JP3820649B2 - Image processing apparatus and image processing method - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an image processing apparatus and an image processing method for accurately counting the number of processing objects in a state where a plurality of processing objects, for example, a plurality of cases containing compact discs are arranged. .
[0002]
[Prior art]
As a processing target for image processing, for example, a case for accommodating a compact disc will be described. A transparent plastic case that accommodates a compact disc (hereinafter referred to as a CD) contains a display label that displays the contents of information recorded on the CD. This label is made of paper or other materials, but the printing colors vary. In a manufacturing factory, a plurality of cases of a specific type are collected and packed in one box. It is necessary to count the quantity of these cases in this boxed state.
[0003]
Conventionally, the number of cases packaged in this way is counted visually by an operator or by the following electrical process. That is, as shown in FIG. 14, a plurality of cases C that are products housed in the box B are arranged in the Y direction. At this time, a plurality of cases C are imaged by the camera, and the brightness of the arranged cases C is obtained. FIG. 14A shows an input screen input to the camera, and FIG. 14B shows luminance fluctuation in the Y direction. The luminance changes along the Y direction, and the number of cases C is determined from the edge of this waveform.
[0004]
[Problems to be solved by the invention]
However, there may be a gap D between adjacent cases C and C. In addition, when the color of the label accommodated in the case C is gray, black, or amber, the difference between the luminance of the case C and the luminance of the gap D on the input screen captured by the camera. Has a problem that an error occurs in the count number of case C.
Accordingly, an object of the present invention is to provide an image processing apparatus and an image processing method capable of solving the above-described problems and reliably counting the number of processing objects in a state where a plurality of processing objects are arranged.
[0005]
[Means for Solving the Problems]
In the present invention, the above object is a plurality of Case In an array of these multiple Case Is an image processing device for counting the number of Case An imaging device that obtains a plurality of images, and a plurality of images obtained by the imaging device Case Images of Case For each of a plurality of block images divided in a second direction intersecting the first direction From one given case Create a master image and create a predetermined image within each block image. Case Compare the amount of luminance dispersion from the predetermined amount according to the amount of dispersion block Select an image and select multiple images within the selected block image. Case Each image in the master image and block image Case The amount of correlation with the image is calculated, and based on the obtained amount of correlation, a plurality of Case This is achieved by an image processing device comprising a counting means for counting the number of
[0006]
In the present invention, the imaging device includes a plurality of Case Get the image. The counting means includes a plurality of counts obtained by the imaging device. Case Images of Case For each of a plurality of block images divided in a second direction intersecting the first direction From one given case Create a master image. The counting means is a predetermined unit in each block image. Case The luminance dispersion amount is compared, a predetermined block image is selected according to the dispersion amount, and a plurality of blocks in the selected block image are selected. Case Are sequentially compared with the master image.
As a result, the counting means is configured so that each of the master image and the block image Case The amount of correlation with the image of Case We will count the number of
In this case, Case If there is a gap between Case Even if it is difficult to distinguish between Case The number of can be counted.
[0007]
DETAILED DESCRIPTION OF THE INVENTION
DESCRIPTION OF EXEMPLARY EMBODIMENTS Hereinafter, preferred embodiments of the invention will be described in detail with reference to the accompanying drawings.
The embodiment described below is a preferred specific example of the present invention, and thus various technically preferable limitations are given. However, the scope of the present invention is particularly limited in the following description. Unless otherwise stated, the present invention is not limited to these forms.
[0008]
FIG. 1 shows an example of an image processing apparatus 10 according to the present invention and a case to be processed. The image processing apparatus 10 includes a camera 12 as an imaging apparatus and a counting unit 14.
As the camera 12, for example, a CCD (charge coupled device camera, solid-state imaging device camera) can be adopted.
Cases C are arranged in the Y direction, and the plurality of cases C constitute a case group CG. The case group CG is accommodated in a box B for accommodating. This box B is used in a manufacturing factory for manufacturing a CD and storing it in a case C.
The camera 12 images the case group CG and the box B and sends an image signal GS to the counting means 14.
In the image processing apparatus 10, the counting means 14 accurately counts the number of cases C constituting the case group CG based on the image signal GS obtained by the camera 12.
[0009]
FIG. 2A illustrates an example of luminance distribution in the Y direction when the camera 12 images the case group CG of the case C in the box B. In this case, a convex portion 16 of luminance is formed corresponding to each case C. However, simply taking such a luminance distribution, when the luminance of the case C and the luminance of the gap D formed between the case C and the case C are substantially the same, the gap D is also one. There is a possibility that the case C is counted and an error occurs in the count number of case C.
The embodiment of the present invention prevents such an error in counting the quantity of the case C in the case C.
[0010]
FIG. 3 shows a detailed configuration of the image processing apparatus 10 shown in FIG.
As described above, the image processing apparatus 10 includes the general camera 12 and the counting unit 14 and further includes other elements.
The counting means 14 is disposed in the image processing board 22, and the image processing board 22 includes the counting means 14 and a D / A converter (digital / analog converter) 27. The image processing board 22 belongs to the main body 30 of the image processing apparatus 10, and the main body 30 includes an image processing board 22, a host CPU (central processing unit) 31, and a program memory 40.
[0011]
An image signal GS (analog signal) of the camera 12 is input to an A / D converter (analog / digital converter) 24 of the counting means 14 in the image processing board 22. The counting means 14 includes a converter 24, an image memory 15A, an image memory 15B, an image processor 16, and an image CPU (central processing unit) 18.
The image signal GSD converted from an analog signal to a digital signal by the A / D converter 24 is sent to the image memory 15A. The image signal GSD is a video signal obtained by imaging the case group CG and the box B in FIG. 1, and the image memory 15A stores the image signal GSD.
[0012]
The image memory 15 </ b> A is connected to the D / A converter 27 and the image processor 16. Therefore, the image signal GSD recorded in the image memory 15A can be digital / analog converted by the D / A converter 27 and displayed on the image processing monitor 9. That is, the image processing monitor 9 displays an image (see FIG. 2) viewed from above the case group CG and the box B.
The image signal GSD stored in the image memory 15A can be sent to the image processor 16.
For example, as illustrated in FIG. 11, the image processor 16 has a second direction (Y direction) that intersects the first direction (Y direction) in which the case C (processing target) in the image signal GSD stored in the image memory 15 </ b> A is arranged. A plurality of block images (L1 to L6) are created according to the X direction), and master images (MR1 to MR6) are obtained from predetermined processing objects in each block image. The image memory 15B stores master images (MR1 to MR6) in the respective block images (L1 to L6).
As a result, the image processor 16 sequentially compares the images of the cases C of the case group CG and the master images (MR1 to MR6) in the block images (L1 to L6) stored in the image memory 15A.
[0013]
The image CPU 18 obtains the amount of correlation between the master image (MR1 to MR6) and the image of each case C in the block images (L1 to L6), and calculates the number of processing targets based on the obtained amount of correlation. Count.
The image CPU 18 has a function of displaying the operation performed by the image processor 16 on the image processing monitor 9 through the D / A converter 27. The image processing monitor 9 has a correlation value as shown in FIG. Display the graph.
[0014]
The host CPU 31 is connected to a program memory 40, a program monitor 42, a mouse 43, which is a pointing device, and a keyboard 44. The host CPU 31 transmits a program in the program memory 40 to the image CPU 18 and transmits a user command. In addition to the program, the user command is data such as a standard value and an inspection area, and a program pause command. The host CPU 31 can receive processing data such as the result of counting the number of cases C as described above from the image CPU, and can print out the data with the printer 31a.
The program monitor 42 is used for display for setting standard values and inspection areas, and for displaying detailed data (dispersion values, correlation values) for each block and each point.
[0015]
The keyboard 44 is used by the user to set standard values. In order to set the examination area, the mouse 43 specifically clicks the start point (upper left point) of the examination area, and then clicks the end point (lower right point).
[0016]
Next, referring to FIG. 10 as an example, a capturing method in which the camera 12 of FIG. 1 captures an image of the case group CG will be described.
In FIG. 10, in the box B, a total of nine cases C and one other case C are arranged in the Y direction. That is, the rectangular portion of the case C corresponds to the back C1 portion of the case C as shown in FIG. When the lid C2 of the cassette C is opened in the direction of the arrow, the label R can be accommodated. The case C is made of a transparent plastic, accommodates a label R, and can accommodate a CD. The label R has the display part PR1 positioned inside the back C1 of the case C, so that the contents of information (title and the like) recorded on the internal CD can be displayed by any expression method such as characters and colors.
[0017]
Returning to FIG. 10, the label R1 is stored in the nine cases C of the arranged cases C, but inside the other case C (which is substantially located at the center). Contains a label R2. The label R2 and the label R1 indicate different display contents. In addition, gaps B1, B2, and B3 exist between the box B and the case C.
[0018]
The image signal GS of the case group CG captured by the camera 12 in FIG. 1 is stored in the image memory 15A in FIG. , And is captured by the image processor 16.
The image processor 16 divides the image into block images L1 to L6 as illustrated in FIG. Moreover, the image processor 16 arbitrarily sets the master images MR1, MR2, MR3, MR4, MR5, MR6 in each of the block images L1 to L6. As described above, the master images MR1 to MR6 are stored in the image memory 15B.
For example, in the block image L1, the image processor 16 sequentially compares the master image MR1 and the image of each case C, and determines the correlation amount between the master image MR1 and the image of each case C in the block image L1 as Y. The number of cases C in the block image L1 is counted based on the obtained correlation amount sequentially.
[0019]
Here, the correlation amount between the master image and the case image in the block image will be described with reference to FIGS.
FIG. 4 shows an example in which the luminance correlation is taken when the master image moves from the left case C to the right case C. A black circle marker CM is formed on the right case C and the left case C. This marker CM is formed on the display portion R1 of the label R in FIG.
[0020]
In the case C on the left side of FIG. 4, when the area of the master image is sequentially moved in the Y direction, it moves as shown in FIGS. 5 (A) to 5 (K). In this way, in FIG. 5A, the correlation amount regarding the marker CM in the case C on the left side is maximized, and the regions are sequentially shifted by one unit as shown in FIGS. 5B to 5J. The amount of correlation changes. As shown in FIG. 5K, the correlation amount is maximized again for the mark CM in the right case.
The expression of the correlation amount (coefficient) is shown in FIG. If the reference case C is a reference region and the target case C is a partial region, N data Mi of the reference region starting from (X1, Y1) and (U, V) are the starting points. When comparing a partial area Ii composed of N pieces of data, the correlation amount (correlation coefficient) R (U, V) can be expressed as shown in FIG.
[0021]
Next, the image processing method will be described with reference to FIGS. 7, 8 and 9 to 12.
FIG. 7 shows that the processing method to be performed later differs depending on whether or not the case C of the case group CG as shown in FIG.
FIG. 8 shows the contents of the normal process SS in FIG.
FIG. 9 shows an example of a product in the case where the feature appears as a whole product (case C) and an example of a product in which no feature appears as the whole product.
In FIG. 9A, as an example of a product in which features as a whole product appear, the label R1 of the case C includes the title name TR and the already described black circle mark CM.
[0022]
On the other hand, in FIGS. 9 (B-1) to (B-3), as an example of a product in which no feature appears as a whole product, one mark CM is a case C as shown in FIG. 9 (B-1). This is the case where it is formed over the entire (label R1) or the case where a plurality of marks CM are formed over the entire case C (label R1) as shown in FIG. 9B-2. As another example, as shown in FIG. 9B-2, a color having the same brightness level as the gaps B1 to B3 as shown in FIG. 12 is applied over the entire case C (label R1). There is an example in which the portion TS and the extremely small title name TR1 are included.
[0023]
FIG. 10 shows an example in which the case C (label R1) of FIG.
FIG. 11 shows an example in which the case C (label R1) of FIG.
FIG. 12 shows an example in which a case C (label R1) in which a plurality of black circle marks CM of FIG.
FIG. 13 shows an example in which the case C (label R1) having the part TS and the title name TR1 in FIG.
[0024]
In FIG. 7, in step ST <b> 1, it is determined whether or not a feature appears as the entire product (case C). This determination is made by the operator. As a result, when the characteristics of the entire product appear, the process proceeds to step ST2, and a master image MR1 (master waveform) of the entire product (case C) illustrated in FIG. 10 is taken. The above-mentioned master image MR1 of the entire product is, for example, the entire image of the case C at the extreme end. This master image MR1 is stored in the image memory 15B of FIG. Then, the process proceeds to the normal process SS.
Since the label R1 of the case C in FIG. 10 has the mark CM and the title name TR, the overall luminance of the label R1 is an example in which there is a large difference from the luminance of the gaps B1, B2, and B3.
The normal process SS is shown in FIG. 8, and first, the camera 12 of FIG. 1 captures an image of the case group CG (step SS1). Then, for example, the image processor 16 in FIG. 3 extracts the start edge SE1 of the case C at the first end (step SS2).
[0025]
Next, the start case C at the end is correlated (step SS3). The image processor 16 compares the master image MR1 with the entire image of each case C in the image memory 15A by shifting the master image MR1 in the Y direction, and the image CPU 18 follows the start case C along the Y direction. The correlations of the images of the plurality of cases C arranged in this way are calculated (steps SS4 and SS5).
[0026]
When image comparison is performed from the start case C to the last case C (step SS6), the image CPU 18 in FIG. 3 detects a change in the correlation value with respect to the position in the Y direction as shown in step SS7.
If there is a change in the correlation value, the image CPU 18 counts the appearance of the correlation values CC1 to CC9 exceeding the predetermined threshold value in FIG. 10 as shown in steps SS8 and SS9, and ends the counting operation. To do. Otherwise, if there is no change in step SS8, the process ends without counting as in step SS10.
In this way, the image processor 16 compares the master image of the entire case C with the image of each case C by the step ST2 of FIG. 7 and the normal processing SS of FIG. 8, and the image CPU 18 compares these correlations. By taking and counting, the quantity of case C can be counted reliably.
[0027]
Next, in step ST1 of FIG. 7, in the case where no feature appears as a whole of case C, that is, in the case of the example as shown in FIGS. 9B-1, B-2, and B-3, The process after step ST3 of FIG. 7 is performed.
9B-1), B-2, and B-3 are different from the case where the entire luminance of each case C and the luminance of the gaps B1, B2, and B3 are different. This is a case where almost no occurs.
[0028]
That is, as shown in FIGS. 11 to 13, the image processor 16 of FIG. 3 divides the case group CG into, for example, block images L1 to L6 (step ST3). Then, master images (MR1 to MR6) are taken for each block image (L1 to L6) (step ST4). In the example of FIG. 11, the master image MR1 is obtained from one case C of the block image L1, the master image MR2 is obtained from the next block image L2, and the master image MR3 is obtained from the block image L3. Thus, master images MR1 to MR6 can be obtained from the block images L1 to L6, respectively. These master images MR1 are stored in the image memory 15B by the image processor 16.
[0029]
The image processor 16 in FIG. 3 compares the luminance dispersion amounts in the block images L1 to L6 as shown in FIGS. 11 to 13 to determine whether there is a block image having an extremely low dispersion amount. (Step ST5).
The luminance dispersion amount is an amount represented by an expression as shown in FIG. 6 and is an amount representing the degree of variation in the data in the region.
For example, in the case of FIG. 11, when the amount of dispersion between the block images L1 to L6 is compared, there is a block image with an extremely low amount of dispersion, so the process proceeds to steps ST5 and ST8. That is, in the case of FIG. 11, as shown in FIGS. 11B and 11C, the block images L <b> 2 to L <b> 6 have extremely small luminance dispersion amounts compared to the block image L <b> 1. On the other hand, in the examples of FIGS. 12 and 13, since it is considered that there is no extremely low block image with respect to the amount of dispersion between the block images L1 to L6, the process proceeds from step ST5 to ST6.
[0030]
First, in the example of FIG. 11, the process proceeds from step ST5 to ST8, and the image CPU 18 determines whether or not there is a difference in the amount of dispersion between block images. In the example of FIG. 11, there is a difference in the amount of luminance dispersion between block images, that is, the block image L1 has a black circle mark CM. Therefore, unlike the other block images L2 to L6, The image L1 is a block image with a high luminance dispersion amount. Accordingly, in step ST10, one block image L1 having a high dispersion amount is selected, and the process proceeds to the normal process SS of FIG.
[0031]
That is, by capturing an image in step SS1 of FIG. 8, block images L1 to L6 are stored in the image memory 15A, and the image processor 16 stores the master image MR1 and the like of the block image L1 in the image memory 15B. Then, the image processor 16 extracts the start edge SE of FIG. 11 in step SS2 of FIG. 8, and correlates the start case C with the image of the master image MR1 in step SS3. Then, the image processor 16 sequentially compares the image corresponding to each case C of the block image L1 and the master image MR1, that is, sequentially shifts the master image MR1 in the Y direction, thereby matching the image of the corresponding case C and the master image MR1. The correlation of the image MR1 is calculated (steps SS4 and SS5).
[0032]
When the comparison of the images in the Y direction is completed (step SS6), the image CPU 18 detects a change in the correlation value in the Y direction as shown in FIG. If the correlation value is changed as shown in FIG. 11B, the image CPU 18 counts (steps SS8 and SS9). As an example of this count, the count values CC1, CC2, CC3,..., CC8 are counted as shown in FIG.
[0033]
If the image CPU 18 does not detect a change in the correlation value in step SS10 in the example of FIG. 12, the positions B1, B2, B3 of the gaps B1, B2, and B3 and the different types of cases C (label R2) in FIG. It does not count at P2, P3, P4 (step SS10).
In the example of FIG. 11, since the label R2 of the sixth case C from the right is different from the label R1 of the other case C and has the mark CM1, the correlation value does not reach a certain threshold and is not counted. Also, the gaps B1, B2, and B3 are not counted because the correlation value does not reach the threshold value.
Therefore, the number of a total of nine cases C (label R1) can be correctly counted except for different types of cases C (label R2).
[0034]
Next, returning to step ST8 in FIG. 7, a case is considered in which there is no difference in the luminance dispersion amount between the block images L1 to L6 as shown in FIG. 12 and FIG.
In this case, the process proceeds to step ST9, normal process SS, and step ST7. As this example, an example as shown in FIG. 12 can be considered. In this case, for example, all the block images L1 to L6 are counted. Thereafter, the normal process SS is performed in the manner as described above, and the most frequently counted number in the block images L1 to L6 is adopted as the count value of case C of the case group CG.
[0035]
Next, in step ST5 of FIG. 7, if there is a block image with an extremely low dispersion amount, steps ST6 and ST7 are performed. As an example of this, the example of FIG. 13 can be considered. In this example, the block images L1, L2, L5, and L6 have extremely low luminance dispersion amounts, and several block images with relatively high dispersion amounts are selected. For example, in FIG. 13, the block images L3 and L4 are selected, the normal process SS of FIG. 8 is performed, and the most frequently counted number is adopted as the count value of case C of the case group CG.
As described above, even if there is a gap between the cases as in the labels R1 and R2 of the case C, the number of cases can be reliably counted.
[0036]
The method of counting the number of cases C using the correlation amount in this way is similar to the way that a person counts by relying on the design of the back (back) of the case C. By observing the change in the correlation amount, the number of cases can be counted reliably in the same manner as a human being. In other words, by calculating the amount of correlation between the master image of the specified area (also referred to as the master area waveform) and the block image (also referred to as the block area waveform) taken in advance, the work is performed along the Y direction. The change of the correlation value is graphed (schematically) as shown in FIGS.
[0037]
Then, the peak position of the correlation value (for example, the count positions CC1, CC2,... In FIG. 11) is obtained while considering the width of the case C in the Y direction, and the image CPU 18 in FIG. . Or it counts using the periodicity of the waveform.
Even if the gaps between the cases C are not uniform and the sizes of the gaps of the product rods are different, the gaps are not counted, for example, as shown in FIG. Can be counted reliably.
[0038]
By the way, the present invention is not limited to the above embodiment.
In the above-described embodiment, the description has been given using the case C that accommodates a compact disc (CD) and a label for displaying the contents. However, the present invention is not limited to this, and other types of cases such as a diameter of 8 cm are used. It can also be used for counting cases used for CDs, mini-discs (MD), other magneto-optical discs, floppy discs, and other disc-shaped high-density recording media.
In addition to such fields, cases in other fields, such as cassette tape cases, or cases in other fields, are also used to count products of other types. It is possible to apply.
[0039]
【The invention's effect】
As described above, according to the present invention, each of the master image and the block image Case The amount of correlation with the image of Case With these arranged Case Can be counted reliably.
[Brief description of the drawings]
FIG. 1 is a diagram showing a preferred embodiment and an object to be processed of an image processing apparatus according to the present invention.
2 is a diagram illustrating an example of an image obtained by the image processing apparatus of FIG. 1 and its luminance.
FIG. 3 is a diagram showing a detailed configuration of the image processing apparatus in FIG. 1;
FIG. 4 is a diagram for explaining movement of a region for obtaining a correlation between luminances of a master image and a target image.
FIG. 5 is a diagram showing a change in correlation amount when the region in FIG. 4 is moved.
FIG. 6 is a diagram illustrating an example of an expression of a correlation amount (correlation coefficient).
FIG. 7 is a flowchart showing a preferred embodiment of an image processing method according to the present invention.
FIG. 8 is a flowchart showing an example of normal processing in FIG.
FIG. 9 is a diagram showing an example of a product.
FIG. 10 is a diagram illustrating a state in which product examples are arranged, and an example of output of correlation values in the arrangement of the product examples.
FIG. 11 is a diagram illustrating a state in which product examples are arranged, and an example of outputting correlation values in the arrangement of the product examples.
FIG. 12 is a diagram illustrating a state in which product examples are arranged and an example of output of correlation values in the arrangement of the product examples.
FIG. 13 is a diagram illustrating a state in which product examples are arranged, and an example of output of correlation values in the arrangement of the product examples.
FIG. 14 is a diagram illustrating an example of conventional image processing.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 10 ... Image processing apparatus, 12 ... Camera (imaging apparatus), 14 ... Count means, L1-L6 ... Block image, MR ... Master image, C ... Case (product processing object) ), R1, R2 ... label, CG ... case group (processing object), B ... box

Claims (4)

An image processing device for counting the number of cases in a state where a plurality of cases are arranged,
An imaging device for obtaining images of a plurality of cases ;
A plurality of case images obtained by the imaging device are divided into a second direction that intersects the first direction in which the cases are arranged, and a master image is created from a predetermined one case for each of a plurality of block images. Compare the amount of luminance dispersion from a given case in the block image, select a given block image according to the amount of dispersion, and sequentially compare the images of multiple cases and the master image in the selected block image An image processing apparatus comprising: a count unit that obtains a correlation amount between a master image and an image of each case in a block image, and counts the number of a plurality of cases based on the obtained correlation amount.   The image processing apparatus according to claim 1, wherein when the correlation amount in each block image is larger than a predetermined threshold, the number of processing targets is counted.   The image processing apparatus according to claim 1, further comprising: a display device that displays a plurality of images to be processed, a plurality of block images, and a graph representing a correlation amount between each of the image to be processed in the master image and the block image. An image processing method for counting the number of multiple cases in a state where multiple cases are arranged,
Get multiple case images,
Images of a plurality of cases, to create a master image from a predetermined one case for each of a plurality of block image is divided in a second direction the case intersects the first direction are arranged,
Compares the amount of luminance dispersion from a given case within each block image, selects a given block image according to the amount of dispersion, and sequentially compares the images of multiple cases within the selected block image with the master image. The amount of correlation between the master image and the image of each case in the block image is obtained,
Based on the obtained correlation amount, count the number of multiple cases ,
An image processing method.

Images