JPH0793553A - Printing and inspecting device for ink jet printer - Google Patents

Abstract

PURPOSE:To excellently perform the quality inspection of a printing character string including the fluctuation components in a time base direction which is printed by an ink jet system for a moving work which is being carried. CONSTITUTION:A printing character string is fetched as data in a printing character input part 10 and this data is converted into the feature amount vector data in a printing direction (time base direction) in a feature amount conversion part 11. The DP matching of the first stage is performed by this feature amount vector data and the feature amount vector data of a reference character string in a character delimiter detection part 12 and the delimiter detection of an inputted character string is performed, normalizing the time in the time base direction. Further, a printing character quality inspection is performed in a one-character internal inspection part 13 every delimited and detected character, using the DP matching processing of the second stage. Therefore, because the time normalizing function to be the function of the DP matching is used, the inspection of the printing character string including time base fluctuation can be exactly performed.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an ink jet printer print inspection device, and more particularly to an ink jet printer print inspection device for inspecting the quality of a print character string printed by an ink jet system.

[0002]

2. Description of the Related Art In the conventional print inspection of this kind, it is customary to inspect the print quality by using a general image processing method. In other words, printing is performed by the correlation matching method that performs simple correlation matching processing with the reference character pattern for each print character, or the method that extracts individual dots that make up the print character and performs matching processing with the dot pattern of the reference data. We are inspecting the quality.

[0003]

With such an ink jet method, for example, when printing is performed on a work such as a beverage can case sequentially conveyed by a conveyor, the characteristics of the ink jet printer printing method are as follows. The change in the time axis direction, that is, the deformation of the print with respect to the movement direction of the work, is particularly large, and individual dots also flow in the work movement direction, and the joining state between adjacent characters tends to occur frequently.

However, since the above-mentioned problems caused by such movement of the work are not taken into consideration in the above-described printing quality inspection method, the character cutting processing can be performed by dividing the characters. It is difficult, and it is difficult to extract each dot unless the individual dots are separated from each other.Therefore, in the above-mentioned conventional print inspection method, it is possible to inspect the print result on a moving workpiece with high reliability and high quality. It has the drawback that it cannot be done.

An object of the present invention is to provide a print inspection device for an ink jet printer, which can inspect a print result on a moving work with high reliability and high quality.

[0006]

A print inspection device for an ink jet printer according to the present invention comprises a reading means for reading a print character string by an ink jet system and a dynamic programming matching (hereinafter referred to as a dynamic programming matching method) for reading data by the reading means. , DP matching) method, and character delimiter detection means for performing delimiter detection processing in character units.

Another ink jet printer print inspection apparatus according to the present invention is further characterized by further comprising inspection means for inspecting the delimiter detection characters by the character delimitation detection means by a DP matching method.

[0008]

In the present invention, a continuous print character string consisting of a combination of dots by the ink jet method is taken in as dot pattern data, and individual character dot configurations are set in advance as reference patterns. Are sequentially converted to the feature amount in the printing direction, and the character feature delimiter detection processing is performed by using the converted feature amount by the DP matching method while eliminating the error of the time axis fluctuation, and the reference pattern is also applied to each character unit after the delimiter detection. It is also configured to detect a match / mismatch by comparing with.

[0009]

Embodiments of the present invention will be described in detail below with reference to the drawings.

FIG. 1 shows a functional block of a print inspection apparatus for an ink jet printer according to the present invention as a processing flow.

In step 10, a continuous print character string consisting of a combination of dots is fetched as image data, for example,
A printed character input unit such as a video camera is shown, and the state of dots, which are the constituent elements of each character, is replaced with binary data for output. This dot pattern is, for example, the dot pattern shown in FIG. 3 (in the figure, the dot pattern of "8" is shown).
Is.

Histogram data of the dot printing amount in the height direction of the dot pattern (vertical direction in the drawing in FIG. 3) is created, and the height of the character to be processed is detected.

In step 11, the data obtained by the print character input unit 11 is used as dot presence / absence information (1 or 0) or print level in the print direction (in FIG. 3, the horizontal direction of the drawing and the time axis direction). Dot density information (0
(Up to 255 steps) are converted into vector data formed by assembling the number of dots, which is output as feature amount data in the printing direction.

In step 12, by using the feature amount data of the print character string and the feature amount data of the preset reference character string, delimiter detection of each character unit of the print character string is performed by the DP matching method. . This DP matching method is a well-known method in the technical field such as voice recognition processing,
This is a method of performing time normalization that expands and contracts the time axis non-linearly so that the same phonemes of the continuous input speech pattern and the same phoneme of the reference pattern correspond exactly. It is called.

This DP matching is a method of temporally associating two patterns (feature vector series) that should correspond to a reference pattern and an input pattern, and for example, Tokai University Press. , Sadahiro Furui, "Digital Speech Processing," p. 162-1
67.

Also in the present invention, the DP matching processing is performed using the series of feature quantity vectors (obtained in step 11) in the printing direction (time axis direction) of the reference data pattern and the input character string data pattern. The character delimiter is detected.

In step 13, 1 is set for each delimited character.
Although the inside of the character is inspected, the DP matching method is also used in this inspection. In step 14, the cause of failure is analyzed, and in step 15, the result is output.

FIG. 2 is a processing flow showing the details of the DP matching method. Let A be the reference data pattern of the character string,
The pixels of the reference data in the time axis direction are i = 1, 2, ...
I, the data pattern of the input character string (print character string) is B, and the time-axis direction pixels of this input character string are j = 1,
2, ..., J, A = a1, a2, ..., ai, ..., aI B = b1, b2, ..., bj, ..., bJ, which is a feature vector sequence.

Note that ai and bj are pixel positions i,
It is dot pattern information in the height direction of j (see FIG. 3). I and J correspond to the lengths of the character string reference data A and the input character string data.

In the ink jet printer print inspection apparatus, the input data pattern B is time-normalized by DP matching with respect to the reference data pattern A, and pattern matching is performed using the following equation (1), and the reference data pattern ai is the most input. Find bj that is close to the data pattern.

In the time axis pixel i, the control of which state to move to next is determined optimally, the total sum of the distances ai and bj is minimized, and the similarity between the reference data and the input data is spread over the entire pattern. It is calculated as the minimum value of the distance. This minimum value can be efficiently performed by the DP matching processing shown in FIG.

In step 20, initial conditions are set, and the function of the following expression (1) is defined as an expression for pattern matching.

D (i, j) = Σ (aik XOR bjk) (1) Here, Σ represents the sum of k = 0 to k, and XOR represents the exclusive OR. Aik is the k-th element of ai,
bjk indicates the k-th element of bj, and k is the total number of elements.

Then, in steps 22 to 25, the function d
The recurrence formula g (i, j) is calculated using (i, j) until j> J. The calculation of this recurrence formula is performed in the direction of passage of time. In this case, the constraint condition (condition for preventing extreme expansion and contraction) is j−r <i · J / I <j + r, where r is a constant. Then, in step 26, the distance D (A, B) = g (I, J) between the patterns is obtained, and the smaller this value is, the better the correspondence between A and B is. (A, B) = g (I,
J) is obtained.

Note that g (i, j) is from (1,1) to (i,
The sum of the distances up to j) is minimized. Expression (1) is g (i-1, j), g (i-1, j) obtained for j, j-1, j-2 of the (i-1) th pixel, respectively.
j-1) and g (i-1, j-2), g (i.j) in the state j of the i-th pixel is obtained.

The combination of the recurrence formula (1) may be slightly different depending on the allowable range and the DP matching method.

FIG. 4 is a flow chart of the first stage DP matching processing for detecting the optimum character delimiter from the input character string data. In this process, step 3
The DP matching processing of 0, 34, 39 is performed three times in total, and the DP matching processing flow shown in FIG. 2 is basically applied to each DP matching.

In the DP matching of steps 30 to 33, each character delimiter position m is determined from the inter-pattern distance D (A, B) of A and B and the character delimiter information of the reference data A described in the processing method of FIG. (M = 1 to M: M is the number of characters)
Calculate gm (i, j) in. Assuming that the number of reference pixels in the time axis direction for one character is x, the reference character delimiter pixel im for the mth character is im = x · m.

In the DP matching processing shown in the next steps 34 and 39, since the character spacing of the input data is not constant like the character spacing of the reference data, the reference character delimiter pixel im obtained in the previous step is compared. And the character delimiter pixels im-r and im + that are moved back and forth by the constant r in the time axis direction.
gm '(i, j), gm "(i, j) for each of r
Is calculated and gm (i, j), gm ′ (i, i obtained by the DP matching process of each preceding stage in steps 30 and 34 are calculated.
The final character delimiter detection is performed while making a comparison with j).

In FIG. 5, a second-stage DP matching process is performed on each of the characters delimited by the first-stage DP matching process shown in FIG. 4 to inspect the character quality within one character. It is a thing.

In step 52 of FIG. 5, if the matching result d (i, j) of the equation (1) is larger than the set value P1 and the degree of pattern matching is extremely bad, the dot deformation amount is calculated in step 53. Is judged to be excessive, and it is determined to be defective. Then, in step 56, one dot constitutional unit (n: 0 to N) calculated from the reference data.
Gn (i, j) is calculated, and if it is larger than the set value P2, that is, the similarity is low (the distance between A and B is larger than the set value), it is determined in step 57 that a dot is missing. It is bad. The processing of steps 50 to 59 described above is the DP matching processing.

Next, in step 61, the value of g (I, J) at the character delimiter position is set to the set value P3 as in step 56.
Erroneous judgment is made by comparing with
The respective failure results obtained in the above steps 53, 57 and 62 are output to the outside.

[0033]

As described above, according to the present invention, DP
Since the matching process is used for the quality inspection of the print character string for inkjet printers, the quality inspection of the print character string including the fluctuation component in the time axis direction printed on the work being conveyed and moved by the conveyor is accurate. It has the effect of being carried out promptly.

[Brief description of drawings]

FIG. 1 is a diagram showing a function of an apparatus according to an embodiment of the present invention as a processing flow.

FIG. 2 is a processing flow chart of DP matching applied to the present invention.

FIG. 3 is a diagram showing a numeral 8 as a dot pattern.

FIG. 4 is a processing flow diagram of a first stage DP matching when detecting a character break in the embodiment of the present invention.

FIG. 5 is a processing flow diagram of DP matching in the second stage at the time of inspecting the character quality within one character in the embodiment of the present invention.

[Explanation of symbols]

 10 print character input unit 11 print direction feature amount conversion unit 12 character matching detection DP matching processing unit 13 1 character internal inspection DP matching processing unit 14 defect analysis unit 15 result output unit

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Midori Ogusa 5-7-1, Shiba, Minato-ku, Tokyo Inside NEC Corporation (72) Inventor Kenichi Matsumura 5-7-1, Shiba, Minato-ku, Tokyo Japan Electric company (72) Inventor Takeshi Fukuhara 1-33-9 Mukojima, Sumida-ku, Tokyo Asahi Breweries, Ltd. Tokyo (72) Inventor Soichiro Maeda 1-23-1 Azumabashi, Sumida-ku, Tokyo Asa Hibir Co., Ltd. Engineering Department

Claims (4)

[Claims] 1. A reading means for reading a character string to be printed by an ink jet method, and a character delimiter detecting means for performing a delimiter detecting process for each character by a dynamic programming matching method which is a dynamic programming method. A print inspection device for an inkjet printer, which is characterized in that 2. The character delimiter detection means converts the read data into feature amount data sequentially in the printing direction, and the feature amount data and the feature amount data of a reference character string corresponding to the print character string. The print inspection device for an ink jet printer according to claim 1, further comprising means for performing a character-based delimiter detection process by a dynamic programming matching method. 3. The print inspection device for an ink jet printer according to claim 1, further comprising an inspection unit that inspects the delimiter detection character by the character delimitation detection unit by a dynamic programming matching method. 4. The inkjet printer according to claim 3, wherein the inspecting unit has a unit for inspecting a deformation amount of the delimiter detection character, an incomplete dot inspection, and a typographical error inspection by a dynamic programming matching method. Print inspection device.