JPH0250282A - Data reading device - Google Patents

Abstract

PURPOSE:To easily and speedily execute the reading and recognition of encoding picture data by calculating a vertical direction sampling position and a horizontal direction sampling position out of the data of a read picture and sampling the data of the picture in sampling coordinates. CONSTITUTION:The recording sheet of the picture to include a mesh-shaped pattern, for which the data are encoded according to brightness and darkness to be selectively formed in respective meshes, is used as a recording medium. Further, in order to recognize the brightness and darkness of the respective meshes from the picture which is read by an image sensor means, a hardware is used to detect the vertical and horizontal data sampling positions of the mesh-shaped pattern and to sample the picture data at the position. Thus, data recognition can be executed at a high speed to the encoding picture of high density.

Description

[Detailed description of the invention]

[Technical Field of the Invention] The present invention relates to a data reading device (image input device a) that captures an encoded image and reads data encoded in the image. [Background] Barcode technology is known as a technology for capturing encoded images from a recording medium (recording sheet) and reproducing binary data. However, in the case of barcodes, it is difficult to increase the recording density due to their structure, so they are not suitable for inputting large amounts of data. JP-A-53-73026 describes 1Xj (for example, 3×3
) in a matrix in which one of the squares is darkened and the other squares are brightened, the technology is first read in to recognize the pattern of brightness and darkness of the eyes. It is thought that the capacity of the coded image of this matrix can be increased by increasing the number of eyes included in the matrix. However, in the above-mentioned publication, since software is first used to recognize the brightness and darkness patterns of the eyes, pattern recognition requires a considerable amount of time and is expensive. An example using a similar matrix image is magazine I1.
0. May 1988 issue, pages 121-125 [Dump list to be read with image scanner]. [Object of the Invention] Therefore, the main object of the present invention is to provide a data reading device that can easily and quickly read and recognize encoded image data with a relatively high recording density. Another object of the present invention is to address the problems that occur when encoded image data is input sequentially (one line at a time) using a hand-held image scanner, such as erroneous recognition of data due to changes in scanning speed, direction, etc. An object of the present invention is to provide a data reading device that solves the problem. [Structure, operation, and development of the invention] A data reading device according to the present invention is an image including a mesh pattern in which data is encoded by brightness and darkness selectively formed in each mesh, and in which the mesh pattern is Encoding a recording sheet on which an image is formed that includes a first mark for indicating a data sampling position in the vertical direction of the mesh pattern and a second mark for indicating a data sampling position in the horizontal direction of the mesh pattern. Used as an image recording medium. The single data reading device includes an image sensor means for reading the image on the recording sheet, and also serves as a hardware means for recognizing the brightness and darkness of each mesh. vertical sampling position calculation circuit means for detecting one mark and calculating the data sampling position in the vertical direction; and detecting the second mark from the data of the read image and calculating the data sampling position in the horizontal direction. and data sampling circuit means for sampling data of the image at sampling coordinates defined by the calculated data sampling positions in the vertical and horizontal directions. The above image can be made to have a sufficiently high recording density by minimizing the mesh size. Further, since image recognition a (recognition of the brightness and darkness of each mesh) is performed by hardware that detects a sampling point and samples data at that point, it can be performed in a short time. The first mark may take various forms, e.g.
The mark is composed of a pair of bars extending along the upper and lower sides of the mesh pattern. In this case, the vertical data sampling position calculating means includes an interval measuring circuit means for measuring the interval between the pair of bars included in the read image data, and a position where the measured interval between the bars is divided approximately equally. and a dividing circuit means for calculating the data sampling position in the vertical direction. According to this configuration, even when the image sensor means captures an image on the recording sheet obliquely, the correct vertical data sampling position can be maintained. This is because the read image is only tilted, and the relative positional relationship between the pair of bars and the mesh pattern is preserved. The distance between a pair of bars changes depending on the value measured in a direction perpendicular to the bars and the value measured in a direction not perpendicular to the bars, but the rate of change is determined by measuring the distance between the top and bottom sides of the mesh. Therefore, the interval between the dividing points of the one-dividing circuit means can always correspond to the interval between the vertically adjacent meshes. As another example of the first mark, a row of special patterns may be incorporated into the mesh pattern. Place it by shifting it. Then, during the data reading operation, each edge of this special pattern, that is, the position where the pattern changes from bright to dark or from dark to bright, is detected. This detection position represents a data sampling position in the vertical direction. The image sensor means may be of a line type, for example, while scanning (moving) the recording sheet from one side of the net-like pattern to the other side, image data of one line at a time ( (almost vertical line image data). When such a line-type image sensor means is used, the image data at the data sampling position in the horizontal direction is line image data given from the image sensor means at a certain point in time, and knowing the timing of arrival of this data is essential for the horizontal direction. This is equivalent to knowing the data sampling position. Therefore, in one configuration example, the border (vertical border) of the bright and dark meshes adjacent to the left and right included in the mesh pattern is used as the second mark, and the horizontal direction The sampling position calculation circuit means includes mesh boundary detection circuit means for detecting the boundary, and circuit means for detecting a timing at a predetermined time interval after the boundary is detected by this means as the timing of the data sampling position in the horizontal direction. configured. The above time interval is determined by the line type sensor means scanning laterally (
It is constant if the speed at which it is moved (manually or mechanically) is constant. However, even when the scanning speed varies, when a constant time interval is used, the sampling position in the horizontal direction may shift from the center of the mesh, and there is a possibility that incorrect data will be sampled. Therefore, in the configuration example, a lateral speed detection circuit means for detecting the scanning speed of the image sensor means for horizontally scanning the mesh pattern is used, and the lateral data sampling position calculation means is used to detect the detected scanning speed and the horizontal direction. The data sampling position in the horizontal direction is calculated from the boundary position of meshes adjacent to . A rotary encoder can be used as the lateral speed detection circuit means. In another configuration example for compensating for fluctuations in the moving speed of the line image sensor means, the mesh pattern includes a special pattern, that is, a special pattern in which at least one horizontal row of mesh alternately repeats brightness and darkness. In order to know the data sampling position in the horizontal direction of each mesh, a boundary detection circuit means for detecting the boundary between adjacent bright and dark meshes in each row in the horizontal direction in the mesh pattern; special boundary detection circuit means for detecting the boundary between the bright and dark mesh; and a lateral circuit for detecting the horizontal scanning speed of the image sensor means by measuring the time between the boundaries detected by the special boundary detection circuit means. directional scanning speed detection circuit means, determining the horizontal data sampling position of each mesh in the mesh pattern from the boundary timing detected by the boundary detection circuit means and the scanning speed detected by the lateral scanning speed detection circuit means. lateral sampling position calculation circuit means for calculating the horizontal sampling position. As another example of a second mark for indicating the data sampling position in the lateral direction of the reticulated pattern. Between each row of the mesh pattern in the horizontal direction, rows of meshes (special pattern) that alternately repeat bright and dark patterns are arranged so as to be shifted by half of the mesh. Then, when reading data, the positions where these special patterns change from bright to dark and from dark to bright are detected. The detected position for the special pattern in a certain column is used as the data sampling position in the horizontal direction for the mesh pattern in the adjacent column. For this configuration,
There is no need for circuit means for detecting the change position (boundary) of brightness and darkness in columns other than the columns of the special pattern. In order to save storage capacity, data sampling is performed by sampling data at the vertical sampling position calculated by the vertical sampling position calculation circuit means from the image data sent from the image sensor means. It is preferable to sample data at the horizontal sampling position calculated by the horizontal sampling position calculation circuit means from among this vertical sampling data (temporarily stored as sampling data in the horizontal direction and discarding other data). . [Embodiment] Hereinafter, an embodiment in which the present invention is applied to inputting performance data of an electronic musical instrument will be described. Fig. 1 is a conceptual diagram of the whole, in which an image EI in which performance data (score data) is encoded is printed with high density in the margin of a musical score S, and this musical score and data image EI are scanned by a data reading device DR. The data is then encoded and transferred to the electronic musical instrument MI for performance. An example of the encoded image EI is shown in FIG. 2. As shown in FIG.
P, and each mesh is bright (b) or dark (■), and along the upper and lower sides of the mesh pattern MP,
An upper bar A-1 and a lower bar A-2 are printed. As will be described later, this pair of bars A-1 and A-2 is used to determine sampling points in the vertical direction of the mesh pattern MP. In the mesh pattern MP, the two vertical rows at the "left end" are a checkerboard pattern, and the two horizontal rows at the "top end" (2
The two lines, ``line'' and ``bottom'', also have a checkered pattern. These patterns indicate edges. The number of rows of the mesh pattern MP is 12 here, so the leftmost two columns and 8 rows excluding the top and bottom are the data use area (area decoded by the data reading device DR). In the data usage area, one note is expressed in two vertical columns, the odd-numbered column from the left represents pitch data, and the light and dark pattern of the three meshes from the top is the octave code (O
C), the brightness of the following four meshes is the scale code (S
C), the brightness of the eighth mesh represents parity, the even-numbered columns of the data use area represent the note length, the brightness and darkness of the seven meshes from the top represent the time code TC, and the eighth It is parity. FIG. 3 shows an example of the code system when the dark mesh is "l" and the light mesh is "0". FIG. 4 shows the overall configuration of the data reading device. The image data read from the sensor part l is sent to the control circuit part 2.
This data is converted into 1-bit 71-dot data (indicating the brightness and darkness of each mesh) and written into the memory 3. This data is sent to the CPU 4.
and read the data of each application (in this case,
performance data) and transfer it to an external device (in this case, an electronic musical instrument). The sensor part 1 is, for example, a contact type line image sensor, and an LED array l-3 as a light source emits light, and the image is converted into an electrical signal by the sensor element array 1-1 through a rod lens array 1-2. The width of some lines of the sensor depends on the usage conditions, but as shown in Figure 5,
When scanning the sensor part l from the left end to the right end of the mesh pattern MP of the encoded image EI, the width must be such that all the encoded images are captured even when scanning is performed at the expected maximum slope. The part surrounded by the dotted line in FIG. 5 is taken in as one image data. Since the memory 3 and CPU 4 in FIG. 5 are common ones, their explanation will be omitted. The control circuit unit 2 is a main part of the data reading device, and here, the brightness and darkness of each mesh in the mesh pattern of the encoded image EI is recognized. Two configuration examples will be shown as the control circuit unit 2, and will be explained below. Example of Control Circuit Unit (1) FIG. 6 shows a control circuit unit 2 according to a first configuration example. Image data is inputted to this control circuit unit 2 from the sensor unit l as described above. Here, it is assumed that the sensor part l scans the encoded image EI from left to right as shown in FIG. FIG. 7 is shown for illustrative purposes, in this figure. This shows a case where the encoded image EI is scanned diagonally,
If the actual size of the mesh is 0.5m x 0.5mm, then 16
It has been enlarged twice. Therefore, per lII■, 1
Assuming that the sensor part l has a resolution of 6 trees, in FIG. 7, the brightness and darkness will be read every [■]. line 7-3
．． 7-4 represents one line of image input data of the sensor part 1. In FIG. 6, 5DATA is serial image input data from sensor part 1, in this case approximately 160 bits/line. This input data 5DATA is a flip-flop FF operated by clock φ1 (see Figure 10).
After passing through I, it becomes an input to the 2,0 bit shift register 2-1. Output of 20-bit shift register 2-1 and inverter I
Input data 5DATA via NV1 is R of NAND configuration.
The signal is input to OMI, and a signal FT (see FIG. 1θ) is generated. This signal PT indicates that the input data 5DATA is 20
These circuit elements become row active when continuous bit data changes to white 1-bit data.
and point 7-2) are detected. In FIG. 6, unless otherwise specified, latches and registers perform read operations using clock φ1 and output operations using clock φ2 (see FIG. 10).
Along with φ1 inverted at NV2, it passes through the gate Gl,
It is input to the circuit 2-3 whose main elements are an S-R flip-flop and a D flip-flop, and the edge 7 of the upper and lower bars
Signal D becomes high active from -1 to edge 7-2
ON (see Figure 10). In addition, circuit 2-3
is reset by inputting a signal SH (see FIG. 1O) generated every time one line is inputted from the sensor part 1 via an inverter INV3. The signal DON is a counter 2-1 that is reset by the signal SH.
is input, and while DON is “H”, that is, the upper bar A
The clock Φl is counted up from the edge 7-1 of -1 to the edge 7-2 of the lower bar A-2. On the other hand, circuit 2
-3 signal DON passed through FF2 and another signal TPT of circuit 2-3 (Q output of D flip-flop)
through gate G2, and at the later edge 7-2.
This signal is switched to N and is input to the 8-bit latch 2-5 as a load control signal. The count output of the counter 2-4 at this point is the previous edge 7-1 (7
-7) to the subsequent edge 7-2 (7-8), and this data is taken into latch 2-5. Note that the signal that has passed through gate G2 is passed through gate G3 together with the signal that has passed through inverter INV4 of clock φ1, and becomes a signal φL (valid line completion signal) that generates a high-level pulse at the later timing of edge 7-2. (1st
(See figure O). The circuit on the output side of latch 2-5 (2-6, 2-7, 2
-8, 2-9, INV5. G4) is a(1) edge 7
This circuit divides the distance from -1 to the subsequent edge 7-2 into approximately 16 equal parts, thereby obtaining the timing signal φS (see FIG. 10) at the sampling position in the vertical direction of the mesh pattern MP. The sampling position is 7-6 in Figure 7.
It is indicated by (...). 16 means that in the case of the image data in FIG. 7, the width of bar A-2 is three meshes, and there are 12 lines of meshes between the upper and lower bars A-1 and A-2. This is because the distance between the bars A-1 and A-2 and the mesh pattern is half a mesh. therefore. If the timing from the detection time of edge 7-1 to the detection time of edge 7-2 is obtained by dividing it into 16 equal parts, the first 12 timings will represent the vertical sampling position of the mesh for 12 lines. become. This circuit (2-6, 2-7, 2-8, 2-9, INV5
．． G4) will be described in detail. The upper 4 bits of the latch 2-5 become the input to the half adder (HA) 2-6, and the output becomes the load data for the subtracter 2-7. This subtracter 2-7 decrements the data by 1 every clock φ1, and loads the data of the half adder 2-6 every time a poll occurs. Also, the pollo output is '7' along with the signal passed through the inverter INV5 of the clock φ1.
'-) G4 and becomes a vertical sampling signal φS. On the other hand, the lower 4 bits of the 8-bit latch 2-5 become one side input of a 4-bit full adder (FA), and its output is taken into the latch 2-9 for each signal φS.
The output of -9 becomes the other input of full adder 2-8. The carry output of the full adder 2-8 becomes the LSB input of the 4-bit half adder 2-6. For example, line 7 in FIG.
-4, there is 130 bits of image data from edge 7-7 to edge 7-8, this value of 130 is counted by counter 2-4, and “
The data is latched as "10000010". .phi.S is output once for every upper 4 bits "1000' = 8.phi.". However, since there are the lower 4 bits "001O", the carry is transferred from the full adder 2-8 to the half adder 2 every 8φS.
-6, so the subtractor 2- which calculates φS
The load data of 7 is incremented by +1. Therefore, in that cycle, φS is output with a delay of φ1 minute. This vertical sampling signal φS operates the 4-bit shift register 2-1O and the 12-bit shift register 2-11 at the subsequent stage to take in (sample) the image input data from FFl. The output of 2-11 is latched into the 12-bit latch 2-12 by the timing signal φ[ of the trailing edge 7-8. The data that enters the latch 2-12 at this time is the data at the sampling position in the vertical direction for each mesh out of the current one line of image data 5DATA. The output of the latch 2-12 is sent to the 12-channel sampling circuit 2-13, where the data at the horizontal sampling position (data indicating the brightness/darkness (0/1) of each mesh) is selected and stored in the memory. Written in 3. Sample circuit 2-
The details of 13 will be described later. FIG. 8 represents the sampled data. That is, the upper data is the contents of the latch 2-12 for each line (φL). For example, the first line is the data of the sampling position when scanning the line 6-3 in Figure 7, and similarly, the seventh line is the data of the line 7-4, and the bottom line is the data of the line 7-5. It shows vertical sampling data when scanning. FIG. 8 shows the data actually written to the memory 3,
The sample circuit 2-13 (FIG. 9) performs the conversion during this time. In FIG. 9, the counter 9-1 is for counting the number of successive bits "O" or "l" (the number of lines φL) with respect to the vertical column of data on the upper side in FIG. It is reset by E, X-0R9-2 when switching from θ" to "l-" or from "1" to O". In other words, when EX-OR9-2 detects a mismatch between the vertical sampling data of the current line from the corresponding 1-bit latch 2-12 and the vertical sampling data of the previous line from FF3, which provides l-line delay. The counter 9-1 is reset. The output of the counter 9-1 is input to the ROM 2, and the ROM 2 is stored for 4 lines when the output of the counter 9-1 is 4 or 12, that is, after the output from the latch 2-12 changes to "θ" or "1". Alternatively, the horizontal sampling signal φhs is generated after a time of 12 lines (4, 12 means modulo 16). The output of latch 2-12 changes from “0” to “1” or “
Changing from 1" to "θ" corresponds to looking at the switching (boundary) from the left and right adjacent dark meshes to the bright meshes, or vice versa, in the mesh pattern MP. In FIG. 9, the circuit 9-3 is provided in the sample circuit 2-13 of the odd channel (odd column from the left in FIG. 8), and the circuit 9-3 is provided in the sampling circuit 2-13 in the even channel.
4 is used. This circuit 9-3.9-4 is for detecting the start of data. That is, as shown in FIG.
The meshes in the rows on the layer are arranged from dark (“1”) to light (“O”), while the meshes in even rows are arranged from bright (“0”) along the scanning direction (Figure 5). ”) to dark “1”. Therefore, odd channel circuit 9-3
Now, detect this change from dark ``1'' to bright ``°0'' at the left end (starting end) from the data of latch 2-12 of the current line and the data of latch 2-12 of the previous line, and then detect the change in the even channel circuit 9. -4 detects the change from bright to dark at the left edge.The output of circuit 9-3.9-4 is reset by the signal generated when the data reading device starts scanning the image.
F4 is set to output a signal EN indicating that data is being scanned. After the data starts, the signal φ (gate) passes through G5, and the clock φw (occurs at the falling edge of the signal DON).
A write signal φ is sent from gate G5 to memory 3 at the timing of
1 is generated, and the contents (DATA) of FF3 at that time are written to memory 3. The signal φns passing through G5 increments the address counter 9-5 of the memory 3 and reads the next address (ADDRES) of the memory 3.
S). FIG. 10 shows a time chart of the main signals. In the figure, ROM is a signal indicating that the data reading device is scanning an image, and 5RON generates one pulse at the start of scanning. φ- is the rising edge of the signal DON (upper bar A-
This is a clock generated when the edge 7-1 of the edge 7-1 passes. The other details have been explained with reference to FIGS. 6 and 7, and will therefore be omitted. In this way, the control circuit section 2 of the first configuration example (FIGS. 6 and 9)
(Fig. 2) receives 5 DATA of image data from the sensor part 1 that moves the encoded image from left to right and reads the image line by line as shown in Fig. 2, and converts it into image data for one line. Edges 7-1 and 7-2 of the included upper bar A-1 and lower bar A-2 are detected, the interval between them is calculated, and the interval is divided almost equally to perform vertical sampling of the mesh pattern MP. Detect the position (2-1, ROM1
．． 2-3 to 2-9, etc.), and the vertical sampling data is extracted from the image data of the next line using the detected vertical sampling position signal φS (2-1o, 2-11.2). FF3.9-
9-1), and measure a predetermined time interval from the time of detection. ROM2), the arrival timing of that time interval is taken as the timing of the sampling position in the horizontal direction,
Corresponding vertical sampling data is extracted and written into the memory 3 (by G5, G6.9-5, FF3, etc.). Therefore, the control circuit section 2 of the first configuration example can quickly recognize the brightness and darkness of each mesh in the mesh pattern MP. Further, even when the sensor part 1 moves diagonally to scan the encoded image EI as shown in FIG. 7, no problem occurs in data recognition because the sampling position in the vertical direction is secured. However, since the sampling position in the horizontal direction is set at the arrival timing of a predetermined time interval between the boundary forces of the bright and dark meshes adjacent to each other on the left and right, in the case of a line-type image sensor in which the sensor part 1 is manually moved. is the mesh pattern ML
Since the scanning speed in the horizontal direction varies, the data sampling position in the horizontal direction may shift by the amount of hair, leading to incorrect data recognition. A control circuit section (second configuration example) that can also deal with this problem will be described below. The control circuit section of the second configuration example includes a circuit that detects the horizontal scanning speed of the sensor part l from the special pattern included in the mesh pattern MP. This special pattern is the mesh pattern M shown in FIG.
These are the patterns shown at the "upper end" and "lower end" of P. That is, a mesh pattern that alternately repeats brightness and darkness along the lateral direction is used. Although brightness and darkness are not shown in the data area in FIG. 11, the description will be made assuming that the data area changes from bright to dark at least every two horizontal meshes. For example, 1 bit is represented by two data meshes, and bit 1 is

[or], bit 0 is ■ or 2, and is the left neighbor of the 1-bit mesh pair. The encoded image EI should be recorded so that a bright and dark boundary is formed between it and the mesh on the right.6For example, when recording bit 1 on the left of [(bit 1), select pattern []. , select cabinet when recording bit 0. Other operating conditions (scanning direction, etc.) are the same as in the first configuration example. The control circuit section 2M of the second configuration example is shown in FIG. 12.The configuration for detecting the vertical sampling position and the configuration for sampling image data at the vertical sampling position are the control circuit of the first configuration example! It is the same as 2. The difference is 2
The difference is that a channel scanning speed detection circuit 2-14 is provided, and the sample circuit 2-13M is changed. The two-channel scanning speed detection circuits 2-14 have the same configuration, and the first one of the 12-bit latches 2-12
In response to the input from the bit and the 11th bit, the horizontal scanning speed of the sensor part 1 is detected from the patterns in the 1st and 11th rows in the mesh pattern MP shown in FIG. FIG. 13 shows the configuration of the scanning speed detection circuit 2-14. The mismatch between the vertically sampled image data of the current line from the latch 2-12 of the 1st or 11th bit and the vertically sampled image data of the previous line that has passed through FF5, that is, the 1st bit which is the special pattern in FIG. The boundary between the adjacent bright and dark meshes on the left and right of the row or the 11th row is E.
It is detected by the X-OR14-1, and the signal passes through the flip-flop 14-3 to reset the counter 14-4, and passes through the gate 14-2 to generate the signal φN (clock for latch) at the timing of clock φ1. do. The counter 14-4 counts the number of lines in which bright (bit 0) or dark (bit 1) continues from the boundary using the line signal φE. Here. The bright and dark boundaries in the special pattern (first and eleventh rows in FIG. 11) occur for each mesh. Therefore, the count value of the counter 14-4 represents the time from when a certain boundary is detected to when the next boundary is detected. This is nothing but detecting the horizontal scanning speed of the sensor unit 1.The count value of the counter 14-4 is input to one side of the full adder (FA) 14-5 and added to the previous value. , the value obtained by averaging the addition results by the divider 14-6 is latched in the latch 14-7 at the boundary timing φ open. The output n of the latch 14-7 is sent as the previous value to the other input of the full adder (FA) 14-5, as well as the divider by 1 14-8 and the latch 14-8 of φN operation.
34n signal is generated. Furthermore, the latch 14
The output n of -7 and the output of the % divider 14-8 are completely added together! ! (
FA) Added by 14-0, latch 14-1 of φ round operation
1, a 3/2n signal is generated. Here, the %n signal represents the time required to move half the mesh at the current lateral scanning speed of sensor part 1, and the 3/2n signal represents the time required to move half the mesh at the current lateral scanning speed. number plus the time required to move to the station. These two signals %n and "/2n are sample circuit 2-13
M is used to detect a horizontal sampling position signal. The configuration of the sample circuit 2-13M is shown in FIG. 14. The difference from the sample circuit 2-13 (FIG. 9) of the 1st #I example is that the configuration of the sample circuit 2-13M is A coincidence circuit 9M compares the %n signal and the 3/2n@ signal, which change according to the scanning speed, with the output of the counter 9-1 to detect a coincidence.
is used. That is, the matching circuit 9M detects the bright/dark boundary of the data rows in the mesh pattern MP by the EX-OR 9-2, and then starts the counter 9-2.
1 detects the timing when y2n time or 3/2n time has elapsed and generates a horizontal sampling position signal φn.
generate s. As mentioned above, %n time corresponds to half the mesh movement. 3/2fi time corresponds to the movement of one mesh plus half. On the other hand, the data rows (the third row to the first θ row in FIG. 11) change from bright to dark or from dark to bright at a period of at least two meshes. Therefore, the coincidence signal φ of the coincidence circuit 9M
ns specifies the approximate center position of each mesh in the data row, regardless of fluctuations in scanning speed. Note that there are two scanning speed detection circuits 2-14, and the output of the scanning speed detection circuit 2-14 for the special pattern at the upper end (FIG. 11) (touch n, 3/2n) is a sample for the upper half data row. The output of the scanning speed detection circuit 2-14 for the lower end special node is input to the sample circuit 2-13M for the lower half data row. Instead, the output of the scanning speed detection circuit 2-14 that scanned the special turn first is sent to all the sample circuits 2-1.
The output of the scanning speed detection circuit 2-14 that is input to 3M and used to scan a special pattern later may be disabled.
For other points, see sample circuit 2-13 in Figure 14.
Since M is the same as the sample circuit 2-13 in FIG. 9, its explanation will be omitted. In this way, the control circuit section 2M of the second configuration example detects the horizontal scanning speed of the sensor part l from the special pattern included in the mesh pattern ML, and uses the detected scanning speed and the data row of the mesh pattern ML. Since the data sampling position in the horizontal direction is detected from the boundary of the bright and dark meshes detected in relation to This makes it possible to reliably recognize data even in encoded images from a line-type image sensor that scans at a variable moving speed. [Modifications] This concludes the description of the embodiments, but these embodiments are merely illustrative, and various modifications and changes can be easily made. For example, the scanning speed dependent signals %n and 3/2n input to the coincidence circuit 9M in FIG. 14 may be generated from the output (scanning speed signal) of a rotary encoder provided in the sensor part 1. Regarding the control circuit section 2M of the second configuration example, it is assumed that the data row of the mesh pattern MP changes from bright to dark or from dark to bright with a cycle of at least two meshes. For example, in FIG. 14, the output of the counter 9-1 is a signal n(
A circuit for resetting the counter 9-1 when a match occurs (see FIG. 13) may be added, and a match circuit 9M may be used to detect a match between the output of the counter 9-1 and the signal 54n. Also, in Figure 11, the special pattern is two lines at the top,
Two lines are formed at the bottom edge, but instead of this, one line of special pattern (mesh that repeats light and dark alternately) may be formed as the center line of the net-like pattern and extended longer to the left and right than the other lines. Yes, in this case, the scanning speed detection circuit 2-14 may have one channel. Further, although the net-like pattern shown in one figure has a rectangular shape as a whole, it may have a parallelogram shape. Further, the image data input to the control circuit section 2 may be data transferred from the image memory of the encoded image EI. In addition, various modifications and changes are possible. [Effect of the Invention J] The data reading device according to claim 1 uses as a recording medium an image recording sheet including a mesh pattern in which data is encoded by brightness and darkness selectively formed in each mesh, and further includes an image sensor. - Hardware that detects the vertical and horizontal data sampling positions of the pattern and samples the image data at those positions in order to recognize the brightness and darkness of each mesh from the image read by means (vertical data sampling position Since the calculation circuit means, the horizontal data sampling position calculation circuit means, and the data sampling circuit means are used, high-density encoded images can be data recognized at high speed. In the data reading device according to claim 2, a pair of bars are formed on the image along the upper and lower sides of the mesh pattern as marks for indicating the sampling position in the vertical direction of the mesh pattern. A position calculating circuit measures the distance between the pair of 8's and divides the distance into approximately equal parts to detect sampling positions in the vertical direction. Therefore, even when the image sensor means reads the encoded image of the recording sheet at an angle, the data sampling position in the vertical direction can be ensured at the correct position. Further, in the data reading device according to the second aspect of the present invention, while moving from one side of the mesh pattern toward the other side as the image sensor means! A line-type sensor is used to read the image line by line, and in connection with this, the horizontal data sampling position calculation means detects the boundary between the bright and dark meshes adjacent to the left and right included in the mesh pattern, and The timing at a predetermined interval from the detection is detected as the timing of the horizontal data sampling position. Therefore, the construction is simple and, as long as the moving (scanning) speed of the image sensor means is constant, a correct lateral sampling position can be obtained. However, it is not suitable for image sensor means whose moving speed is variable. On the other hand, in claim 3, since the scanning speed detection circuit means for detecting the scanning speed in the horizontal direction of the image sensor means is used, the horizontal direction is always correct even for images from the image sensor means whose scanning speed is uneven. The directional data sampling position can be detected and reliable data recognition can be performed. Claim 4 cites a preferable configuration of the scanning speed detection circuit, according to which, in the mesh pattern, rows of meshes that alternately repeat brightness and darkness along the horizontal direction are formed as a special pattern; The scanning speed of the image sensor means is detected by detecting the boundaries between brightness and darkness in the special pattern and measuring the time from one boundary to the other. Therefore, compared to using a device (for example, a rotary encoder) that detects the scanning speed of the image sensor means independently of one image, this method has the advantage of being inexpensive, simple in construction, and highly accurate.

[Brief explanation of the drawing]

Fig. 1 is an overall schematic diagram when the present invention is applied to the input of performance data of an electronic musical instrument, Fig. 2 is a diagram showing an example of an encoded image of a musical score recorded on a recording sheet, and Fig. 3 is a diagram showing an example of an encoded image of a musical score recorded on a recording sheet. A diagram showing a code system related to musical notes, FIG. 4 is an overall configuration diagram of a data reading device according to an embodiment of the present invention, FIG. 5 is a diagram used to explain the line width and scanning direction of a part of the sensor, and FIG. 6 is a diagram showing an example of the configuration of the control circuit section shown in FIG. 4, FIG. 7 is an enlarged view of the encoded image used to explain the operation of the control circuit section, and FIG. 8 is a diagram showing the control circuit section. A diagram showing sampled image data, FIG. 9 is a configuration diagram of the sample circuit shown in FIG. 6, FIG. 10 is a time chart of the main signals used in the control circuit, and FIG. 11 is a mesh pattern of the image. This diagram emphasizes the special pattern in , and is used to explain another configuration example of the control circuit section shown in FIG. 4. FIG. 12 is another configuration example of the control circuit section shown in FIG. 4. Figure t513 is a configuration diagram of the scanning speed detection circuit in Figure 12,
FIG. 14 is a configuration diagram of the sample circuit of FIG. 12. l...Sensor part, 2,2M...Control circuit part, 2-13.2-13M...Sample circuit, 2-14...Scanning Speed detection circuit.

Claims (4)

[Claims] (1) An image that includes a net-like pattern in which data is encoded by brightness and darkness selectively formed in each mesh, and in this image there is a mark for indicating the data sampling position in the vertical direction of the net-like pattern. image sensor means for reading the image from a recording sheet on which an image including one mark and a second mark for indicating a data sampling position in the lateral direction of the mesh pattern is formed;
vertical sampling position calculation circuit means for detecting the first mark from data of the image read by the image sensor means and calculating the data sampling position in the vertical direction; horizontal sampling position calculation means for detecting the second mark from the data of the image and calculating a data sampling position in the horizontal direction of the mesh pattern; the vertical sampling position calculation circuit means; and the horizontal sampling It is characterized by comprising data sampling circuit means for sampling data of the image at data sampling coordinates defined by the data sampling position in the vertical direction and the data sampling position in the horizontal direction calculated by the position calculation means. Data reader. (2) In the data reading device according to claim 1, the image sensor means reads the image one line at a time in a substantially vertical direction from one side of the mesh pattern to the other side, and the first mark A pair of bars are formed along the upper and lower sides of the mesh pattern, and the vertical data sampling position calculation circuit means includes an interval measuring circuit means for measuring the interval between the pair of bars, and this interval measuring circuit. dividing circuit means for calculating the data sampling position in the vertical direction by dividing the data approximately equal to the distance between the pair of bars measured by the means, and the second mark is adjacent to the left and right included in the mesh pattern. It is a vertical boundary between bright and dark meshes, and the horizontal sampling position calculation circuit means includes a mesh boundary detection circuit means for detecting the boundary, and a predetermined time period after the boundary is detected by the mesh boundary detection circuit means. A data reading device comprising circuit means for detecting the timing of the interval as the timing of the data sampling position in the horizontal direction. (3) An image that includes a net-like pattern whose overall shape is a substantially rectangular shape, and in which data is encoded by selectively forming brightness and darkness in each mesh, and in which the vertical direction of the net-like pattern is The above-mentioned image is obtained from a recording sheet on which an image including a first mark for indicating a data sampling position and a second mark for indicating a boundary position between horizontally adjacent meshes of the above-mentioned mesh pattern is formed. an image sensor means for reading each line in a substantially vertical direction from one side of the mesh pattern to the other side; and detecting the first mark from data of the image read by the image sensor means. vertical sampling position calculation circuit means for calculating the data sampling position in the vertical direction; and a boundary for detecting the boundary position by detecting the second mark from data of the image read by the image sensor means. position detection circuit means; lateral speed detection circuit means for detecting the scanning speed of scanning the mesh pattern in the lateral direction; and the boundary position detected by the lateral boundary position detection circuit means and the lateral speed detection circuit means. horizontal sampling position calculation means for calculating the data sampling position in the horizontal direction of the mesh pattern from the scanning speed detected by the scanning speed; A data reading device comprising: data sampling circuit means for sampling data of the image at data sampling coordinates defined by the directional data sampling position and the lateral data sampling position. (4) An image that includes a net-like pattern that is generally rectangular in its entirety and encodes data by selectively forming brightness and darkness in each mesh, and in this image, vertical data of the net-like pattern is included. The above-mentioned image is obtained from a recording sheet on which an image is formed, which includes a mark for indicating the sampling position, and in which at least one horizontal line of the above-mentioned mesh pattern has a special pattern of alternating brightness and darkness. an image sensor means that scans horizontally from one side of the pattern to the other side and reads almost vertical lines one by one, and detects the mark from the data of the image read by the image sensor means. vertical sampling position calculation circuit means for calculating the data sampling position in the vertical direction of each mesh of the mesh pattern; boundary detection circuit means for detecting the boundary between adjacent bright and dark meshes in the row; and detecting the boundary between adjacent bright and dark meshes in the special pattern from data of the image read by the image sensor means. a special boundary detection circuit means for detecting the horizontal direction of the mesh pattern by measuring the time from when a boundary is detected to when the next boundary is detected by the special boundary detection circuit means; scanning speed detection circuit means, horizontal data sampling of each mesh of the mesh pattern from the timing of the boundary detected by the boundary detection circuit means and the horizontal scanning speed detected by the horizontal scanning speed detection means; defined by horizontal sampling position calculation means for calculating the position, and the vertical data sampling position and the horizontal data sampling position calculated by the vertical sampling position calculation means and the horizontal sampling position calculation means. data sampling circuit means for sampling the data of the image at the sampling coordinates of each mesh.