JPH01278364A - Method for measuring wire stroke in wire dot printer - Google Patents

Abstract

PURPOSE:To eliminate the adverse influence of noise to prevent an erroneous detection caused by noise, by a method wherein, data is sorted by labelling into a plurality of common areas, a horizontal projection distribution is applied to upper two areas with respect to the number of pixels, and the upper end and the lower ends of a pin are detected. CONSTITUTION:Because noises N1-N3 are caused by the application of dust generated from a wire etc. to a separator etc., the area of the noise is normally smaller than that of an upper end E1 or a lower end E2. With respect to image data which has been already vertically differentiated, a common area made of continuous pixels is detected and labeled with the same label, e.g. a label number. Then, the data is sorted by every area. Next, upper two areas with respect to the number of pixels included therein are selected from among the labeled areas, and a horizontal projection distribution is applied to the two areas. In this manner, since noise caused by the dust of a separator can be eliminated, the possibility of the stroke calculation using a noise as an end point is eliminated, and the reliability can be enhanced.

Description

[Detailed Description of the Invention] Summary: To provide a wire stroke measuring method for a wire dot printer that prevents erroneous stroke length measurement due to noise introduced during pin photography. For this purpose, the upper and lower end points of the pin are detected by taking the difference between the captured image data when the pin is protruding and when the pin is retracting, binarizing and performing vertical differentiation processing, and then taking the horizontal projection distribution. death,
In the wire stroke measurement method of a wire dot printer that measures the stroke of the pin, after vertical differentiation processing, the data is subjected to labeling processing to be classified into multiple identical regions, and the top two regions with the highest number of pixels are On the other hand, by taking a horizontal projection distribution, noise is eliminated and erroneous detection due to noise is prevented.

FIELD OF THE INVENTION The present invention relates to a wire stroke measuring method for a wire dot printer.

A wire dot printer uses a printer head having a plurality of thin wires to strike a platen through printing paper and an ink ribbon to form characters and figures on the printing paper using schools of fish. Simultaneous copying is possible, and running costs are low because it does not require any type of printing paper, so it is widely used as an output device for office automation equipment and personal computers. The print quality of a wire dot printer is determined by the stroke of the wire (
Since the distance between the retracted position and the protruded position of the wire greatly depends on the distance between the retracted position and the protruded position of the wire, it is necessary to measure whether the wire extends within a predetermined allowable stroke when assembling the printer head.

Some wire stroke measurement methods use a camera to take images of the wire when it is protruding and when it is being retracted, and these images are processed to measure the stroke. Due to the influence of attached dirt and the like, noise may be mixed into the image data, resulting in erroneous measurements, and an effective noise removal method is desired.

On the other hand, in the field of image processing, clusters or continuity of image data distributed on an image plane are detected, and each cluster of image patterns is numbered (labeled) to form a cluster of original image data. A labeling process is used to enable image data to be inspected effectively and quickly.

Conventional Technology Regarding the configuration of printer heads, 9-pin printer heads have been mainly used in the past, but 24-pin printer heads are now being adopted to improve printing quality. Along with this, the conventional 9-pin wire stroke measuring device cannot be applied as is, so the inventors have developed a wire stroke measuring device as described below.
A wire stroke measuring device for 4 bins has been proposed.

The printer head has a first & second pin row made up of 12 pins and a second pin row made up of 12 pins, and is attached to a jig (head mounting base).

A separator such as a thin white plastic plate is provided between the first row of pins and the second row of pins using a jig. A first camera is provided to photograph the pin row, and on the other side of the separator, a second illumination device and a second camera are provided to photograph the second row of pins.

FIGS. 9 and 10 are diagrams showing the processing procedure for images taken by the camera, with FIG. 9 being a flowchart and FIG. 10 being a schematic diagram. First, in step 202, pinned image A is captured, and in step 204, pinned image B is captured. Next, the difference A-B between the pinned image A and the pin saved image B is taken, and the extended pin portion is cut out (step 205). This is binarized and vertically differentiated (step 206), a horizontal projection distribution (histogram) is taken (step 207), and extraction points are read out (step 208). Next, in step 209, the upper end point H1 and the lower end point H2 are detected, and by multiplying the difference by the length per pixel,
I am trying to measure the wire stroke. here,
Regarding the calculation of the upper end point H1 and the lower end point H2 in step 209, for example, in the case of Hl, as shown in FIG. 8, the vicinity of the end point has an impulse-like peak, so
A certain threshold is set, the coordinates Pl, P2 where this threshold cuts the impulse are detected, and this P+,
Let the midpoint of P2 be the end point H1. The same applies to the end point H2. The above process is performed one after another for 24 pins (step 210), and based on the results, a pass/fail judgment is made (step 211). If the pin is defective, the position of the pin is adjusted as appropriate in step 212.

Problems to be Solved by the Invention In this way, in the past, a separator was inserted between the first pin row and the second pin row to make the contrast between the pin image and the background clearer, and to make the contrast between the captured image and the background part clearer. , the edge was detected by differential processing in the vertical direction. However, this thin plate is easily contaminated due to the influence of dust attached to the wire, and as a result, the image after vertical differential processing may end up containing noise, as shown in Figure 11. When the horizontal projection distribution is taken, it becomes as shown in FIG. 12, and there is a problem in that the noise portion is detected as an end point and an erroneous measurement result is produced. Note that setting a high threshold value may not be able to detect a necessary portion, so it is not a good way to avoid the influence of noise.

The present invention has been made in view of these points, and its purpose is to provide a wire stroke for a wire dot printer that does not cause erroneous stroke length measurements due to noise introduced during pin photography. The objective is to provide a measurement method.

Measures to solve the problem After taking the difference between the captured image data when the pin is protruding and when the pin is retracting, and performing binarization and vertical differentiation processing, the data is subjected to labeling processing to classify it into multiple identical regions, and the pixel The horizontal projection distribution is taken for the top two areas with the largest numbers, and the top and bottom points of the pins are detected. This eliminates the influence of noise, prevents false detection due to noise, and solves the above-mentioned problems.

Function To explain with reference to Fig. 1, the noises NI to N are caused by dust from wires etc. adhering to the separator etc., so normally the noises NI to N are generated at the upper end point E1 and the lower end point E.
Its area is small compared to 2. Focusing on this, we perform a labeling process on the image data after vertical differential processing, which detects the same area of consecutive pixels and attaches the same label to it. Classify into. Next, among the labeled regions, the top two that contain the largest number of pixels are selected,
A horizontal projection distribution is taken for these two regions (second
(see figure). This makes it possible to eliminate noise caused by dirt on the separator, etc., so that the stroke is no longer calculated using noise as an end point, as in the past, and its reliability can be improved.

Embodiments The present invention will be explained in detail below based on embodiments shown in the drawings.

FIG. 4 is a block diagram of a 24-pin measurement system to which the present invention is applied, FIG. 5 is a block diagram of the main parts of FIG. 4, and FIG. 6 is a view taken in the direction of arrow A in FIG. The printer head 20 is attached to a jig (head mount) 26. The printer head 20 has a first pin row 22 made up of 12 pins and a second pin row 2'4 made up of 12 pins (see FIG. 6). A separator 28 made of a white plastic thin plate or the like is provided between the first row of pins 22 and the second row of pins 24 using a jig 27.

On one side of the separator 28, a first lighting device 30 and a first lighting device 30 are provided.
A first camera 32 is provided to photograph the stage pin row 22, and a second illumination device 34 is provided on the other side of the separator 28.
A second camera 36 for photographing the second row of pins 24 is provided.

The video signals of the cameras 32 and 36 are processed by the grayscale image processor 40.
and image processing is performed. Camera images of cameras 32, 36 can be monitored by monitor 42. 44 is a personal computer, which connects the rcRT 48 via a bus 46.
, keyboard) is connected to a 50° floppy disk 52. Furthermore, a grayscale image processor 40 is connected to the bus 46 of the personal computer 44 . The grayscale image processor 40 is controlled by a personal computer 44, and can read and write data to and from the personal computer 44. Switching of cameras 32 , 36 is controlled by personal computer 44 via grayscale image processor 40 . Further, a pin driver 58 and lighting devices 30 and 34 are connected to the bus 46 of the personal computer 44 via Pf056.

With this configuration, lighting of the lighting devices 30 and 34 and loading and unloading of the printer head 20 can be controlled by the personal computer 44 via the PIO 56. still,
The first and second lighting devices 30 and 34 are the main lighting devices 30a and 34a and the lower lighting device 3, respectively, to make illumination uniform.
It is divided into 0b and 34b (see Figure 5).

FIG. 3 is a flowchart of measurement processing performed by the apparatus configured as described above. First, the bottle is ejected via the pin driver 58 by a bottle drive signal from the personal computer 44 (
Step 101), a pinned image (RA) is captured in step 102.

Similarly, the pin is retracted (step 103), and step 1
At 04, the pin evacuation image B is captured. Next, the difference A-B between the pinned image A and the pinned image B is taken, and the extended bin portion is cut out (step 105).
. In step 106, as shown in FIG. 7, the difference A-B between the pinned image A and the pinned image B is binarized, and isolated pixels of the noise occurring in the image data are divided into isolated pixels. (corresponding ones) is performed, and then vertical differentiation is performed to extract images of the top and bottom ends of the bins. After this is binarized again and isolated points are removed, the process proceeds to step 107 in FIG.

In step 107, the image data is labeled (classified) for each continuous region, and in step 108, the top two labeled regions with the largest number of pixels are selected, and the other regions are cropped.

Next, a horizontal projection distribution (histogram) is taken for the two areas with a large number of pixels (step 109), extraction points are read out (step 110), and step 11
1, the upper end point Hr and the lower end point H2 are detected, and the stroke of the pin is calculated.

In step 111, regarding the calculation of the upper end point H1 and the lower end point H2, for example, in the case of Hl, as shown in FIG. Coordinates P to cut
3. Detect Pz and set the midpoint of this P and P2 as the end point H
+. The end point H2 is also obtained in the same manner.

Also, as another method, the coordinate P at which the impulse is cut is
The end point H2 is determined by taking a weighted average between l and Px as shown in the following formula.

P.

p near address (horizontal distribution coordinate position) G,: Number of pixels at p address The end point H2 is found in the same way.

The wire stroke can be measured by multiplying the difference between the end points H1 and H2 thus obtained by the length per pixel.

The above processing is performed one after another for 24 pins (step 112),
Based on the results, a pass/fail judgment is made (step 113),
If the pin is defective, the position of the pin is adjusted appropriately in step 114.

The labeling process performed in step 107 is to detect the same area of consecutive pixels in the image and attach the same label to it. For example, the i-image data labeling method (patent application The algorithm of No. 62-!50298) is configured using the hardware method shown in the same application, and is realized by adding this to the grayscale image processor 4o. Thereby, high-speed processing can be performed in real time, an increase in measurement time due to an increase in processing procedures can be prevented, and stable results can be obtained.

Effects of the Invention As described in detail above, even if noise occurs in image data due to wire dirt, etc., this invention is processed by labeling, and the top two with the largest number of pixels are selected. Since the wire stroke is determined based on , it is possible to obtain accurate results without the influence of noise.

[Brief explanation of the drawing]

FIG. 1 and FIG. 2 are diagrams for explaining the present invention in detail, and FIG. 1 is an explanatory diagram of an image subjected to labeling processing,
FIG. 2 is a diagram showing the horizontal projection distribution for the top two areas with the largest number of pixels. FIG. 3 is a flowchart showing the processing procedure of an embodiment of the present invention, FIG. 4 is a configuration diagram of a 24-pin wire stroke measuring device suitable for carrying out the present invention, and FIG. 5 is the main part configuration of FIG. 4. 6 is a view taken in the direction of arrow A in FIG. 5, FIG. 7 is a detailed flowchart of step 106 in FIG. 9 and 10 are flowcharts and schematic diagrams showing the processing procedure of the prior art, and FIGS. 11 and 12 are diagrams for explaining the problems of the prior art. 20... Printer head, 22... First stage pin row, 24... Second stage pin row, 28... Separator, 30... First lighting device, 32... First camera, 34 River 2nd lighting device, 3
6... Second camera, 40... Grayscale image processor, 44... Personal computer.

Claims (1)

[Claims] Difference between captured image data when the pin is protruded and when the pin is retracted is taken (105), and binarization and vertical differentiation processing is performed (106).
In the wire stroke measuring method for a wire dot printer, the vertical projection distribution is then taken (109) to detect the upper and lower end points of the pin (111) and the stroke of the pin is measured. After the directional differentiation process (106), the data is labeled and classified into multiple identical areas (107), and the horizontal projection distribution is taken for the top two areas with the largest number of pixels (108, 109). A wire stroke measurement method for a wire dot printer characterized by eliminating noise and preventing false detection due to noise.