JPS6155690A - Detection of 2-d shape positional deviation - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of Industrial Application This invention is directed to a two-dimensional shape displayed on the screen of a cathode ray tube (hereinafter abbreviated as CRT), in which small shape elements such as phosphors are distributed discretely. The present invention relates to a method for detecting positional deviation of a complicated two-dimensional shape formed by an alignment pattern.

Conventional structure and its problems Conventional color CRT red, green, blue (hereinafter abbreviated as ROB)
As an automatic color CRT misconvergence inspection method that automatically inspects whether the beams from each electron gun in the color layer are converging on the same point, the test pattern displayed on the CRT screen is detected by a pattern input device. A method has been adopted in which the detected two-dimensional image signal is stored in a frame buffer or converted into run-length coded data, and then this image information is processed to recognize the position of the test pattern.

However, methods that use frame buffers and run-length encoded data to recognize the position of two-dimensional shapes,
Generally, it has the disadvantage of slow processing speed.

Therefore, it is conceivable to use a method of recognizing the position of a two-dimensional shape by projection processing, which has a faster processing speed than the above two methods, but in projection processing, two-dimensional original image signals are
Because it is compressed into dimensional data, it is difficult to recognize the position of a complex two-dimensional shape consisting of a large number of discretely distributed shape elements like phosphors on a CRT screen, that is, a two-dimensional shape composed of aligned patterns. Met.

OBJECTS OF THE INVENTION An object of the present invention is to overcome the above-mentioned problems and provide a method for detecting positional deviation of a two-dimensional shape, which can perform position recognition of a two-dimensional shape composed of alignment patterns at high speed.

Structure of the Invention In order to achieve this object, the method for detecting positional deviation of a two-dimensional shape according to the present invention detects a complicated two-dimensional shape consisting of an alignment pattern using a pattern input device, converts it into a two-dimensional image signal, The position of the two-dimensional shape is recognized from projection data obtained by sending the two-dimensional image signal to a pattern recognition device, binarizing the obtained two-dimensional image signal, and projecting it onto a pair of reference axes.

The feature of this method is that a pattern recognition device cuts out a plurality of small regions including the boundaries of the two-dimensional shape to be recognized from a binarized two-dimensional image signal, and identifies the centroid of each small region and each The measurement position of the two-dimensional shape is determined from the sum of the areas of the shape elements included in the small region, and the positional deviation of the two-dimensional shape is detected by calculating the difference from the standard position stored in advance. be.

Therefore, by cutting out multiple small regions from a complex two-dimensional shape composed of alignment patterns and reducing the number of shape elements handled in one recognition process, complex two-dimensional shapes can be processed using fast projection processing. It becomes possible to recognize it.

DESCRIPTION OF EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 shows a schematic configuration of a color CRT misconvergence automatic inspection apparatus which is an embodiment of the present invention. When a test pattern 2 consisting of nine circular spots generated by the pattern generation circuit 1 is displayed on the screen of the CRT 3, three television cameras 4, 5, 6
A numerical control circuit 7 controls the position of a numerical control table 8 (hereinafter referred to as an NC table) on which a CRT 3 is mounted so that the test pattern 2 is imaged. One of the three television cameras 4, 5, and 6 (for example, television camera 4) is sequentially selected by the camera selection circuit 9 and controlled by the camera control circuit 10. Then, the image signal of the circular spot in the test pattern 2 detected by the selected television camera 4 is converted into an analog-to-digital converter (
It enters a circuit 11 (hereinafter abbreviated as A/D conversion), where it is binarized by a binarization control circuit 2, a small area including the boundary of the circular spot is cut out by a window frame control circuit 13, and then projected. A two-dimensional image signal is information-compressed into a pair of one-dimensional image data by a processing circuit 14, which is processed by a central processing unit (hereinafter abbreviated as CPU).

A microcomputer (pattern recognition device: CPU chip is MNI610 or MN1) consisting of read-only memory, read/write memory, input/output ports, etc.
613).

As a pattern recognition device, the main controller 15 or operation! A determination control circuit, that is, a CPU 17, to which instructions are given from 16, a position recognition circuit 18 that recognizes the measured position of the circular spot from the projection data of the circular spot, a standard position memory circuit 19 that stores the standard position of the circular spot, and a circular spot. It is comprised of a measurement position memory circuit 20 for storing the measurement position of , a memory control circuit 21 for controlling the two memory circuits 19 and 20, and a recognition circuit 22 for comparing the standard position and the measurement position.

The recognition result of the circular spot is sent to the main controller 15, and each time one round of recognition is completed, the main controller 15 sends a sequence signal to the sequence control circuit 23.
The sequence control circuit 23 is a numerical control circuit 5. Controls the pattern generation circuit 1.

Next, the operation of the pattern recognition device configured as described above will be explained. FIG. 2 shows a test pattern for automatic color CRT misconvergence inspection consisting of nine circular spots, and FIG. 3 shows a recognition processing procedure for the test pattern.

First, three television cameras 4 to 6 are sent to A- in Figure 2.
After being positioned so as to image the three circular spots above (step 1), the television camera 4 is selected by the camera selection circuit 9 [step 2] and the image signal of the circular spot CI+ is detected. And circular pants l” CI
The position recognition of I is performed sequentially for each color of RGB [step 3], the recognition data is sent to the main controller 15 [step 4], and the recognition process for the circular spot CII is completed.

Next, the television camera 5 is selected by the camera selection circuit 9, and the circular spot 7) is recognized.

In this way, the three circular spots on line A-A are CI
I, C2+, C31 are recognized in order from the top, and A-A
The processing for the circular subo-nod on the line ends.

Then, the NC table 8 moves and the process moves on to the circular spot on the line B-B. The inspection of one CRT3 is performed in order from the left (or from the right) A-A, B-B, c-
circular spot on cvA (or C-C, B-B, A
- circular spot on line A) and end. As shown in FIG. 2, a relative coordinate system XRYi for each circular spot is set as a coordinate system for recognizing circular spots.

Furthermore, the operation of the pattern recognition device will be explained using FIGS. 4 to 7, but before that, the relationship between each figure will be described. FIG. 4 is an enlarged view of the circular spot constituting the test pattern, and shows the two-dimensional shape to be recognized, that is, the positional relationship between the circular spot and the RGB phosphor that is the alignment pattern, and the plurality of small regions (Ri ,R2,R3,R4
)It is shown. Figure 6 shows the small area R in Figure 4 above.
2 is enlarged, showing the light-emitting phosphors (E2□, E2□, R23, R24, E2S) that are the shape elements of the alignment pattern.
It is shown. Further, FIG. 5 shows a procedure for recognizing the position of one circular spot, and FIG. 7 shows a procedure for detecting the centroid of one small area.

The operating procedure of the pattern recognition device is as shown in FIGS. 4 and 5. First, four small head regions Rz (
1≦i≦4) [Step 11], each small region R
The process starts by detecting the centroid G=(Xt+)'t) of i [step 12].

The procedure for detecting the centroid Gi of each small region Ri is shown in FIGS. 6 and 7 (in the case of i=2 in FIG. 6), the small region R
All m light emitting phosphors Ei contained in i, (1≦j
≦m, Fig. 6 shows the case where m=5), the centroid g ij (X
ij + Y ij) and its area Sij are determined from projection data obtained by projecting the two-dimensional image signal onto a pair of reference axes [Step 21], and the luminescent phosphor E
From the centroid g = i (X + j, Y =;) of i j and its area S ij, the centroid CB (x,) of the small region Ri.

yi) [Step 22], small region R.

The formula for calculating the centroid Gi (Xi + Yi) is as follows.

However, S is the area of the light-emitting phosphor without chipping.

Returning to the explanation of FIGS. 4 and 5, the centroid G of each capitula Ri obtained as described above.

(xt, yt) and the sum St of the area S ij of the light-emitting phosphor E ij included in the small region, the measurement position (xe, yc) of the circular spot is determined as shown in the following equation [Step 13].

However, S, = Σ Sij (5) If The measurement position (xc) of the circular spot is determined as described above.

yc) and the standard position (x3゜y
・) Calculate the difference Δx and Δy, and calculate the difference ΔX.

Let Δy be the positional deviation amount of the circular pattern. This Δx
, Δy are given by the following equation.

ΔX=Xe−X, ......(6)Δ)'=7
cm)'s ・・・・・・(7) Standard position (xs,
yi is calculated and stored for each RGB color by the above method using a color CRT whose misconvergence amount is guaranteed to be within a standard. Therefore, detection of the positional deviation of the circular spot is performed for each color of RGB, and Δx, which is calculated by equations (6) and (7),
Δy is also calculated three times for each RGB color, and the measurement position and positional deviation amount ΔX for each RGB color are the recognition results.

Δy is sent to the main controller 15 [Step 14
].

In the above embodiment, the test pattern 2 is made up of nine circular spots, but the number may be any number as long as it is not extremely large or small. The shape of the spot may also be triangular, square, or star-shaped.

In addition, there are 3 TV cameras (TV camera 4° 5.6)
Although the CRT 3 is moved by the NC table 8, the number of television cameras may be nine, and the NC table 8 and the numerical control circuit 7 may be omitted.

Effects of the Invention As described above, the present invention can generate a complex two-dimensional shape by cutting out a plurality of small regions encompassing the boundary of the shape from a complex two-dimensional shape composed of alignment patterns to be recognized. This method simplifies the process and enables position recognition of the two-dimensional shape by projection processing with a high processing speed, and has great practical effects.

[Brief explanation of drawings]

FIG. 1 shows a color C for detecting positional deviation of a two-dimensional shape composed of aligned patterns based on an embodiment of the present invention.
FIG. 2 is an explanatory diagram showing a test pattern displayed on a CRT for color CRT misconvergence automatic inspection, which is an embodiment, and FIG. 2 is a flowchart showing the test pattern inspection procedure; FIG. 4 is an explanatory diagram showing the positional relationship between the circular spot forming the test pattern, the phosphor that is the alignment pattern, and the small area to be cut out;
FIG. 5 is a flowchart showing the procedure for determining the circular sporator measurement position, FIG. 6 is an enlarged view of the small region R2 in FIG. 4,
FIG. 7 is a flowchart showing the centroid detection procedure for one small area. 2...Test pattern, 4,5.6...TV camera, 9...Camera selection circuit, 11...A/D conversion circuit, 12...Binarization control circuit, 13... Window frame control circuit, 14... Projection processing circuit, 17... Judgment control circuit, 18... Position recognition circuit, 19... Standard position memory circuit, 20... Measured position memory circuit, 22...・Identification circuit Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7

Claims (2)

[Claims] (1) A two-dimensional shape consisting of discretely distributed small shape elements is detected by a pattern input device, converted into a two-dimensional image signal, and further, the two-dimensional image signal is binarized and projected onto a pair of reference axes. From the projection data obtained by this, n small regions R_i (1
≦i≦n), and the centroid (x_i
, y_i) and the total area S_i_j of the m shape elements E_i_j (1≦j≦m) included in each small region S_
From i, the measurement position of the above two-dimensional shape (x_c, y_c)
, and calculate the above measurement position (x_c, y_c) and the standard position (x_s, y_s) of the above two-dimensional shape stored in advance.
Based on equations (A) and (B), Δx, Δ
A method for detecting positional deviation of a two-dimensional shape, characterized in that positional deviation of the two-dimensional shape is detected by calculating y. Δx=x_c-x_s...(A) Δy=y_c-y_s...(B) (2) The measurement position (x_c, y_c) of the two-dimensional shape is determined by the centroid (x_i, y_i) of the n small regions R_i and the area S_ of the m shape elements E_i_j included in each small region.
Total sum S_i of i_j (=Σ＾m_j_=_1S_i_j
), the centroid (x_i, y_i) of the small region R_i is given by the equations (C) and (D), and the centroid (x_i, y_i) of the small region R_i is
The centroid (x
_i_j, y_i_j) and the area S_i_j of the shape element, the method for detecting a positional deviation of a two-dimensional shape according to claim (1) is given by equations (E) and (F). x_c=(Σ^n_i_=_1x_i・S_i)/(Σ
^n_i_=_1S_i)・・・・・・(C)y_c=
(Σ^n_i_=_1y_i・S_i)/(Σ^n_i
＿＝＿１S＿i）・・・・・・(D)x＿i＝(Σ＾m
＿j＿=_1x_i_j・S_i_j)/(mSp)・
...(E)y_i=(Σ＾m_j_=_1y_i
_j·S_i_j)/(mSp) (F) where Sp is the area of the shape element without chipping.