TWM500316U - Control device for reducing misjudgment of recognizing photographed dice point - Google Patents

Description

Control device for reducing misjudgment of photography to identify dice points

This creation is a control device for reducing the number of dice judgments to identify the number of dice points. The image of the dice is processed, analyzed, and discriminated by three or more sets of recognition programs to improve the correct rate of discriminating points.

According to Fig. 1, the scorpion type is a square block, and the corners are chamfered to form a "chamfered area 5" area, and each black dot has a spacing of 4. The method of recognizing the number of dice points is as shown in Fig. 2. A computer host has an identification program. After the camera captures the images of dice 1, 2, and 3, the dice 1, 2, and 3 images are transmitted to the host computer. The identification program of the computer host is to process, analyze, and discriminate the number of dice 1, 2, and 3 images into a graphic file, turn it into a dot, compare it with the dice point in the file, etc. The problem is that the number of points is easy to occur:

(1) As shown in Fig. 1, each corner of the dice 1, 2, and 3 has a chamfered surface to form a "chamfered area 5" area, and the chamfered area 5 produces a "reflective" effect when illuminated by light. It is an area, and it is easy to make the identification program a "one point", which means that it will cause misjudgment.

(2) If the light happens to be reflected at the "six points (as shown in Figure 1 of Figure 1)", the distance between the dice is 4, and the distance between the two is small, which will make it impossible to discriminate the number of points.

(3) If the two dice 2 and 3 are in contact with each other in Figure 1, the distance between one of the six dice 3 may be judged as the other two dice 2 due to the shortened distance. In other words, it may be misjudged as "one point, three points". Five o'clock."

(4) As shown in Figure 1, one point dice 1, two dice 2, six o'clock dice 3, because a little dice 1 and two dice 2 have a chamfered area 5 reflection, and will be mistakenly judged as "two points ( The reflection of the chamfering area 5 is classified as the number of points of the dice 1), two points, six points, or misjudged as "two points, three points, six points".

(5) In other words, since the knowledge is only identified by the "single identification program", the rate of error is high.

Please refer to Figures 3 and 4 for connecting a camera to a computer. The computer has first, second, third and fourth identification programs. The four groups of identification programs are available for selection. The camera is aligned with the dice. The image is captured and the dice image can be transmitted to the host computer; the dice image is selected by the host computer to select three or more sets of identification programs for processing, analyzing, and judging the number of points, and the computer host array identification program is It is designed to convert images into grayscale maps, search contours, and filter out contours of dots. Each group has extinction, binarization, connected contour cutting, three-heart algorithm, distance algorithm and edge-calculation algorithm. The function of the mask can be used to eliminate the reflection of the gray scale diagram of the dice, to form a black and white contrast, to divide and analyze the contour of the dot, and to calculate the counting of each dot spacing, so as to achieve the effect of correctly determining the number of dice points.

1, 2, 3‧‧‧ 骰子

4‧‧‧ spacing

5‧‧‧Chamfering area

Figure 1, the diagram of the dice

Figure 2, a flow chart of the prior art

Figure 3, the flow chart of this creation

Figure 4, schematic diagram of the creation identification program group

First, the function of the multi-group identification program of this creation (as shown in Figure 4) is explained:

Gray-scale map: The color image of the dice photographed by the photo is converted into a gray scale map. The gray scale map is composed of a plurality of small black dots, and the black dot can be used as a digital signal comparison.

Extinction: The reflection of the chamfered area 5 (as shown in Fig. 1) of the dice, or other reflections, is eliminated, and the "reflective area" is prevented from being misidentified as the "point area" of the dice.

Binarization: The gray-scale image is given the effect of "black and white contrast". In other words, the overall surface of the gray-scale image is light gray, and the dice is confused. Therefore, it is necessary to strengthen the black-and-white contrast processing. The dots of the scorpion are black and the other areas are white.

Connected contour cutting: As shown in Figure 1, the dice 3, because the distance between the four dots of the dice 3 is smaller, it is necessary to cut the dots into separate points to avoid multiple chaos. Misjudgment.

Search for outlines: Search for the circular outline of each cut circle and calibrate the fixed point.

Filtering the outline of the dot: searching for a contour that is not a circle, for example, the reflection of the chamfered area 5, or other reflections, the reflective area is non-circular, and the non-circular outline is excluded from counting points. .

The three-hearted algorithm: the three-hearted system refers to the "center, center of gravity, and center of the circle". The distribution of 1~6 points of a scorpion can be determined by the geometrical calculation of the center, center of gravity, and center of the triangle.

Distance algorithm: When multiple pairs of dice are used, the distance between the dots is used as the basis for calculation. The distance beyond two points is another point.

Edge-finding algorithm: Based on the distance between the edge of the scorpion body and the dot, the distance beyond the proper one is a point.

Figure mask: combines the contour of the searched dot with the dice pattern originally created in the database to form a real circle.

This creation can be provided with four sets of recognition programs: as shown in Fig. 4, the first identification program has a gray scale map, extinction, binarization, connected contour cut, search contour, filter unless dot contour, three heart calculation The processing function of the method; the second identification program has a grayscale map, a edge-finding algorithm, a search contour, a filter function for filtering the outline of the dot, and a distance algorithm; the third recognition program has a grayscale map and a search contour , mask masking, binarization, connected contour cutting, filtering unless the dot contour, distance algorithm processing function; the fourth identification program has a grayscale map, search contour, graph mask, edge algorithm, filtering Unless the dot contour, the three-heart algorithm handles the function.

The role of this creation will be described in detail as follows: Please refer to Figure 3 again. The number of identification programs to be used in the computer host must be three or more sets, and the group to be used depends on the site. Depending on the brightness of the light, for example, the first, third and fourth identification programs are used for brighter light, and the second, third and fourth identification programs are used for darker light. Take the first, third and fourth identification programs and identify the dice in the first picture as an example: after the camera takes a dice image, the first recognition program converts the image into a grayscale image, and then extinction, binarization, and continuous contour cutting. , search for contours, filter unless the contour of the dot, the processing and analysis of the three-hearted algorithm; the third recognition program also converts the image into grayscale image, search contour, map mask, binarization, connected contour cutting, filtering The contour of the dot and the processing and analysis of the distance algorithm; the fourth recognition program also converts the image into a grayscale map, a search contour, a graph mask, a edge-measure algorithm, a filter unless the dot contour, and a three-heart algorithm are processed. Analysis The first, third, and fourth identification programs are processed and analyzed, and the number of points of the dice can be separately determined. At this time, the points determined by the three sets of identification programs are compared, and the majority of the points are used as a reference. For example, the first and fourth identification programs are judged as "9 points" and the third identification program is judged as "10 points". The "9 points" is the standard. In other words, it can be said that the first, third and fourth identification programs vote. The number of votes is based on the majority of the points considered. This is the unique idea of using three or more sets of recognition programs. The multiple identifications of the "majority identification program" are used to reduce the error rate of the points. The effect.

In summary, the control device of this creation is indeed an unprecedented creation, and has the effect of improving the correct rate of discriminating points. It is in line with the patent regulations, and please grant a patent.

1, 2, 3‧‧‧ 骰子

4‧‧‧ spacing

5‧‧‧Chamfering area

Claims (2)

A control device for reducing the number of false positives to identify a number of dice points, wherein the computer host has a camera, wherein the computer has a first identification program, a second identification program, a third identification program, and a fourth identification program. The identification program is available for selection; the camera captures the image of the dice and transmits the dice image to the host computer; the dice image is selected by the host computer according to the brightness of the scene, and the three or more groups are selected by the host computer. The identification program processes, analyzes, and determines the number of points. The host computer selects the number of points determined by the number of identification programs as the reference, and enhances the correct discrimination of the number of dice points by simultaneously identifying the plurality of identification programs. The control device for reducing the number of misidentification photography identification dice points according to the first aspect of the patent application, wherein the first identification program has a grayscale map, extinction, binarization, connected contour cutting, search contour, filtering unless the dot The processing function of the contour and the three-hearted algorithm; the second identification program has a grayscale map, a side algorithm, a search contour, a filtering function of the dot contour, and a distance algorithm, and the third recognition program has a graying function Step diagram, search contour, map mask, binarization, connected contour cut, filter unless the dot contour, distance algorithm processing function, the fourth identification program has a gray scale map, search contour, map mask, filter Unless the dot contour, the three-heart algorithm handles the function.