CN103913461A - TFT-LCD lighting automatic optical inspection based image processing method - Google Patents

Abstract

The invention relates to the technical field of image detection, and especially relates to a TFT-LCD lighting automatic optical inspection based image processing method. The image processing method comprises the following steps: acquiring images of a display screen in different modes; preprocessing the acquired images to remove the background interference; carrying out mask treatment on the preprocessed images; and carrying out defect extraction and calculation. The TFT-LCD lighting automatic optical inspection based image processing method uses a plurality of automatic optical inspection modes, can realize the normal detection of spot, line and Mura defects, and can realize the accurate positioning of light spot and light line defects to a sub-pixel, and the classification of the Mura defects according to an SEMU criterion. The image processing method has the advantages of simple flow, and effective ensure of the detection rate.

Description

Image processing method for automatic optical detection of TFT-LCD lighting

Technical Field

The invention relates to the technical field of image detection, in particular to an image processing method for automatic optical detection of TFT-LCD lighting.

Background

A TFT-LCD (Thin Film transistor-Liquid Crystal Display) is an electro-optical Display device that realizes Display by controlling the orientation of Liquid Crystal molecules having refractive index anisotropy to vary the transmittance through a Liquid Crystal panel, and a Thin Film Transistor (TFT) plays a switching role therein. Compared with a traditional Cathode Ray Tube (CRT) display, the Cathode Ray Tube (CRT) display has the characteristics of low power consumption, long service life, no radiation, small volume and the like. In recent years, the liquid crystal display has been made to stand out from the intense competition among a plurality of flat panel displays, becomes a new generation of mainstream display, and is widely applied to the display fields of notebook computers, desktop displays, liquid crystal televisions, mobile display terminals and the like.

With the rapid development of TFT-LCD technology, the improvement of product quality and the reduction of cost become the goals pursued by the industry. The most effective and direct way to reduce the cost is to improve the yield of the product. The traditional method for detecting the product yield adopts manual operation, so that the workload is high, and the influence of subjective factors of detection personnel is easily caused, so that the detection efficiency and precision cannot be ensured. Especially, with the continuous improvement of the automation degree of the production process, the manual visual inspection can not meet the requirements of the current industrial field.

Machine vision technology based on image processing technology, which has been rapidly developed in recent years, can solve this problem. The automatic optical detection equipment can greatly improve the production efficiency and the automation degree of production, and the machine vision is easy to realize information integration, thereby meeting the requirements of digital and automatic production.

The invention is based on the theory of machine vision, and applies an image vision algorithm to detect the defects of the TFT-LCD screen, thereby achieving the purposes of high speed, high precision and labor cost saving.

Disclosure of Invention

Technical problem to be solved

The invention aims to solve the technical problem of providing an image processing method for automatic optical detection of TFT-LCD lighting so as to overcome the defects of low detection speed and low precision caused by manual detection in the prior art.

(II) technical scheme

In order to solve the above technical problem, the present invention provides an image processing method for automatic optical detection of TFT-LCD lighting, comprising:

s1, collecting images of the display screen in different modes;

step S2, preprocessing the collected image and removing background interference;

step S3, performing mask processing on the preprocessed image;

and step S4, defect extraction calculation.

Further, between step S1 and step S2, the method further includes:

and carrying out boundary positioning on the acquired image to determine a defect detection image area.

Further, the preprocessing the image specifically includes: and filtering the image by adopting a Gabor filtering operator.

Further, the Gabor filter operator includes:

<math> <mrow> <mi>g</mi> <mrow> <mo>(</mo> <mi>x</mi> <mo>,</mo> <mi>y</mi> <mo>,</mo> <mi>f</mi> <mo>,</mo> <mi>&theta;</mi> <mo>)</mo> </mrow> <mo>=</mo> <mfrac> <mn>1</mn> <mi>ab</mi> </mfrac> <mi>exp</mi> <mo>[</mo> <mo>-</mo> <mi>&pi;</mi> <mrow> <mo>(</mo> <mfrac> <msubsup> <mi>x</mi> <mi>r</mi> <mn>2</mn> </msubsup> <msup> <mi>a</mi> <mn>2</mn> </msup> </mfrac> <mo>+</mo> <mfrac> <msubsup> <mi>y</mi> <mi>r</mi> <mn>2</mn> </msubsup> <msup> <mi>b</mi> <mn>2</mn> </msup> </mfrac> <mo>)</mo> </mrow> <mo>]</mo> <mo>[</mo> <mi>exp</mi> <mrow> <mo>(</mo> <mi>i</mi> <mn>2</mn> <mi>&pi;</mi> <msub> <mi>fx</mi> <mi>r</mi> </msub> <mo>)</mo> </mrow> <mo>-</mo> <mi>exp</mi> <mrow> <mo>(</mo> <mo>-</mo> <mfrac> <msup> <mi>&pi;</mi> <mn>2</mn> </msup> <mn>2</mn> </mfrac> <mo>)</mo> </mrow> <mo>]</mo> </mrow> </math>

x_r=xcosθ+ysinθ，y_r=-xsinθ+ycosθ

a = b =1/f, f: center frequency, θ: and (4) filtering direction.

Further, the boundary positioning specifically adopts a binarization method to separate a defect detection image area from a non-detection image area, and a binarization threshold value adopts a global threshold value as m;

further, the defect extraction calculation specifically includes:

$\mathrm{SEMU}=\frac{\left|C_x\right|}{C_{\mathrm{jnd}}}=\frac{\left|C_x\right|}{(\frac{1.97}{S_x^{0.33}}+0.72)}$

wherein,

S_xis the area of the Mura defect;

C_jndcontrast for the human eye with minimal perceptible difference;

C_xis the average contrast of the Mura region.

(III) advantageous effects

The image processing method for the automatic optical detection of the TFT-LCD lighting provided by the embodiment of the invention applies a plurality of automatic optical detection modes, can normally detect the defects of points, lines and Mura, and can accurately position the defects of bright points and bright lines to one sub-pixel and classify the Mura defects according to SEMU criteria. The image processing flow is simple and convenient, and the detection rate can be effectively ensured.

Drawings

FIG. 1 is a flow chart of an image processing method for automatic optical inspection of TFT-LCD lighting according to an embodiment of the present invention;

FIGS. 2 (a) - (d) are schematic diagrams illustrating the boundary positioning of images according to the embodiment of the present invention;

FIGS. 3 (a) - (b) are schematic diagrams illustrating filtering processing performed on an image according to an embodiment of the present invention;

FIGS. 4 (a) - (c) are schematic diagrams illustrating the masking process performed on the image according to the embodiment of the present invention;

fig. 5 is a schematic diagram of a TFT pixel structure.

Detailed Description

The following detailed description of embodiments of the present invention is provided in connection with the accompanying drawings and examples. The following examples are intended to illustrate the invention but are not intended to limit the scope of the invention.

The invention adopts an area array CCD camera to collect and light images of a TFT-LCD screen in different modes of black, gray, white, red, green, blue and the like, transmits the collected images to a computer and enters an image processing process. In the image processing process, the actual detection region ROI of the image is extracted, then the defects in all modes are detected, the defects are accurately positioned, the defect information is recorded, and the defect information is transmitted to a repair section.

As shown in fig. 1, the image processing method for automatic optical detection of TFT-LCD lighting according to the embodiment of the present invention specifically includes:

step 1, carrying out boundary positioning on the acquired image.

The area array CDD collects images under each mode of the TFT-LCD screen, the unlighted edge part and the workbench part for placing the screen are also collected at the same time, and an actual detection region ROI needs to be extracted firstly, and then the next image processing step is carried out.

Specifically, referring to fig. 2 (a) - (d), the ROI region of the TFT-LCD image is well defined on the color scale than the black background, the ROI region is separated from the background region by using a binarization method, and the binarization threshold can be set as m by using a global threshold;

in the gray-scale images (0-255), the non-detection area is set to be black, namely 0, and the detection area is set to be white, namely 255, and reference is made to a binarization effect graph in the graph.

Boundary positioning is carried out in a binary image, a screen workbench for placing a TFT-LCD is not completely horizontal, a boundary such as a left edge is firstly found, N edge coordinate points are found, the height H and the width W of the image are fitted by a least square method, and an angle needing rotation correction is calculated.

The coordinates (X, Y) of these N points were then calculated to obtain the following results

<math> <munder> <mrow> <mi>&alpha;</mi> <mo>=</mo> <mfrac> <mrow> <mo>(</mo> <mi>&Sigma;y</mi> <mo>)</mo> </mrow> <mi>n</mi> </mfrac> <mo>-</mo> <mfrac> <mrow> <mi>&beta;</mi> <mrow> <mo>(</mo> <mi>&Sigma;X</mi> <mo>)</mo> </mrow> </mrow> <mi>n</mi> </mfrac> </mrow> <mo>&OverBar;</mo> </munder> </math>

<math> <munder> <mrow> <mi>&beta;</mi> <mo>=</mo> <mfrac> <mrow> <mi>n&Sigma;XY</mi> <mo>-</mo> <mi>&Sigma;X&Sigma;Y</mi> </mrow> <mrow> <mi>n&Sigma;</mi> <msup> <mi>X</mi> <mn>2</mn> </msup> <mo>-</mo> <msup> <mrow> <mo>(</mo> <mi>&Sigma;X</mi> <mo>)</mo> </mrow> <mn>2</mn> </msup> </mrow> </mfrac> </mrow> <mo>&OverBar;</mo> </munder> </math>

<math> <mrow> <mover> <mi>Y</mi> <mo>^</mo> </mover> <mo>=</mo> <mi>&alpha;</mi> <mo>+</mo> <mi>&beta;</mi> <mover> <mi>X</mi> <mo>^</mo> </mover> </mrow> </math>

The image is angularly rotated by the value of β, and the image is corrected to a horizontal, vertical, and non-oblique image, referring to fig. 2 (c). Only one rectangular frame in the image, i.e. the detection region ROI, can then apply the method of finding rectangular blocks, find the number total =1 of rectangular frames, extract four boundaries of the detection region, refer to 2 (d).

Step S2, preprocessing the ROI region of the detected image, and making a basis for defect extraction and defect calculation. Image preprocessing includes operations such as Gabor filtering, image segmentation, morphology, etc.

As shown in FIG. 3 (a), the captured TFT-LCD image shows a regular arrangement of repetitions

x_r＝xcosθ+ysinθ，y_r＝-xsinθ+ycosθ

By applying the above formula, the processing results in an image with uniform background and obvious defects as shown in fig. 3 (b).

Step S3, a mask process is performed on the preprocessed image.

The manufacturing process of the TFT-LCD is very complicated, has hundreds of processes, and can be divided into three basic steps in combination:

front substrate fabrication (Array process), middle liquid crystal panel assembly (Cell process), and back Module assembly (Module process). The embodiment of the invention aims at the detection of the TFT in the Celll process, and the automatic optical detection is used for lighting the TFT-LCD screen, so that the upper polarizer and the lower polarizer of the polarizer are required to be placed on the TFT-LCD screen, and the automatic optical detection is used for lighting, so that the colors of black, gray, red, green, blue and the like are displayed. Air bubbles sometimes remain in the polarizer, and during the detection process, each TFT screen causes 'pseudo-defects', so that the part presenting the air bubbles needs to be recorded, and the area is considered to be defect-free before processing images every time.

Specifically, first, a mask region is set on the user side interface, as shown in fig. 4 (a), then the above operations of boundary positioning and preprocessing are performed on the image to obtain a detected defect image, as shown in fig. 4 (b), at this time, the background and the moire of the image are filtered out, and a defect is left, and a changed region is removed from the defect image according to the setting of the mask region, that is, a region with the same color level as the background is set in the defect image, as shown in fig. 4 (c).

In the steps, the mask effect is adopted to remove the self defects of the non-TFT-LCD screen, so that the real defects are reserved, and the false detection rate is reduced.

Step S4, extracting and calculating defects;

after the above processing, only the defect information is retained in the image and is clearly separated from the background. And applying an edge searching operator to the image to search defect information, and calculating the attributes of the searched defects, including the area, the length, the width and the gray value. As shown in fig. 5, one TFT pixel includes three R, G, B sub-pixels, and for the following repair segment, the defect information is located in one sub-pixel, that is, the coordinate position of the bright spot and the bright line is recorded.

Mura defect graded statistics is carried out according to SEMI (semiconductor Equipment and Materials International) which is an International organization for semiconductor equipment and Materials, and in 2002, SEMI D31-1102 standard, namely SEMU index, for detecting Mura defect in image quality of liquid crystal display is provided.

$C_{\mathrm{jnd}}=(\frac{1.97}{S_x^{0.33}}+0.72)$

$\mathrm{SEMU}=\frac{\left|C_x\right|}{C_{\mathrm{jnd}}}=\frac{\left|C_x\right|}{(\frac{1.97}{S_x^{0.33}}+0.72)}$

Wherein,

S_xis the area of the Mura defect;

C_jndcontrast for the human eye with minimal perceptible difference;

C_xis the average contrast of the Mura region.

In the step, the TFT-LCD screen is applied to a plurality of automatic optical detection modes, so that the defects of points, lines and Mura can be normally detected, and the defects of bright spots and bright lines can be accurately positioned to one sub-pixel and the defects of Mura can be classified according to SEMU criteria. The treatment process is simple and convenient, and the detectable rate is ensured.

The embodiment of the invention provides an image processing method for automatic optical detection of TFT-LCD lighting, which adopts a backlight source of an electrical measuring instrument to light a TFT-LCD screen, applies automatic optical detection to automatically switch and detect black, gray, white, red, green, blue and the like of interfaces, acquires images by a planar array CCD camera under different modes, detects the defects of bright spots, bright lines, dark spots, dark lines and Mura caused by electricity, and finally extracts defect information, stores and transmits the defect information and finishes detection.

The algorithm of the invention has complete design functions, and solves the problems of off-line and on-line TFT lighting automatic optical detection defect detection.

The above description is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, various modifications and substitutions can be made without departing from the technical principle of the present invention, and these modifications and substitutions should also be regarded as the protection scope of the present invention.

Claims (6)

1. An image processing method for automatic optical detection of TFT-LCD lighting is characterized by comprising the following steps:

s1, collecting images of the display screen in different modes;

step S2, preprocessing the collected image and removing background interference;

step S3, performing mask processing on the preprocessed image;

and step S4, defect extraction calculation.

2. The image processing method as claimed in claim 1, further comprising, between step S1 and step S2:

and carrying out boundary positioning on the acquired image to determine a defect detection image area.

3. The image processing method according to claim 1, wherein the preprocessing the image specifically comprises: and filtering the image by adopting a Gabor filtering operator.

4. The image processing method of claim 3, wherein the Gabor filter operator comprises:

<math> <mrow> <mi>g</mi> <mrow> <mo>(</mo> <mi>x</mi> <mo>,</mo> <mi>y</mi> <mo>,</mo> <mi>f</mi> <mo>,</mo> <mi>&theta;</mi> <mo>)</mo> </mrow> <mo>=</mo> <mfrac> <mn>1</mn> <mi>ab</mi> </mfrac> <mi>exp</mi> <mo>[</mo> <mo>-</mo> <mi>&pi;</mi> <mrow> <mo>(</mo> <mfrac> <msubsup> <mi>x</mi> <mi>r</mi> <mn>2</mn> </msubsup> <msup> <mi>a</mi> <mn>2</mn> </msup> </mfrac> <mo>+</mo> <mfrac> <msubsup> <mi>y</mi> <mi>r</mi> <mn>2</mn> </msubsup> <msup> <mi>b</mi> <mn>2</mn> </msup> </mfrac> <mo>)</mo> </mrow> <mo>]</mo> <mo>[</mo> <mi>exp</mi> <mrow> <mo>(</mo> <mi>i</mi> <mn>2</mn> <mi>&pi;</mi> <msub> <mi>fx</mi> <mi>r</mi> </msub> <mo>)</mo> </mrow> <mo>-</mo> <mi>exp</mi> <mrow> <mo>(</mo> <mo>-</mo> <mfrac> <msup> <mi>&pi;</mi> <mn>2</mn> </msup> <mn>2</mn> </mfrac> <mo>)</mo> </mrow> <mo>]</mo> </mrow> </math>

x_r=xcosθ+ysinθ，y_r=-xsinθ+ycosθ

a = b =1/f, f: center frequency, θ: and (4) filtering direction.

5. The image processing method according to claim 4, wherein the boundary positioning specifically employs a binarization method to separate a defect detection image area from a non-detection image area, and a binarization threshold is set to m by using a global threshold;

6. the light-emitting apparatus according to claim 1, wherein the defect extraction calculation is specifically: $C_{\mathrm{jnd}}=(\frac{1.97}{S_x^{0.33}}+0.72)$

$\mathrm{SEMU}=\frac{\left|C_x\right|}{C_{\mathrm{jnd}}}=\frac{\left|C_x\right|}{(\frac{1.97}{S_x^{0.33}}+0.72)}$

wherein,

S_xis the area of the Mura defect;

C_jndcontrast for the human eye with minimal perceptible difference;

C_xis the average contrast of the Mura region.