Rule-Based

Before teaching a system to "think" like a human, you often just need it to follow strict, unbreakable laws. In industrial automation, this is called Rule-Based Intelligence. Instead of training neural networks with thousands of sample images, our Edge-Vision-4.0 software allows you to set clear, mathematical parameters to detect defects instantly.

Step 1

Analyze defect images

We look at actual photos of defects collected from your production line to examine exactly how they appear on screen.

Step 2

Define your quality standards

You specify exactly when an anomaly should be treated as a real defect and when it can be safely ignored, giving you total control over your inspection criteria.

Step 3

Set the exact parameters

We translate your standards into strict mathematical rules using four clear parameters: minimum/maximum width, height, shape (rectangular or not), and the exact contrast value.

Step 4

Categorize and name the type

The specific rule is saved and assigned a unique name or letter code (type). This ensures the software classifies and logs the defect correctly every time it passes the camera.

Step 5

Deploy with 100% insight

The rules go live instantly. Because you defined the parameters yourself, you have complete insight into the system's logic—there is no algorithmic guesswork, just absolute predictability.

How Rule-Based Vision Works

The system inspects the continuous material web in real time by looking at the pixels through a set of strict mathematical calculators. It filters out anomalies based on three main pillars:

  • Contrast & Intensity (Gold Vest Value): The software continuously checks the surface brightness. If a spot is suddenly a specific percentage darker or lighter than your perfect material baseline (like a dark oil stain or a bright light hole), it is immediately flagged.
  • Size & Pixel Count: To prevent false alarms from harmless loose dust particles, you can set strict size limits. The system will safely ignore minor specs but triggers an alarm the exact millisecond a defect crosses your minimum length or width thresholds.
  • Geometry & Direction: Defects often have specific shapes. For example, a scratch caused by a mechanical roller always forms a long, thin vertical or horizontal line. Rule-based setups excel at filtering shapes based on their aspect ratio and orientation along the material flow.

Why Choose Rule-Based Setups?

While modern learning AI gets a lot of attention, rule-based classification remains the absolute backbone of factory quality control for three major reasons:

  • Instant Setup: You do not need a long commissioning phase to collect data. You enter the dimensions and contrast limits, and the system is ready to inspect from day one.
  • 100% Predictable Outcomes: There is no "black box" mystery. If a defect meets the mathematical criteria, it is caught. If it doesn't, it passes. This absolute transparency makes troubleshooting incredibly easy for operators.
  • Ultra-High Processing Speeds: Because processing basic math requires very little computing power, rule-based systems are lightning-fast. They can track ultra-high-speed production lines without a single millisecond of delay or motion blur.

The Best of Both Worlds: Hybrid Inspection

The most reliable production lines use a hybrid setup: rule-based filters instantly catch the obvious, high-speed defects (like holes and edge cracks), while our AI models run simultaneously to analyze complex, irregular surface textures.