Skip to main content

Inspection Setup & ROI Types

This page explains the fundamental concepts behind defining inspection areas and organizing AI-based inspection workflows in the OV20i system.

ROI (Region of Interest) Fundamentals​

What are ROIs?​

ROIs (Regions of Interest) define rectangular areas where the AI model will perform inspection analysis. These regions focus processing power on specific locations, improving both accuracy and efficiency.

Core ROI Concept:

  • AI Model Workspace - ROIs serve as the reference positions where AI models analyze image data
  • Processing Efficiency - Focused analysis improves inspection speed
  • Result Precision - Targeted analysis reduces false positives
  • Multiple Regions - Multiple ROIs can be defined for comprehensive inspection

image.png

ROI Types in OV20i​

Rectangular ROI Only​

The OV20i uses rectangular ROIs as the single ROI type for all inspection scenarios.

Why Rectangular:

  • Universal Application - Suitable for most industrial inspection needs
  • AI Optimized - Rectangular regions optimize model processing
  • Simple Implementation - Straightforward setup and management
  • Alignment Compatible - Works seamlessly with template alignment system

ROI Naming and Identification​

Default Naming: "New ROI"​

Each new ROI starts with the default name "New ROI" which can and should be changed to provide meaningful identification.

Why Change ROI Names:

  • System Integration - Named ROIs are easier to reference in automation
  • Maintenance - Clear names simplify troubleshooting
  • Documentation - Meaningful names improve system understanding

Naming Examples:

  • "Bolt_Tightness_Check"
  • "Label_Presence_Verify"
  • "Surface_Quality_ROI"
  • "Component_Alignment"

image.png

Inspection Types: Grouping and Organization​

Default: "Inspection Type 1"​

The system creates "Inspection Type 1" as the default grouping name, which should be changed to reflect the actual inspection purpose.

Inspection Type Function:

  • ROI Grouping - Organizes related ROIs under logical categories
  • Result Organization - Groups inspection outcomes by type
  • Workflow Management - Simplifies complex multi-area inspections

image.png

Critical: Node-RED Integration​

Why Inspection Type Names Matter: It is important to change the Inspection Type name because Node-RED uses this name for output publication. The name becomes the identifier for data routing and system integration.

Node-RED Output Impact:

  • Data Publishing - Results are published using Inspection Type names
  • External Integration - Other systems receive data with these identifiers
  • Automation Logic - Node-RED flows reference these names for routing

Effective Naming Strategy:

  • Process-Based: "Quality_Control", "Dimensional_Check"
  • Feature-Based: "Surface_Inspection", "Component_Verification"
  • Location-Based: "Top_Panel_Check", "Side_Assembly"
  • Avoid Generic: Replace "Inspection Type 1" with specific identifiers

Setup Workflow​

Configuration Process​

Essential Steps:

  1. Inspection Type Naming - Change from "Inspection Type 1" to meaningful name
  2. ROI Creation - Add rectangular regions using + Add Inspection Region
  3. ROI Naming - Change from "New ROI" to descriptive names
  4. Positioning - Size and position ROIs appropriately
  5. Validation - Verify setup meets inspection requirements

Multiple ROI Management​

Organization Strategy:

  • Logical Grouping - Group related ROIs under appropriate Inspection Types
  • Consistent Naming - Use systematic naming conventions
  • System Integration - Consider Node-RED output requirements

Integration Considerations​

Template Alignment Integration​

Alignment Benefits:

  • Automatic Adjustment - ROIs move with part alignment corrections
  • Consistent Coverage - ROIs maintain relative position to part features
  • Position Tolerance - Part variations don't affect inspection accuracy

Node-RED Compatibility​

Integration Requirements:

  • Unique Names - Ensure Inspection Type names are unique for outputs
  • Naming Standards - Establish conventions for system integration
  • Output Testing - Verify Node-RED receives correct identifiers

Best Practices​

Effective ROI Configuration​

Key Guidelines:

  • Descriptive Naming - Both ROI and Inspection Type names should be meaningful
  • Appropriate Sizing - Balance feature coverage with processing efficiency
  • Strategic Organization - Group related ROIs logically

System Integration​

Planning Considerations:

  • Node-RED Requirements - Plan Inspection Type names for data routing
  • Documentation - Record ROI purposes and integration points
  • Testing - Validate that outputs reach intended systems correctly

🔗 See Also​