JPH0465429B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a method of cutting out key tops, which is particularly suitable for use in a keyboard inspection device.

[Conventional technology]

Recently, keyboard inspection devices (keyboard checkers) that automatically inspect the key tops arranged on a keyboard have begun to appear. this is,
For example, a plurality of key tops are imaged by a predetermined image pickup device, and the image signals are processed by a predetermined processing device to check whether the key tops are mounted or not. In this case, in order to extract individual keytops from the captured keyboard image, it is necessary to determine the binarization level for this purpose, but conventionally this work has been done exclusively manually. In other words, the operator manually determines the cutting level for each keytop while looking at the keyboard monitor image. The results determined in this way are saved as a "settings file" and used as a reference for subsequent checks.

[Problem that the invention seeks to solve]

However, if the above-mentioned work is performed manually, it is not only complicated and time-consuming, but also has problems inherent to manual methods in terms of accuracy and reliability, so automation is strongly desired. .

[Means for solving problems]

This invention measures the area ratio histogram of the grayscale image when capturing and processing a plurality of keytops, and determines a threshold value for cutting out the keytop surface from the background based on this histogram. The aim is to automate the process and eliminate the drawbacks caused by the manual process described above.

[Effect]

FIG. 1 is a main diagram that best represents the features of this invention. Note that A in the figure shows the relationship between an ideal key top and its image signal (video signal), and B in the figure is an area ratio histogram corresponding to A.

Assume now that the keytop image obtained through an imaging device such as a television camera is ideal, as shown in FIG. In other words, the image of the keyboard is divided into four parts (regions): the top surface A of the keytops, the side surface B of the keytops, the border C between keytops, and the character D displayed on the keytops, and the gray value (white or black level) of each part is )E are substantially constant and take the values of T _BK , T _BS , T _BB , and T _BP , respectively. That is, here, D, A, B, and C are bright in this order (T _BP > T _BK > T _BS > T _BB ), and therefore the character D is a white character (bright character; bright character). Furthermore, when the character D is a black character (dark character; dark character), the shading value T _BP of the character D is smaller than T _BK , and the relationship between these shading values E varies depending on the type of keyboard or key top. Although they are different, it has been confirmed through numerous experiments that at least the relationship T _BK > T _BS > T _BB exists, that is, the top surface A of the key top is brighter than its side B, and the side B is brighter than its boundary C. It is being

Now, consider a predetermined image observation area (window) with a certain width W _D and height _HD , and if the ratio of the area S representing the key top surface A occupying within this area is the area ratio P _A , then this The area ratio P _A is expressed as P _A = 100 × S / W _D × H _D [%] ...(1), and this value is the
It can be seen that it changes depending on the value conversion level (here, 256 levels from 0 to 255 are selectable).
In other words, if the binarization level is small, the area ratio P _A will be large, and if the binarization level is large, the area ratio P _A
becomes smaller. Therefore, for example, for an image represented as shown in Figure 1A, the relationship between the area ratio P _A occupied within a predetermined area representing the key top surface and the binarization threshold is as shown in Figure 1B. , hereinafter this will be referred to as an area ratio histogram. Note that P _AP , P _AK , P _AS ,
P _AB indicates the area percentage of each portion D, A, B, and C in A of the figure, respectively.

The area ratio histogram obtained in this way has a slight difference for each keytop image, but it is known that it generally has the same tendency as shown in Figure A. Therefore, if such an area ratio histogram can be obtained, it is possible to The basic concept of the present invention is that the binarization level for cutting out the keytop surface from the background can be determined using the method.

〔Example〕

FIG. 2 is a block diagram showing a keyboard inspection device to which the present invention is applied. In the same figure, 1
is a processing device, 2 is an imaging device such as a television camera,
3 is a TV monitor, 4 is a keyboard to be inspected,
5 is an XY table, and 6 is an XY table controller. The processing device 1 includes a master processor unit 7, a CRT display 9, a man-machine interface 8 consisting of a keyboard 10, a floppy disk 11, a printer interface 12, an image input control processor 13, image processing processors 14 and 15, and an XY table control. Interface 1
6, binarization threshold generation circuit 17, binarization section 1
8, a monochrome control section 19, a window generation section 20, a feature extraction section 21, image memories 22, 23, and the like. Here, the master processor 7 gives predetermined commands to the image input control processor 13 and image processing processors 14 and 15, which are slave processors, to perform overall sequence control. Two image processing processors are used to improve processing speed, but one can be omitted in some cases.

The operation will be briefly explained below.

First, when the keyboard 4 to be inspected is placed on the XY table 5, the master processor 7
gives a movement command to the controller 6 through the XY table control interface 16. As a result, the XY table 5 operates, the keyboard 4 is moved to a predetermined position, and the television camera 2 images the key top surface. In this case,
It goes without saying that the television camera 2 may be moved instead of the keyboard 4. Next, the master processor 7 specifies a binarization threshold value for the image input control processor 13, so the processor 13 transfers this to a binarization threshold generation circuit (which generates a 256-level threshold value). ) 17 to generate the binarization threshold as instructed. The binarization unit 18 includes a television camera 2
The image signal from is binarized based on this binarization threshold. Here, the bright part (white part) of the image
A command as to whether to import the image or the dark part (black part) is issued from the master processor 7, and is applied to the monochrome control section 19 via the processor 13, where a predetermined operation such as inversion is performed. The window generation unit 20 determines the area (window) in which the binary image should be captured, and the feature extraction unit 21 compresses the binary image data within the specified area and transfers it to the image memories 22 and 23 via DMA ( Write in direct memory access) mode. Incidentally, such image input control is performed by the processor 13, and commands for this purpose are given from the master processor 7. Further, the master processor 7 issues pattern calculation processing commands to the image processing processors 14 and 15. As a result, the processor 14
Alternatively, step 15 creates an area ratio histogram as described above, and uses this to determine a threshold for keytop extraction. It goes without saying that in order to create this histogram, it is necessary to obtain the area ratio for each binarization level. Further, a man-machine interface 8 consisting of a CRT display 9, a keyboard 10, etc. is provided for the operator to interact with the keyboard inspection device, and a floppy disk 11, printer interface 12, etc. are provided as input/output devices. ing.

Next, a method for cutting out key tops according to the present invention will be explained below.

Now, assuming that the area ratio histogram obtained by the image processing processor 14 or 15 is expressed as shown in FIG. 3, by analogy with the histogram in FIG. In other words, the threshold values T ₁ and T ₂ points can be considered to correspond to the previous T _BS and T _BB , respectively. Here, we only want to cut out the top part of the key top, so the point we need to find is T _BK in Figure 1 (b).
This point corresponds to . Therefore, since the area ratio P _AK of the top surface of the key top is usually considered to be greater than P _AKL (20 to 40%), first, find the minimum point T ₀ of the threshold value that is greater than P _AKL (for example, 30%). Determine from the histogram.

On the other hand, the threshold T ₁ is expressed by the following equation: L _A (T _i )
This value is considered as the point at which the difference of the second-order differential coefficient is greater than or equal to a predetermined value L _AL (≈0.1 to 1.0%), and this is determined by trial and error from the histogram.

L _A (T _i )=P _A (T _i +ΔT _o ) −2P _A (T _i )+P _A (T _i −ΔT _o ) ……(2) Here, T _i is the threshold value (=0 to 255 ) and
ΔT _o represents the threshold interval (≈5 to 20).

The points corresponding to T _BK above are these T ₀ ,
Since it is considered to be between _T1 points, a value T _BKE that internally divides the distance between these points into a predetermined ratio n _K1 :n _K2 is determined as shown in the following equation.

T _BKE =n _K2・T ₀ +n _K1・T ₁ /n _K1 +n _K2 ...(3) The threshold value T _BKE shown in equation (3) is the cutting level for separating the keytop surface from the background. It turns out that.

Therefore, assuming that an example keyboard image is obtained as shown in FIG. 4, a pattern cut out from this at a binarization level of T _BKE is expressed as shown in FIG. 5, for example. Here, the left point, right point, apex of the circumscribing frame circumscribing each pattern as shown in FIG.
Let the coordinates of the bottom point be X _WL , X _WR , Y _WT , Y _WB , and the width and height of the pattern be W _W and H _W , respectively.
A relationship as shown in the following equation is established between these.

W _W =X _WR −X _WL +1 …(4) H _W =Y _WB −Y _WT +1 …(5) The reason for adding +1 here is that if the left point and right point coordinates are equal to each other ( This is to avoid the width and height of the pattern expressed by equations 4) and (5) becoming zero. Here, let W _WL and H _WL be the minimum values of the key top outline, and consider the key top pattern to be one that satisfies the following equation.

W _W ≧W _WL , H _W ≧H _WL (6) In other words, patterns that do not satisfy the relationship of equation (6) are ignored as not key tops. In this way, those that are only half visible in Figure 5 are eliminated by this equation (6), so in the end, 4 keytops will be cut out for Figure 4 as shown in Figure 5. .

In addition, according to the present invention, even when a plurality of keytops of different colors exist in one screen, the keytops can be extracted by repeating the above method. In addition, in the above, a rectangular window was used to extract the keytops, but by using a window that fills the outermost part of the pattern obtained with the extraction threshold, the keytops found in the first The area ratio histogram is created again for the other items, and the keytop extraction threshold is determined based on this to determine the second keytop window. By repeating such operations, it is possible to extract keytops even if there are keytops with different gradation values in one screen. In addition, in the above, the key top surface was cut out using the points T ₀ and T ₁ , but the third
Point T ₂ in the figure (corresponding to T _BB in Figure 1 B) is connected to point T ₁
By calculating in the same way as in the case of , and internally dividing the area between points T ₁ and T ₂ at a predetermined ratio, it is possible to cut out the side surface of the key top (see part A in Figure 1 or part B in Figure 4). can.

〔effect〕

According to this invention, by creating an area ratio histogram, the binarization level for keytop extraction can be automatically determined.
This has the advantage of significantly reducing manual work and improving reliability. As a result, a highly accurate "configuration file" can be created in a short time.

[Brief explanation of the drawing]

Fig. 1 is a main diagram that best represents the features of this invention, Fig. 2 is a block diagram showing a keyboard inspection device to which this invention is applied, and Fig. 3 is an area ratio histogram for explaining an embodiment of this invention. , 4th
The figure is a reference diagram showing an example of a keyboard image, and FIG. 5 is a reference diagram for explaining a key top cutting pattern. Description of symbols, 1...Processing device, 2...Imaging device (television camera), 3...Monitor TV, 4...Keyboard for inspection, 5...XY table, 6...
XY table controller, 7... Master processor unit, 8... Man-machine interface, 9... CRT display, 10... Keyboard, 11... Floppy disk, 12... Printer interface, 13... Image input Control processor, 14, 15... Image processing processor,
16...XY table control interface, 17
... Binarization threshold generation circuit, 18 ... Binarization section, 19 ... Monochrome control section, 20 ... Window generation section, 21 ... Feature extraction section, 22, 23 ... Image memory, A ... Keytop top surface, B...keytop side, C...border between keytops, D...letter displayed on keytop surface, E...shading value.

Claims (1)

[Claims] 1 A cutting method in which multiple keytops on a keyboard are imaged with an imaging device, and the image signals are processed by a processing device to cut out the keytop surface from the background, and a predetermined image observation area is set and the binarization level is For each image binarized according to the level, the area ratio of the area value representing the key top surface to the observation area is calculated to create an area ratio histogram. The first area where this area ratio is greater than or equal to a predetermined value
After determining the point and the second point at which the rate of change is greater than or equal to a predetermined value, the binarization level is determined from the point that internally divides these two points at a predetermined rate, and the determined 2 points are determined. A method for cutting out key tops, characterized in that each key top is cut out based on the digitization level.