and train your team. Your cameras end up covering what you think they’re covering — and working when you actually need them to.