AXIS Object Analytics (AOAS)

AXIS Object Analytics is on-camera AI that pre-classifies objects before they reach Anava. This guide covers AOAS configuration and best practices.

Why Use Object Analytics?

AOAS provides significant benefits when used with Anava:

Benefit	Impact
Fewer false positives	Camera filters non-relevant motion
Lower costs	Only relevant events trigger cloud analysis
Faster response	Less noise to process
Better accuracy	Pre-classified objects provide context

Comparison

AOAS Scenarios

Available Scenarios

Scenario	Description	Use Case
Object in Area	Detect objects entering area	Intrusion, presence
Crossline	Object crosses virtual line	Entry/exit, perimeter
Object Counting	Count objects crossing line	Traffic, occupancy
Time in Area	Object remains for duration	Loitering
Speed	Object moving above threshold	Vehicle speed

Object Classes

Class	Description
Human	People detection
Vehicle	Cars, trucks, etc.
Face	Face detection
License Plate	Vehicle plates

Configuring AOAS on Camera

Step 1: Access Camera Web Interface

1. Navigate to camera IP in browser
2. Log in with admin credentials
3. Go to Apps → AXIS Object Analytics

Step 2: Create Scenario

1. Click + Add Scenario
2. Select scenario type (e.g., "Object in Area")
3. Draw detection area on camera view
4. Configure object filter (Human, Vehicle, etc.)

Step 3: Configure Detection Parameters

Setting	Recommendation
Object Size	Minimum pixel size to detect
Detection Speed	Sensitivity vs accuracy trade-off
Time in Area	For loitering scenarios

Step 4: Test and Verify

1. Enable scenario
2. Walk/drive through detection area
3. Verify events in camera event log
4. Adjust if needed

Using AOAS with Anava

Create Profile with AOAS Trigger

1. In Anava, navigate to group's Profiles
2. Click + Create Profile
3. Set Trigger Type to Object
4. Select Profile matching your AOAS scenario

Trigger Configuration

Trigger:
  Type: Object
  Profile: person  # Matches AOAS scenario name

Available AOAS Profiles

Profile	Description
person	Human detection
vehicle	Vehicle detection
face	Face detection
custom	Named custom scenario

Best Practices

Scene-Appropriate Scenarios

Scene	Recommended Scenario
Entrance	Crossline + Human
Parking	Object in Area + Vehicle
Restricted Area	Object in Area + Human
Perimeter	Crossline + Human + Vehicle

Detection Area Design

Good Areas:
├── Cover entry/exit points
├── Avoid high-traffic paths (unless needed)
├── Include buffer zone for detection time
└── Account for camera angle

Avoid:
├── Areas with constant motion (trees, flags)
├── Reflective surfaces
├── Extreme close-up coverage
└── Areas with lighting changes

Object Size Settings

Distance	Min Object Size
0-10m	10-20% of frame
10-30m	5-10% of frame
30m+	3-5% of frame

Combining AOAS with Anava

AOAS Pre-filter + Anava Intelligence

The optimal setup:

1. AOAS filters to relevant objects only
2. Anava provides deep AI analysis
3. Result: Low false positives + Rich intelligence

Example: Security Setup

AOAS Configuration:

• Scenario: Object in Area
• Object: Human
• Area: Perimeter zone

Anava Skill:

• Pre-filter: (none needed, AOAS handles)
• Analysis: Authorization, threat assessment
• Objects: Person, Weapon (for additional classification)

Example: PPE Compliance

AOAS Configuration:

• Scenario: Crossline
• Object: Human
• Area: Work zone entry

Anava Skill:

• Analysis: PPE detection
• Objects: Hard Hat, Vest, Glasses

Troubleshooting

AOAS Not Triggering

1. Check scenario is enabled

   • Verify in camera Apps interface

2. Check object size

   • Object may be too small
   • Reduce minimum size threshold

3. Check detection area

   • Object must enter defined area
   • Adjust area boundaries

Wrong Objects Detected

1. Check object filter

   • Enable only relevant classes

2. Adjust sensitivity

   • Lower detection speed for accuracy

3. Check lighting

   • Shadows can cause false detection

Events Not Reaching Anava

1. Check profile trigger

   • Trigger type must be "Object"
   • Profile name must match AOAS scenario

2. Check camera connectivity

   • Camera must be online in group
   • MQTT connected

Performance Optimization

Reducing Camera Load

• Limit active scenarios (2-3 per camera)
• Use appropriate detection speed
• Set reasonable object sizes

Reducing Cloud Costs

• Use AOAS for all appropriate cameras
• Combine with Anava scheduling
• Only enable needed object classes

Related Topics

• Trigger Types - All trigger options
• View Areas - Multi-view configuration
• Tuning & Optimization - Reducing false positives