Computer Vision for Detecting Flash and Sink Marks in Plastic Injection Molding

In the plastic injection molding industry, the quality of the final product depends on many factors: process parameters, mold quality, material temperature, and production cycle stability. Even small variations can generate visible defects on molded parts.

Among the most common defects are flash and sink marks, issues that can affect the aesthetics, functionality, and dimensional conformity of plastic components.

Traditionally, the inspection of these defects has relied on manual quality control, but this approach has clear limitations in terms of speed, reliability, and repeatability. For this reason, more and more manufacturers are adopting Computer Vision systems capable of automatically detecting these anomalies directly on the production line.

What Are Flash and Sink Marks in Injection Molding?

During the injection molding process, several types of surface defects may occur.

Flash

Flash is a thin excess of plastic material that appears along the mold parting lines or at junction points where the mold halves meet.

The main causes may include:

  • injection pressure that is too high

  • mold wear or misalignment

  • imperfect mold closure

  • suboptimal process parameters

Flash often requires manual trimming or, in more severe cases, leads to the rejection of the molded part.

Sink Marks

Sink marks are small depressions on the surface of a molded part, often visible in areas where the component has thicker sections.

They can be caused by:

  • uneven cooling

  • insufficient holding pressure

  • non-optimal part design

  • uneven material distribution

This defect can compromise the aesthetic appearance of the product and, in some cases, its structural strength.

Challenges of Manual Inspection

In modern manufacturing environments, injection molding cycles are extremely fast. Parts are produced in high volumes and often handled automatically by robotic systems.

In this context, manual inspection presents several limitations:

  • difficulty inspecting 100% of produced parts

  • variability between operators

  • challenges in detecting very small defects

  • potential slowdowns in the production process

To ensure high quality standards, many manufacturers are implementing automated inspection systems based on Computer Vision.

How Computer Vision Works in Plastic Injection Molding

Computer Vision systems use high-resolution industrial cameras, controlled lighting, and image analysis algorithms to examine every molded part.

The inspection process typically involves four main steps.

1️⃣ Image Acquisition

When the part is removed from the mold or placed on a conveyor belt, one or more cameras capture high-definition images.

Lighting is designed to highlight:

  • edges and parting lines

  • surface defects

  • variations in texture or geometry

This setup makes even very small defects clearly visible.

2️⃣ Image Analysis

Computer Vision algorithms analyze the captured images and compare them with reference models of compliant parts.

The system can identify:

  • excess material along edges

  • surface deformations

  • depressions caused by sink marks

  • variations in profile or geometry

Thanks to AI and Deep Learning, advanced systems can also detect complex defects that are difficult to program manually.

3️⃣ Defect Identification

When an anomaly is detected compared to predefined parameters, the part is classified as non-compliant.

The system can distinguish between:

  • acceptable parts

  • parts with minor aesthetic defects

  • parts that must be rejected

This classification enables more precise quality management.

4️⃣ Automatic Action

Once a defect is detected, the system can automatically trigger several actions:

  • automatic rejection of defective parts

  • alerts to operators

  • recording of defects in the quality management system

  • statistical analysis of detected defects

These insights can also be used to improve injection molding process parameters.

Benefits of Computer Vision in Plastic Injection Molding

Implementing machine vision for quality inspection provides several advantages.

✔ 100% Production Inspection

Every part can be inspected without slowing down the production process.

✔ Higher Accuracy

Cameras can detect extremely small defects that are difficult to identify manually.

✔ Improved Process Stability

Data analysis helps identify variations in process parameters quickly.

✔ Reduced Scrap Rates

Problems are detected immediately, preventing the production of large defective batches.

✔ Automated Quality Control

The system ensures consistent and repeatable inspections regardless of operator involvement.

Typical Applications

Computer Vision is widely used in plastic injection molding across several industries, including:

  • automotive components

  • plastic packaging

  • medical devices

  • consumer electronics

  • household products

In all these applications, automated inspection of surface defects is essential to guarantee product quality and reliability.

The Future: Integration with Intelligent Manufacturing Systems

Vision systems are evolving rapidly thanks to integration with Artificial Intelligence, advanced analytics, and MES systems.

This enables manufacturers to:

  • monitor product quality in real time

  • correlate defects with process parameters

  • predict anomalies before they occur

  • continuously improve production processes

Visual inspection is therefore becoming a key component of smart manufacturing environments.

Conclusion

In plastic injection molding, defects such as flash and sink marks can significantly impact product quality and production costs.

The use of Computer Vision allows manufacturers to automate quality inspection, ensuring fast, accurate, and reliable control over every produced part.

In an increasingly automated and quality-driven industrial landscape, machine vision has become a strategic tool for improving efficiency, reducing waste, and ensuring products consistently meet required standards.

Want to know more? Contact us at info@metalya.it

Previous
Previous

Detecting Deformations in Molded Parts with Computer Vision

Next
Next

Computer Vision for Batch Verification in Food Products