Enterprise Video Management and AI Video Analytics

Section 28 20 00
Enterprise Video Management & AI Video Analytics

Basis Of Design: BluB0X BluSKY
Related Sections:

• 28 13 00 – Enterprise Physical Security Management Platform (PSMP)
• 28 10 00 – Enterprise Access Control System (EACS)
• 28 05 00 – Common Work Results For Electronic Safety And Security

Part 1 – General

1.1 Summary
A. This section defines the requirements for an Enterprise Video Management System (EVMS) combined with AI-based Video Analytics and Intelligence.
B. The system shall provide live monitoring, recording, streaming, snapshot capture, analytics, event generation, and cross-system correlation under centralized PSMP management.
C. Video shall function as an autonomous sensing and intelligence system, not solely as a recording subsystem.

1.2 Purpose And Intent
A. The intent of this specification is to:

1. Establish video as a data-producing and event-generating system

2. Enable AI-driven analytics, reasoning, and autonomous optimization

3. Support secure, low-latency video access on any device, anywhere

4. Ensure long-term adaptability, privacy compliance, and lifecycle flexibility

B. The system shall support execution of analytics at the edge, on dedicated hardware, or in the cloud, based on operational requirements.

1.3 Definitions
A. EVMS – Enterprise Video Management System
B. Snapshot – A still image captured independently or in sequence from video streams
C. ROI / RODI – Region Of Interest / Region Of Disinterest
D. Homography – Spatial mapping between camera views and floor plans

Part 2 – System Description

2.1 Video System Overview
A. The EVMS shall manage:
• Live video streams
• Recorded video
• Local and cloud-mediated streaming
• Snapshots and snapshot sequences
• Video metadata and AI-derived descriptors
• Analytics outputs and correlated events

B. Supported video sources shall include:
• Fixed cameras
• PTZ cameras
• Multi-sensor / panoramic cameras
• Elevator and specialty cameras

2.2 Relationship To PSMP
A. The PSMP shall serve as the system of record for:
• Video-generated events
• Analytics results
• Cross-system correlations

B. Video platforms may be native or integrated, provided all requirements herein are met.

2.3 System Scope
The system shall support:
• Live monitoring and playback
• Secure local and cloud streaming
• Event-based and continuous recording
• Snapshot-driven analytics
• Autonomous video intelligence
• Correlation with access control, alarms, intercoms, identity, and elevators

Part 3 – Video Management Architecture

3.1 Deployment Models
A. The EVMS shall support:

1. On-premises deployments

2. Cloud-hosted deployments

3. Hybrid architectures combining local recording with centralized intelligence and streaming

3.2 Edge, Cloud, And Hybrid Analytics
A. Analytics shall be executable:
• On edge hardware
• On dedicated AI devices
• In centralized or cloud environments

B. Analytics placement shall be configurable by camera, site, or use case.

3.3 Resilience And Scalability
A. Local recording and streaming shall continue during WAN outages.
B. The architecture shall scale from single cameras to enterprise portfolios without redesign.

Part 4 – VMS Platforms And Recording Systems

4.1 Third-Party VMS Integration
A. The system shall support integration with enterprise VMS platforms including, but not limited to:
• Milestone
• Salient Systems
• Avigilon
• Exacq

B. Integration shall not restrict access to video, metadata, snapshots, analytics outputs, or events.

4.2 Native And Edge Video Hardware
A. The system shall support proprietary video recording and AI hardware capable of:
• Video recording
• Streaming
• Snapshot capture
• Edge analytics
• Event generation

B. Native hardware may operate autonomously or in coordination with centralized services.

Part 5 – Camera And Video Source Support

5.1 Camera Types
The EVMS shall support:
• Fixed cameras
• PTZ cameras
• Multi-sensor / panoramic cameras
• Elevator and specialty cameras

5.2 Video Formats And Streams
A. Multiple concurrent streams per camera shall be supported.
B. Supported codecs shall include:
• H.264
• H.265

C. Resolution, frame rate, and bitrate shall be configurable.

Part 6 – Recording, Storage, And Retention

6.1 Recording Modes
The system shall support:
• Continuous recording
• Event-based recording
• Snapshot-only recording
• Scheduled recording

6.2 Snapshot And Event Capture
A. Any system event may trigger snapshot capture, including:
• Motion analytics
• Access control events
• Door forced or held
• Alarms
• Intercom events

B. Snapshot frequency, duration, and sequencing shall be configurable.

6.3 Storage And Retention Governance
A. Retention policies shall be configurable by:
• Camera
• Event type
• Site
• Facility
• Tenant
• Geography

B. Events and associated video may be locked to prevent overwrite.

Part 7 – Natural Language Camera Intent (BluEYES)

7.1 Natural Language Configuration
A. The system shall support defining camera behavior using natural language descriptions.
B. Instructions may describe:
• Objects of interest
• Behaviors of interest
• Time-based conditions
• Required responses

7.2 Autonomous Scene Interpretation
Based on defined intent, the system shall autonomously:
• Identify background reference imagery
• Identify floor or ground regions
• Create ROIs and RODIs
• Determine relevant object classes

7.3 Automated Responses
The system shall support automated actions including:
• Log entries
• Notifications
• Alerts
• Alarms
• Escalation workflows

Part 8 – AI Video Analytics And Intelligence

8.1 Core Analytics
A. The system shall provide:
• Object detection and classification
• Background modeling
• Motion detection with false-alarm reduction
• Dwell and loitering detection

8.2 Temporal And Statistical Intelligence
A. Baseline event frequency shall be calculated by:
• Time of day
• Day of week
• Object type

B. Anomalies shall be detected relative to baseline behavior.

8.3 Movement And Path Analysis
A. The system shall identify:
• Entry points
• Exit points
• Travel paths
• Dwell zones

Part 9 – Homography And Spatial Context

9.1 Camera-To-Floor-Plan Mapping
A. Support homographic mapping between camera views and floor plans.

9.2 Spatial Projection
A. Project camera ROIs onto floor plans and floor plan regions into camera views.

9.3 Natural Language Spatial Descriptions
A. Events shall be describable using named regions meaningful to users.

Part 10 – Event Generation, Snapshots, And Correlation

10.1 Event Triggers
Events may originate from:
• Video analytics
• Access control
• Alarms
• Intercoms

10.2 Autonomous Association
A. Events shall be autonomously associated with relevant cameras and devices.
B. Associations shall be bi-directional.

10.3 Evidence Packaging
A. Events shall include associated snapshots, video segments, and metadata.

Part 11 – Autonomous Camera Optimization

11.1 AI-Driven Optimization
The system shall autonomously optimize:
• Resolution
• Frame rate
• Bitrate
• Recording schedules
• Pre- and post-event capture durations

Optimization shall be based on observed scene behavior.

Part 12 – Streaming And Video Delivery

12.1 Local Streaming
A. Support low-latency video streaming within local networks.

12.2 Cloud Streaming
A. Support secure cloud-mediated streaming to remote devices.
B. No inbound firewall ports shall be required.
C. Outbound-only communication shall be supported.

12.3 Streaming Technologies
A. Support WebRTC and peer-to-peer streaming.
B. Streaming shall function through NAT and firewalls.

12.4 Device And Interface Independence
A. Video shall be viewable:
• In standard web browsers
• On mobile devices
• On desktops and laptops

B. Interfaces shall be responsive and adapt to device capabilities.

Part 13 – Search, Data Lake, And Sharing

13.1 Natural Language Search
A. Support natural language search across video, snapshots, events, and metadata.

13.2 Data Lake Integration
A. Video metadata and analytics outputs shall be available beyond local storage.

13.3 Secure Sharing
A. Support cloud-hosted video with secure, shareable links.
B. No file downloads shall be required to share video.

Part 14 – Over-The-Air Updates And Lifecycle

14.1 Autonomous OTA Updates
A. Support cloud-delivered autonomous updates.
B. Updates shall:
• Be verified prior to activation
• Be transactional
• Support rollback
• Minimize operational disruption

14.2 Continuous Improvement
A. The system shall support ongoing enhancement without manual intervention.

Part 15 – Privacy, Compliance, And Governance

15.1 Privacy Controls
A. Role-based access to live and recorded video.
B. Support masking and redaction.

15.2 Regulatory Compliance
A. Support compliance with:
• GDPR
• U.S. federal and state privacy laws (including California)
• Regional and national privacy regulations

15.3 Geography-Aware Policy Enforcement
A. Retention, analytics, and storage policies shall be configurable by:
• Country
• Region
• Site
• Facility
• Tenant

Part 16 – Operator And Administrative Interfaces

A. Provide dashboards combining:
• Video
• Events
• Analytics
• Maps and floor plans

B. Display camera status, metadata, location, and associated devices.

Part 17 – Submittals And Close-Out

A. Camera schedules
B. Analytics and retention configurations
C. Compliance and privacy policies
D. Spatial mappings and homography documentation

Part 18 – Acceptable Manufacturers

18.1 Basis Of Design
A. Managed through the PSMP; supports integrated third-party VMS platforms and proprietary video + AI hardware.

18.2 Acceptable Alternatives
A. Alternative systems shall meet all requirements of this specification.

End Of Section 28 20 00