ResEngine User Guide

Welcome to the ResEngine User Guide. This comprehensive documentation explains the various actions available in the system, how they behave, and how to configure them effectively.

What is ResEngine?

ResEngine is the intelligent engine that powers your video security and analytics workflows. Think of it as the "brain" that connects your cameras to meaningful actions—detecting events, counting visitors, creating alerts, and sending notifications.

Instead of requiring you to write complex rules or hire developers for every change, ResEngine allows you to configure sophisticated monitoring solutions through a simple interface.

What are Actions?

Actions are the individual building blocks that make up a workflow. Each action does one specific thing, and by combining them, we create intelligent security and analytics solutions.

Examples of Actions:

• Detect People: Uses AI to find people in camera footage
• Count Visitors: Counts how many people enter and exit a location
• Create Violation: Records an incident when something goes wrong
• Send Notification: Alerts the right people when an event occurs

When you create a workflow, you're essentially saying: "First, do this action. Then, based on the result, do that action."

How Workflows Work

A workflow is a series of connected actions. Here's a simple example:

1. Step 1: Run the "Detect People" action on the warehouse camera
2. Decision: Was anyone detected?
   • Yes → Go to Step 3
   • No → End workflow (nothing to do)
3. Step 3: Run "Create Violation" to record the unauthorized access
4. Step 4: The system automatically notifies the security team (based on your Brand settings)

This workflow runs automatically on a schedule (for example, every hour during nighttime hours), and you don't need to monitor it manually.

Action Categories

Actions are organized into categories based on their purpose:

Detection Actions (Image Analysis)

These actions analyze video footage to find and identify things:

• Object Detection: Find people, vehicles, hard hats, or any trained object
• AprilTag Detection: Identify special visual markers (like ID badges)
• Face Detection: Detect human faces
• Mask Detection: Check if someone is wearing required safety gear
• Color Detection: Identify clothing colors (useful for uniform compliance)
• Visitor Counting: Count entries and exits with demographic analysis

Data Processing Actions

These actions read, aggregate, and upload data:

• Zone Dwell Reader: Find events where someone stayed too long in an area
• Visitor Summary Uploader: Send hourly visitor statistics to the cloud
• ANPR: Read license plate numbers from vehicle images

Business Logic Actions

These actions handle alerts and notifications:

• Create Violation: Record an incident with evidence
• Zone Dwell Violation Snapshot: Prepare annotated images for violations

System Actions

These actions perform utility functions:

• Snapshot Stream: Capture a single image from a camera
• Stop: Gracefully end a workflow branch

Key Concepts

Branches

A Branch is a physical location—your store, warehouse, office, or any site with cameras. Each branch has its own cameras, zones, and settings.

Zones

Zones are areas you draw on the camera view to tell the system where to look. For example:

• A "Detection Zone" around a restricted doorway
• An "Entry Line" across the store entrance for counting
• A "Crop Zone" around a cashier's screen for compliance checks

Confidence Threshold

Many detection actions have a Confidence Threshold setting. This is a number between 0 and 1 that tells the AI how "sure" it needs to be before reporting a detection.

• Lower threshold (e.g., 0.2): More sensitive, catches more events, but might have false alarms
• Higher threshold (e.g., 0.7): More strict, fewer false alarms, but might miss some events

Runtime

Max Runtime determines how long an action will look for something before giving up. For example, if you set a detection action to 60 seconds and nothing is found, the action will report "No Detection" and the workflow will continue accordingly.

Understanding Results

Each action produces a Result when it finishes. This result determines what happens next in the workflow.

Common results include:

Result	Meaning	Typical Next Step
detected	Something was found	Proceed to create a violation or alert
no_detection	Nothing was found	End the workflow (nothing to report)
violation_created	An incident record was successfully created	Notifications are triggered automatically
error	Something went wrong	Check the troubleshooting section

Getting Help

If an action isn't behaving as expected:

1. Check the Configuration: Verify that the correct cameras and zones are selected
2. Review the Troubleshooting Section: Each action's documentation includes common issues and solutions
3. Check Camera Connectivity: Many issues are caused by cameras being offline or having network problems
4. Review the Confidence Threshold: If you're missing detections, try lowering the threshold. If you're getting false alarms, try raising it.

Navigation

Use the sidebar to browse the complete Actions Library. Each action has its own page with:

• What it does
• How to configure it
• What results to expect
• Troubleshooting tips