JP2835262B2 - Image processing system - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing system, and more particularly to an image processing system used for detecting an image by pattern matching.

[0002]

2. Description of the Related Art As a method of image pattern recognition, there is a similarity method by pattern matching. Conventionally, there is no concept of providing (masking) a mask (blocking / transmitting) area in reference image data (for example, stored in a memory) used for pattern matching. Therefore, in the case where an irregular image exists in the reference image data, the detection accuracy is reduced, and the image may not be used as the reference image data.

As an example of the above-mentioned matter, if there is a dice image and the number of dice eyes does not matter, and if it is desired to detect only the position of the dice, an unnecessary dice eye image becomes an obstacle. The accuracy of dice position detection is reduced.

In the case where an image to be detected using the reference image data (image to be processed) is an inclined (rotated) die, the reference image (data) is also
Similarly, it is necessary to use a tilted image. In this case, however, the reference image for the image to be tilted cannot be functioned normally unless the mask that covers the dice is also tilted.

[0005]

SUMMARY OF THE INVENTION The present invention has been made in view of the above circumstances. By providing a mask area in reference image data, a portion other than an image of a required portion to be detected from now is cut and covered. It is an object of the present invention to provide an image processing system in which the detection accuracy for a detection pattern is improved, and which can perform good detection even for an inclined processing target image.

[0006]

According to the present invention, there is provided a means for producing masked reference image data by rotating mask image data for substantially masking reference image data in the same manner as the reference image data. An image processing system comprising: Also, the present invention provides a first reference image data generating section obtained by rotating original reference image data, and mask image data for substantially masking the data of the image data generating section. A first mask image data generating unit rotated by the same angle as that described above, and a mask processing unit that substantially masks unnecessary data of the first reference image data by calculation from the data of the two image data generating units. A processing target image data transmitting unit, and a pattern matching processing unit that performs pattern matching processing by comparing data from the processing target image data transmitting unit with data from the mask processing unit and detects whether the pattern matching is successful or not. An image processing system comprising:

That is, the present invention provides a mask pattern area for hiding unnecessary portions of reference image data, thereby substantially eliminating unnecessary portions of reference image data. In this way, the detection accuracy of the pattern matching is increased, but when the processing target image is a rotated image, the reference image data is also used in the same manner, and the mask image data is also used at this time.
Similarly, by rotating the image, reference image data with a mask is prepared, so that the reference image data can be compared with the image to be processed, thereby increasing the detection accuracy for the rotated image to be processed.

[0008]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a configuration diagram showing the embodiment in an operation flow. In FIG. 1, reference numeral 1a denotes a reference-side rotation processing unit that performs a rotation process on the reference image data side;
b denotes a mask-side rotation processing unit that performs processing for rotating mask image data for turning unnecessary data in the reference image data into a state as if it is hidden by the same angle as the reference image data.
Is a reference image data generating unit (for example, a memory of data from the processing unit 1a) which is rotated as described above, and 2 is a mask image data generating unit which multiplies the data of the image data generating unit 1 by a substantial mask. (For example, processing unit 1
b), and the data from the data generator 2 is rotated by the above angle. Reference numeral 3 denotes a mask processing unit that substantially masks unnecessary data of the reference image data generation unit 1 by calculation from the data of both image data generation units 1 and 2, 4 denotes a processing target image data transmission unit, and 5 denotes A pattern matching processing section that performs pattern matching processing by comparing data from the processing target image data sending section with data from the mask processing section, and detects the success or failure of the pattern matching.

The substantial mask in the mask processing section is as follows.
This is substantially equivalent to masking the unnecessary data, such as masking the unnecessary data, invalidating the unnecessary data, or removing the unnecessary data.

FIG. 2 shows the relationship between the reference image data 11 and the mask pattern area 12. The mask pattern area corresponding to the reference image data having the relationship as shown in FIG. The data of the mask pattern 21 to be referred to is stored as shown in the right diagram of FIG. The reference image data 11 and the mask pattern area 12 in FIG. 2 may be considered as one screen or one reference screen. Each square 22 in FIG. 3 indicates a pixel, and the numerical value in each pixel 22 indicates pixel data. The internal configuration of the reference image data 11 also includes a large number of pixels, as in the right diagram of FIG.

In the right diagram of FIG. 3, the data is "0" or "1" here. In the case of "1", the pixel is a masked pixel, and in the case of "0", the pixel is not masked. The gist of the present embodiment is that, when the processing target image data rotated by the set angle is fetched from the outside into the processing target image data sending unit 4, the reference image data from the mask processing unit 3 for the processing target image data is: Since the unnecessary portion of the reference image data is substantially masked, it is possible to improve the detection accuracy of the detection necessary portion at the time of pattern matching in the processing unit 5. This will be described in more detail with reference to FIG.

The data unnecessary part of the reference image data generation part 1 is based on the data of the mask image data generation part 2
By the calculation in the mask processing unit 3, a substantial mask is applied. On the other hand, the data for one screen or one reference from the processing target image data sending unit 4 and the masked data (for one screen or one reference) from the mask processing unit are processed by the pattern matching processing unit 5. The pattern matching process is performed, and a result of the success or failure of the pattern matching is obtained from the processing unit 5.

In such processing, the mask processing unit 3
In the above, both inputs are rotated by the same angle, and the reference image data from the mask processing section 3 is removed from the reference image data which is unnecessary for the pattern matching processing in the pattern matching processing section 5. The detection accuracy at the time is improved.

FIG. 4 is an explanatory view of the case where the image data is rotated. FIG. 4A shows a state where the reference image data is 0 degrees, that is, a state before the rotation. FIG. 4B shows a case where the reference image data after rotation is 45 degrees, that is, the set rotation angle is 45 degrees. In FIG. 4, 31 indicates mask image data (mask pattern), and 32 may be considered as one screen or one reference screen. The portion of the mask pattern 31 may be considered as a group of pixel data “1”, and the surrounding portion may be considered as a group of pixel data “0”. A method for obtaining the rotated reference image data with mask shown in FIG. 4B will be described.

(A) First, the position after rotation (x, y) (where x = x ₁ , x ₂ ..., Y = y ₁ , y ₂ , ...)
Ask for.

(B) Next, a mask pattern value (0, 1) after rotation is obtained for each pixel by a mathematical expression for calculating a mask pattern value after rotation, which will be described later. (C) The above operation is performed for all the pixels of the mask pattern to create a rotation mask pattern.

[0017] Figure 5, in the above (i) term, the equation for determining the position after rotation, when determining the position p ₁ of rotating the position p ₀ angle beta (Fig. 5 only when only one pixel (Not shown). Position p ₀ before rotation (x
₀ , y ₀ ), the angle of r is α. _{First, x 0 = r · cosα ......} (1) y 0 = r · sinα ...... (2) x 1 = r · cos (α + β) ...... (3) y 1 = r · sin (α + β) ...... ( 4) The addition theorem of the trigonometric function is sin (α ± β) = sinα · cosβ ± cosα · sinβ (5) cos (α ± β) = cosα · sinβ ± sinα · cosβ (6) Therefore, from the expressions (3) to (6), x ₁ = r · cos α · cos β−r · sin α · sin β (7) y ₁ = r · sin α · cos β + r · cos α · sin β (8) and _{(2), x 1 = x 0 · cosβ} -y 0 · sinβ ...... (9) y 1 = y 0 · cosβ + x 0 · sinβ ...... because made (10), the equation (9), ( The position after rotation can be obtained by using equation (10).

In the equation for calculating the mask pattern value after rotation, p (x, y) is the value of the pixel to be rotated. n = {p (x, y) + p (x + 1, y) + p (x, y + 1) + p (x + 1, y + 1)} 4 (11) In the equation (11), when n is 0.5 or more, , The value of the mask pattern value after rotation is “1”. Also (1
In the expression (1), when n is less than 0.5, the mask pattern value after rotation is set to “0”.

The present invention is not limited to the above embodiment, but can be applied to various applications. For example, the image data generating units 1 and 2 can be realized as, for example, a memory (storage device), but the image data transmitting unit 4 can also be realized using, for example, a memory serving as a relay.

[0020]

As described above, according to the present invention, a mask pattern area for concealing unnecessary portions of reference image data is provided, and thereby unnecessary portions of reference image data are substantially eliminated. In order to increase the accuracy, if the processing target image is a rotated image, the reference image data is also used in a similar manner, and at this time, the mask image data is also used.
Similarly, by rotating the image to be prepared, so-called reference image data with a mask can be prepared and compared with the image to be processed, so that the detection accuracy for the rotated image to be processed can be improved. It is.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of an embodiment of the present invention.

FIG. 2 is a relationship diagram between reference image data and a mask pattern.

FIG. 3 is a configuration diagram of a mask pattern.

FIG. 4 is an explanatory diagram of rotation of reference image data with a mask pattern.

FIG. 5 is a coordinate diagram used for explaining an equation for obtaining a position after rotation for one pixel.

[Explanation of symbols]

1a: Reference-side rotation processing unit, 1b: Mask-side rotation processing unit, 1 ... Reference image data generation unit, 2 ... Mask image data generation unit, 3 ... Mask processing unit, 4 ... Processing image data transmission unit, 5 ... Pattern Matching processing unit.

Claims (2)

(57) [Claims] 1. An inspection image and a reference image are created.
Means for masking unnecessary portions of the reference image.
Means for creating a mask pattern, and the reference
Create a rotated reference image by rotating the image by a predetermined amount.
Means for rotating the mask pattern by the predetermined amount.
Means for creating a rotated mask pattern;
When the inspection image is rotated by the predetermined amount, the rotation lift
Pattern using the reference image and the rotation mask pattern.
An image processing system comprising means for performing a matching process . 2. The rotation reference image and the rotation
The mask pattern is used before the inspection image is created.
The image processing system according to claim 1, wherein the image processing system is created in advance .