Application of Image Technology in Printing Inspection (Part 2)

1) Surface Inspection Technology

High-precision, high-speed, robust image registration technology

Golden Template generation and comparison technology

Blob Analysis Technology·

Defect Cluster Analysis Technology

. . .

2) Image Measurement Technology

Color measurement

Color imaging calibration

A color space and feature selection

A color change description and calculation

Geometric measurement

An imaging system geometry calibration

Sub-pixel edge detection

Object shape curve fitting

One parameter measurement

3) Image Recognition Technology

Line Feature Robust Extraction

Line feature comprehensive decision

New and old feature extraction of banknotes

Banknotes old and new classifier design

· Face recognition

Number identification

4) Introduction to Surface Inspection Technology

l General procedure of surface inspection system

The surface inspection system is actually a machine vision system. Its flow is:

Figure 2 General process of surface inspection system

l High-precision positioning registration algorithm

a. The role of positioning algorithm in machine vision system

â–  Image feature extraction.

â–  Mainly relative position information, including rotation angle, horizontal and vertical offsets, provides information for various subsequent detection algorithms and belongs to the bottom layer of the system. Offset color, color mixing, wiring, overprinting, printing and other detection algorithms in off-line offset inspection systems all require the use of positioning module data.

â–  Linear simulation of nonlinear deformations. Due to the non-linear deformation of the paper, only a sufficient number of positioning points can realistically simulate the nonlinear deformation in the line segment.

b. Classification of positioning algorithms

â–  Use image grayscale features for positioning

â–  Use image geometry for positioning

â–  Use image space domain information for positioning

l Blob cluster analysis algorithm

a.Blob analysis algorithm

â–  Blob analysis algorithms, also known as speckle analysis algorithms, are often used to extract and classify image features from a target image.

■ By analyzing the graphical characteristics of the Blob unit, a simple pattern of gray information can be rapidly converted into the shape information of the pattern, including the graphic center of mass, the area of ​​the graph, the perimeter of the graph, the minimum rectangle of the graph, and other graphic information.

â–  The Blob analysis algorithm plays an important role in the surface detection. It can judge the true defects and the false defects according to different graphic characteristics.

â–  Blob analysis algorithms can also be used in the field of particle counting.

b. Blob analysis step

i. Blob image segmentation

â– Fixed Threshold

Suitable for images with high contrast and good background consistency.

â–  Variable Threshold

Segmentation is performed using statistical information such as minimum mean, maximum mean, and mean squared error. Suitable for some images with poor contrast and consistency.

Ii. Blob feature extraction

â–  After image segmentation, the grayscale information of the image can be converted into a Blob message queue through the Blob feature extraction process.

â–  According to the requirements of the surface quality detection application, targeted algorithm can be used to extract more Blob shape information.

c. Typical Blob Features

The typical Blob features are: ratio of Blob area to circumscribed rectangle area, Blob extension rate, and the long axis angle of the minimum bounding ellipse.

Source: Print Information