JPS62100093A - Processing device for serial data - Google Patents

Abstract

PURPOSE:To discriminate the quality of goods by providing a window area in a visual field to watch the goods to be conveyed, counting the number of picture elements as a binary output by image picking-up the goods with the image pickup device and comparing with the reference value. CONSTITUTION:In a visual field S to watch the goods to be conveyed, three windows W1, W2 and W3 are specified. An X coordinate data storing device 51 and a Y coordinate data storing device 52 respectively store the X coordinates and the Y coordinates of the respective windows. A window area output means 54 receives the output from a coordinate comparing means 53, respectively obtains the 'AND' of X1, X2 and X3 coordinate data from the storing device 51 and Y1, Y2 and Y3 coordinate data from the comparing means 53 and sends it to a counting means 55 as the window area signal. The counting means 55 examines the serial scanning data outputted from one-dimensional line sensor 4 for one scanning, detects and counts the picture data in it and sends them throughan adding means 56 to a holding means 57. The total value classified by respective windows of the holding means 57 is compared with the reference value classified by respective windows prepared separately and decides the quality of the goods.

Description

[Detailed Description of the Invention] [Industrial Application Field] According to the present invention, 12 articles are transported within a desired field of view by a single transport device such as a n-hair. 7.1: "1" in the imaging output (binarized output) obtained by setting E or a plurality of window areas (windows) and imaging the article with an imaging device through the window area The present invention relates to a discriminating device that determines whether the article is good or bad by counting the number of pixels that take a specific logical value of "l" and comparing the counting result with a pre-prepared reference value. It is a thing,
More specifically, the discriminating device does not use a two-dimensional imaging device such as a television camera as an imaging device, but a one-dimensional imaging device such as that used for facsimile transmission scanning.
The present invention relates to a processing device for processing serial data output from a dimensional line sensor to obtain an output corresponding to a count result to be compared with the reference value when a dimensional line sensor is used.

[Conventional technology]

2. Description of the Related Art In recent years, the application of imaging devices has been increasing for factory automation, which uses an imaging camera to inspect the appearance of products (articles) and automate product quality determination, defect detection, and the like. In this kind of product inspection (determining the quality of an article), the method is to use a two-dimensional camera and set a judgment area called a window (window area) at each vital part of the image that represents the "product" to be inspected. When the inspection target is (
The area within the window is counted and compared with a determination reference value set for each window to determine pass/fail.

According to this method, it is possible to freely select the area for feature extraction by arbitrarily changing the window size and number of positions, etc., so it has a wide range of applications and can flexibly respond to changes in the inspection target. Since each part of the image can be judged independently, it is possible to classify different types of products and inspect only common parts.

Furthermore, in order to compare the area with a reference value, if the size of the underside of the window is determined appropriately, it can be applied to the inspection line without worrying about slight misalignment between the image and the product or small dust.

An example of this window image processing is shown in FIGS. 2 and 3.

For the purpose of checking whether the characters rAJ are printed correctly, FIG. 2 shows windows Wl, W2. An example in which W3 is set is shown. A) is an example that is printed correctly, but in b), the center part of the letter A is filled in, so the area measurement value of window W2 exceeds the standard value, and the lower left of the letter A is missing. Therefore, the area measurement value of window W3 is less than the reference value, and therefore it is considered to be defective from both of these points. In example C), window W2. Although the measured value of the window W3 is correct, the area measured value of the window W1 exceeds the standard value because there is a drop of ink on the upper right corner of the letter A, and is also considered to be defective.

FIG. 3 is an explanatory diagram showing an example of classification and identification of characters rEJ rFJ rLJ, and windows Wl and W2 are set. If the area measurement value of both windows Wl and W2 exceeds the standard value, "E", if only window W1, "F"
, if there is only window W2, it can be determined to be "L".

Conventional processing equipment that uses such a method uses a two-dimensional imaging camera to extract a predetermined area in a two-dimensional image, but it is still expensive and although the method is general-purpose, it cannot be used in general factories. It is not very popular in Japan. One reason for the high price is that it uses a 2D camera (2D imaging device), and the camera body, its scanner, and the memory that stores 2D images are major obstacles to lowering the price. There is.

[Problem that the invention seeks to solve]

Therefore, the present invention uses a one-dimensional line sensor, whose price has rapidly decreased with the spread of facsimiles in recent years, in place of a two-dimensional imaging device, and uses a two-dimensional imaging device with the transport direction of the transported article as another dimension. The problem to be solved is to construct a window method similar to the case using the above method at a low cost, and an object of the present invention is to provide a serial data processing device that is useful for this purpose and is inexpensive and easy to spread.

[Means and effects for solving the problem] In the present invention, a pulse train having a pulse interval inversely proportional to the speed of the transport line that transports the article (that is, if the speed of the transport line is cylinder speed, whether it is high cycle or low speed), (e.g., a low-period pulse train), or if the speed of the conveyance line is considered to be sufficiently constant, a two-dimensional image is created with a normal equal-period clock as one dimension and the scan direction of the -dimensional line sensor as the other dimension. , the plurality of set windows are identified by coordinate data provided along the transport direction, and the total number of pixels for each scan corresponding to the area within the window is calculated by calculation processing. Therefore, instead of an expensive two-dimensional imaging device, we used an inexpensive one-dimensional line sensor to realize a window method similar to that of a two-dimensional imaging device, requiring only an image memory for -scanning. It makes it possible.

The serial data processing device according to the present invention is a device for calculating the total number of pixels corresponding to the area within the window from the serial output data of the primary line sensor through arithmetic processing as described above.

DESCRIPTION OF THE PREFERRED EMBODIMENTS The basic configuration and operating principle of the serial data processing apparatus according to the present invention will be specifically explained with reference to FIGS. 1 and 1A.

Figure 1: Figure 1A shows the basic configuration of the serial data processing device according to the present invention.
This is an explanatory diagram of the coordinates of the ``find area.''

For convenience of explanation, let us first begin with the explanation of the figures.

(In the field of view S where you want to see the object you are looking into,
Assume that we have determined W3. Then, JJ in the Y direction, which is the direction of the conveying force of the article.
Y coordinate interval Yl (between ω and ◎) of each window, Y2
((0,': between O>, Y3 (between ■ and O)
and X coordinate sections XI, X2 . X3 will be obtained as 2 coordinate data.

Dimensions Refer to Figure 1. In the figure, an X coordinate data storage device 51 stores X coordinate data X1 . X2.
X3 is recorded (g). Also, Y coordinate data recording device 5
2 is the Y coordinate data Yl, Y2 . I'm using Y3 as a mud.

1, the coordinate data comparing means 53 receives a pulse train representing the conveyance speed of the article conveyance line generated by a means not shown, counts this pulse train, and calculates the pulse train between the article (the visual field S) and the one-dimensional line sensor 4; If we consider that the object 1, that is, the field of view S, is stationary and the sensor 4 is moving in the Y direction, calculate the relative distance 1 of the sensor 4 (Y coordinate), and calculate this and the memory. The Y coordinate data of each window from the device 52 is compared, and as long as the two match, an output is generated and sent to the window area output means 54.

In addition to receiving the output from the coordinate comparison means 53, the window area output means 54 also receives the X coordinate data from the device 51, and outputs the logical product of the two, for example, the Y1 coordinate data from the comparison means 53. The AND of the The AND is sent to the 8A number means 55 as a window area W3 signal.

The counting means 55 examines the serial scanning data output from the one-dimensional line sensor "IJ4" only while the window area signal is being input, that is, if the window Wl is selected, the counting means 55 examines the serial scanning data outputted from the one-dimensional line sensor "IJ4" by one scanning within the range of the area Wl. , detects the image or data (“l” if it is logical “l”) therein, calculates it, and calculates the number by calculating means 5.
6.\Send.

The adding means 56 adds up the counting results of the pixel data included in each scan within the range of each 1" Indian region from the counting means 55 for each Indian region. ,
The total is held by the holding means 57.

In this case, the window areas Wl, W2 . W3 is set so that even if it is only a partial overlap, there will be no overlapping part. Therefore, the holding means 57 is set so that the sum of the pixel data obtained for each area is held. Kotokonaru 4
Therefore, it is possible to prepare a separate set of metalwork values for each Indian metalwork and compare them with the basic ten values for each window.

The basic configuration and operating principle of the serial data processing device according to the present invention are as follows.

[Example〕 Next, an embodiment of the tree fight will be explained with reference to this figure.

FIG. 1 is a perspective view showing an example of the configuration and arrangement of a serial data processing apparatus according to an embodiment of the present invention and 7 applied to product inspection.

Refer to the same figure. A product 2 conveyed by a heald conveyor 1 is illuminated by a lighting device 3.

The field of view 5 of the one-dimensional line sensor camera 4 is arranged at right angles to the traveling direction 6 of the heald conveyor 1, and the one-dimensional image of the product 2 is sent as serial data to the serial data processing device 7 according to the invention.

The position sensor 8 detects the product 2 and sends ONL and ON-OFF signals to the serial data processing device 7 only while the product 2 passes through the position of the field of view 5. The row encoder 9 sends a pulse train having a pulse interval inversely proportional to the traveling speed of the belt conveyor 1 to the serial data processing device 7.
Since the coordinates of the belt conveyor traveling direction 6 can be known, this allows the predetermined window area on the product 2 to be 1
Know the period during which the dimensional sensor camera 4 is passing through the field of view 5, and count the number of significant bits in the output of the camera 4 (pixel signal of "1" if the logic is "1" in the binary output) during this period. , and can measure an amount proportional to the area of a predetermined window area on the product 2.

After the product 2 has passed, the serial data processing device 7 compares the measurement results for the window area with the reference value to determine the product. If the product is determined to be defective, it sends a signal to the air cylinder driver 10 to move the air cylinder 11. The product 2 is pushed out toward the defective product conveyor 12.

FIG. 5 is a block diagram showing an example of the hardware configuration of the serial data processing device 7 as an embodiment of the present invention.

In the figure, the image sensor driver 13 sends a trigger signal TR indicating the start of scanning and a transfer lock CLK1 to the one-dimensional line image sensor 14.
Analog image No. 13 ■S of product 2 imaged above is output. The analog image signal VS is constant and has a threshold voltage of 2
The signal is input to the serial data processing device 7 as a binarized image signal VVS.

[)MA (Die Ref 1 - Memory Access) Controller 16 receives DMA request REQ from I/O port 17
Outputs a DMA response DMAR in response to the DMA,
Switch the buffers 18 and 19, switch the bus to the DMA controller 16 side, and transfer the binary image signal VVS to the V I
DMA transfer to DEO-RAM20.

The output signal (pulse) of the rotary encoder 9 is CP
The pulses are input to interrupt terminals INTL and INT2 of U21, and the number of pulses from the encoder 9 is counted by interrupt processing. In addition, the CPU 2 t sends the 0N-OFF signal of the position sensor 8 (SAMP signal indicating whether or not the product 2 has arrived) to i/
It can be read through the o boat ef, and the output is the same (it is output as an ACT signal through the i10 port 17 and can drive the air cylinder 11.

In addition, the CPU 21 can receive input of necessary data (ie, coordinate data of a set window area, etc.) from the keyboard input device 22, and performs processing according to a program written in the ROM 22 using the general-purpose RAM 23 as a memory.

FIG. 6 is a timing chart of the main parts of the signals in FIG. Please also refer to the same figure.

The one-dimensional line image sensor 14 uses the trigger 24 as the start signal TR and has a total number of pixels of 2.04 for 1947 minutes.
An analog image signal 25 (VS) of 8 is output, and a binarized image 26 is inputted to the DMA controller 16 as ■■s. When CP[J21 requires binary image data VVS, it outputs a pulse of DMA request 27 (REQDMA) through the I/O boat 17, and at the falling edge of the pulse, a DMA response 2B (CDMAR) is returned, and one line is sent. Data transfer of 2.048 pixels is started.

The binarized image signal 26 (VVS) is transferred to the transfer lock 29 (
CLKI) + 7 V I D E O・RA
The DMA is transferred to M2O, and the data is transferred to M2O.
), it is determined whether the binarized image signal 26 (VMS) is at a high level or a low level, and V I D is determined accordingly.
The image data is stored in the image memory 30 (not shown in FIG. 5) in the EO-RAM 20 as shown in FIG. 6Q.

Figure 7 is shown in Figure 5.1. FIG. 2 is an explanatory diagram showing a specific example of a memory configuration in an embodiment of the present invention.

Each memory area 1 shown in FIG. 7, RAM in FIG.
It is located at 23.

The serial data buffer 3OA reads the DMA-transferred serial data from the VI DEO-RAM 20, and
This is an area for buffering in AM23.

The window X coordinate data storage area 3I stores the coordinates of the start point and end point of each window with the scanning direction of the one-dimensional line sensor I4 as the X coordinate, and stores the number of pixels for one line of the one-dimensional line sensor 1, ↓. The items expressed in order are listed below. Based on this X coordinate data, the data included in the window area is extracted from the input serial data, and the number of significant pixels (pixels of "1" if the logic is "'1") is counted.

In the window's X coordinate data storage area 31, X (
N, 0) is the X coordinate of the starting point of the Nth window (opening point) (the order of the number of pixels in one line l'j of the line sensor 14) and the 1-5SB side. This is the area where the coordinates of
1) There is (A1, also No. N 1.1 Win F'
X coordinate (1 of the line sensor 14) of the ending point of
This means that it is an area where the arrangement of the number of pixels for a line (coordinates on the MSB side in the song) is recorded.

The window X-coordinate data storage area 32 is filled with the number of output pulses from the I:1-Tari encoder 9. This is the area where the X coordinate data of the 7th window is recorded, and the same area 3
2, Y (N, 0) means that it is an area that stores the Y locus IL of the starting point (opening point) of the Nth window, and Y (N, 1) means that it is the 1i[\\ area that also stores the X coordinate of the end point (closing point) of the Nth .:Nouille...'.

The judgment data/data storage area 33 is an area for storing reference value data to be compared with the measurement results for the judged country India. Same area 3.
In 3, REG (N) and THE (N) mean that the measurement result for the Nth win is good if it is within the Ln range of REG (N) i TRE (N), and if it is outside the range. If present, this data means that it is defective. In other words, it can be said that REG (N) is data representing the reference value for the Nth India, and TRE (N) is data representing the allowable error.

The work area 34 consists of work areas A to L, and indicates each work area used when the CPU executes a processing operation, A71J to 1. The subscript (identification symbol) is used as a subscript when repeating the same operation.

The X-coordinate organized data storage area 35 stores the Y-coordinate data stored in the window's X-coordinate data storage area 32,
They are rearranged in order of size, regardless of the window number. In FIG. 1A, the X coordinate of the starting point of window W1 is ■, the X coordinate of the ending point is ◎, and window W2
The X coordinate of the starting point of window W3 is ■, the X coordinate of #%-j'' point is ■, the Y coordinate of the starting point of window W3 is ■, the Y coordinate of the ending point is
(to), but in this case, in the area 35, the size is 11 inches, -), ■, ■, 0. m4 of ○, ■, θ
, two rearranged 'te'1゛Y (1), 'TY (2),
T Y (3), ...T Y (2N).

Now, suppose that the total S value of the output pulses of the ILI-Tari encoder 9 reaches "T'Y (1)", that is, U7 of ■ in Figure 1A, then in this case, the win F'WI opened this evening at Iminge Therefore, when the output pulse count value of the dimensional rotary j=encoder 9 reaches TY(2), that is, the circle in Fig. 1Δ, the window W2 opens at this timing, and in the same way, the output pulse count value becomes TY(3). When this timing is reached, the window W1 is closed, and in the same manner, each inn 1 is opened or closed in order.

Therefore, in 0CW(1) of the opening/closing number data storage area 36, the area 35! 7, corresponding to the TV (1) in the window, remember the number, in this case Wl, and also in the open/close data arrangement area 37 (occi), the window with the Indian number W1 is stored. This means that data for open or closed, in this example, open data will be stored.

The window work area 38 is an area that stores the current opening/closing status of each window, and is modified every moment based on the data in the opening/closing data storage area 37.

The addition work area 39 consists of N areas, and C
When the PU 21 looks at the window work area 38 and learns the open window, for example Wl, it looks at the dimensional window's The number of existing pixels included in the window area Wl for each scan is detected and counted, and the result is displayed in the work area 39, for example, in the AR
It is added to E A (1).

The dimensional timer area 40 has a counting area T I MER(1
) and TIMER (2), the former is an area for counting the output pulses of the row encoder 9, and the latter is an area for counting the clock CLK 1 that drives the one-dimensional line image sensor 14.

The program included in the ROM 22 in FIG. 5 will be explained in accordance with the flowcharts in FIGS. 8 to 14.

FIG. 8 shows the main program, and this routine is normally repeated by the CPU 21.

It is checked in step (1) whether there is any input from the keyboard input device 22. If there is any key input, the process proceeds to Step Nob (2), and after reading input from the keyboard in a keyboard input processing routine as a subroutine, the process proceeds to Step (3).

If there is no keyboard input in step (1), jump to step (3). In step (3), the output signal SAMP of the position sensor 8 is read, and the previously read SAM
Compare with the P signal (step 4).

As a result, in step (5), it is determined whether the previous SAMP signal was low level and the SAMP signal read this time is high level, and if so, it is determined that the SAMP signal has risen,
It is determined that the product 2 has just entered the field of view 5, and the process proceeds to step (6), where an initial processing subroutine is executed to perform the initial setting necessary for future window measurements. If the SAMP signal does not rise,
Proceeding to step (7), it is determined whether the SAMP signal read this time is at a high level.

As a result, if the SAMP signal read this time is at a high level, it means that there is a product to be measured within the field of view, so the process proceeds to step (8) and a measurement process as a subroutine is executed. If the level is low, proceed to step (9), and if the previously read SAMP signal is high level, product 2
It is determined that the timing has just passed the field of view 5, and the process proceeds to step (10), where the measurement results stored in the addition work area 39 and the data recorded in the judgment data storage area 33 are compared, and whether the product is good or not. Execute subroutine determination processing to determine defective products.

If it is not a falling edge, it means that product 2 is completely out of sight, and the process proceeds to RET (return). R.E.
After T, perform other processing (omitted as it is not related to the main point of the present invention), return to 5TART, and then proceed to step (1).
) to (10) are repeated.

FIG. 9 shows a flowchart of keyboard input processing as the subroutine described above.

In step (11), the window coordinates of the number are read into the X (k, J), Y (k, J) areas of the storage area 31.32 (where = 1.2...N, J = 0. 1)
.

The contents of each data are as described above with reference to FIG.

The process proceeds to step (14) of reading the comparison reference value and tolerance stored in the judgment data storage area 33 in the dimension steps (12) and (13). In step (14), it is determined whether the key input has been completed and all necessary data has been input. If it has not been completed yet, return to step (11) and repeat steps (11)-(12)-(13). Manipulate back. Once all necessary data has been input, proceed to step (15).

In step (15), all values of Y (k, J) are examined over the range of = 1°2...N, J = 0.1, and Y (
k.

Assign K from 1 to 2N in descending order of Y)
Y (
k, J) in the area OCW of the opening/closing number data storage area 36.
k in (K), area QC (K) of opening/closing data area 37
, the value of J is memorized. Therefore, TY (K)≦TY(
The relationship is K+1>.

When the output pulses of the rotary encoder 9 are counted by the interrupt input lNTl of the CPU 21 and the value becomes T)'(k), the window of the window number OCW (k) is
If k) is O, it is closed, that is, it is the end point of the window,
If QC(k) is 1, it can be determined that it is open, that is, the starting point of the window.

FIG. 10 shows a flowchart of initial processing as a subroutine. Initial processing is performed by performing seven initial steps immediately before measurement. In this state, all windows are closed, so the window W (k
), O indicating that the window is closed is substituted into the window work area 38 for all values from =l to N. A of the addition work area 39 indicating the measurement area of ``W(k)''
0 is also assigned to REA (I<,) as the initial bribe value.

In step (T7), the timer area 40117) TI
In I~4El, set O to (1) and TIMER (2) as Bricent. In step (18) storage area 35 v
-) 1゛Yck) to the subscript in the work area 34 that specifies the value of (1). L and ゛゛ru.

In step (19), the interrupt input 1NTI is enabled in order to count the pulse input from the low reencoder 9.

FIG. 11 shows the interrupt processing routine INT1 of INT1. In this routine, every time one output pulse from the encoder 9 is input, i N T1-''(20
)...(28) -RET processing is performed only once. At Step A (20), TIMER (1) is raised from 1 to 1. In step (21), T+MEI
Check whether Ma(1) is equal to TV(A). A',! : Subroutine initial processing (Fig. 10) ζ, J, is initially set to 1.

If 1" I MER(1) = TV' (A) T:
If so, go to x-i-tub (22)-, count up A by 1, and write 7 W (OcW (A))? , :open or close is (II'). When it becomes equal to either, the Y coordinate data is the data of the window number (OCW (A) 1), and if the data QC (A) is 0, it is the end point of the window, and it is 1. If so, it means that it is the starting point of the window, and the area W (OCW (
A) This is a process of reducing l by seven times.

After this, return to step (21) and perform T I M R (
1) Determine whether =TY (A). The value at this time is +1 at step (22) compared to the previous determination at step (21). , 2 (21) - (2
2) - (23) are repeated in the Y coordinate data storage area;
(This is because there is a possibility that there are equal values among 2, and if all of these equal values are processed, the determination in step (21) becomes NO, and the process proceeds to step (24).

At Ste Nobu (24) (Yo Step (25) - (26)
-(27) - In order to repeat -(28) N times, set the subscript M of the work area 34 to the first gJl straight 1. Ste,
・6+ to determine whether W(M)-〇 with Tsubu (25)
, V(M) is either O or 1, and if W(M) = 1, it means that win t'M is open, and the judgment is NO, so execute step (26). do. If W (M) = O, then Sue) ~ Notsubu 27) Hejani/Fish M +
1 and 7 and Stenov (28) N times (25) - (26
)-(27)-(28) have been executed, and if they have been executed N times, l"S FF.

Proceed to T to finish 4-nable-chin. If it is completed and not completed, proceed to step 25 and proceed to the next M1. We will process your request.

In step (26), the one-dimensional line image sensor 1
4 is stored in the area VVSB where the current scanning measurement data is stored, and from the window X coordinate data storage area 31 to the window M significant pixel width X (M, I) ~
Knowing (M, O), it is a significant pinot between the boxes!・Count the number, count the area of window 1, AREA (M
). By this processing, the Y-coordinate interval of 21-Ya in the direction of movement and the Y-coordinate interval of 14())i in the one-dimensional direction of the window defined by the X coordinate interval, 11. The number of significant pins within each win is counted.

Note that VVSB (k) is data placed in the RAM 23 (however, = 1.2...2048)
,■I VV transferred to DEO・RAM20 by DMA
This is a copy of S (k) as is in a subroutine (a program within the measurement process shown in FIG. 12, which will be described later). This is to enable the next DMA transfer to be performed during processing.

Figure 12 shows the subroutine (flowchart of measurement processing).
shows. During measurement processing, measurement data is always transferred to VID using DMA.
The data is imported into the EO/RAM 20 and then copied to the RAM 23.

In step (29), the DMA request signal REQDMA is
A pulse is output to the MA controller 16. At the fall of this pulse output 27, the DMA controller 16 sets the DMA response response signal to the DMA cylinder level and holds the high level until the DMA data transfer is completed, so wait until DMAR becomes low level in step (30). Proceed to step (31).

In step (32), set V V S (I) to 1=1 to 20.
48 to V V S B (+), but in order to avoid that lNT1 is input during the copying and read out before V V S B (+) is finished changing, the step (31 ), and after copying is completed, interrupts are enabled in step (33). The reason why the process is made to wait for a certain period of time in step (34) is to thin out DMA transfers and balance interrupt processing.

FIG. 13 shows a subroutine (flowchart of determination output processing). Since this routine is called immediately after the product 2 passes through the field of view 5, the interrupt 1NT1 is disabled in step (35) in order to stop the measurement. In step (36), the following steps (37)-(38)-(3
9) is repeated N times, the subscript ■ is set to the initial value 1. In step (37), the wind measurement data A RE A (1) of the passed product 2 is obtained (where I=1.2...
...N), the standard clay tolerance (REG(1)
±TRE (1)).If not, it is determined that the product is defective and step (40
) to jump to.

If it is, go to step (38) and increase ■ by one until ■ exceeds N (step 39).
Continue the judgment in 37). Only when all the window measurement values are determined to be good, the process advances to RET and the subroutine process ends.

In step (40), a constant is set in TIMER (2). This is the waiting time until the product 2 is transported by the belt conveyor and reaches the front of the air cylinder 11, and T
1NT of interrupt 2 to count I M E R (2)
2 is allowed (step 41). This constant is greater than 1.

FIG. 14 shows a flowchart of the interrupt processing INT2.

In step (42), it is determined whether TIMER(2)=O or not, and if it is 0, the process proceeds to step (45) and interrupt INT2 is prohibited. After this, the program (13th The interrupt 2 routine is not executed until 1NT2 is enabled in Figure 2).

In step (43), it is determined whether TIMER (2) - 1. If it is 1, the air cylinder is driven (step 44), and in step (46), TIMER (2) is determined.
) to count down. If it is not 1 in step (43), it is determined that product 2 has not yet been carried to the position in front of the air cylinder, and the process jumps to step (46).

〔Effect of the invention〕

As explained above, according to the present invention, the two-dimensional imaging camera is replaced with a one-dimensional line image sensor, and
A two-dimensional window area is constructed by the window X coordinate data stored in 1.0 and the window Y coordinate data recorded in area 32 and 1.0, and the X coordinate data group in area 31 is Since the configuration is configured such that predetermined X coordinate data can be selected from among them and addition processing of significant bits can be performed according to the output J pulse from the rotary encoder 9, two-dimensional window image processing can be performed using an inexpensive one-dimensional image sensor. It became so.

In this case, the rotary encoder 9 is not necessarily required, and if the conveyor 2 moves at a sufficiently constant speed, it can be replaced with a fixed-cycle clock.

Furthermore, the SAMP signal indicating the entry of the product (article) into the field of view is not necessarily necessary; for example, it is determined from the binarized image V V S (J) that the product A is reflected in the field of view 5, and it is regarded as the SAMP signal. You can also do that.

Furthermore, the conveyor movement direction may not be perpendicular to the one-dimensional line sensor, and the conveyor movement may not be linear. For example, if a conveyor draws a semicircle, the shape of the window can be expressed in polar coordinates, making it possible to create windows that cannot be easily created using a two-dimensional camera.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing the basic configuration of a serial data processing device according to the present invention, FIG. 1A is an explanatory diagram of the coordinates of a window area within the field of view, and FIGS. 2 and 3 are explanatory diagrams each showing an example of window image processing. , FIG. 4 is a perspective view showing a configuration and arrangement example when an embodiment of the present invention is applied to product inspection, FIG. 5 is a block diagram showing a hardware configuration of an embodiment of the present invention, and FIG. The figure shows a timing chart 1 of the signals of the main parts in FIG. 5, FIG. The figures are flowcharts showing the programs stored in the ROM, and FIG. 8 is a flowchart for the main programmer, FIG. 9 is a flowchart for keyboard input processing, FIG. 10 is a flowchart for initialization processing, and FIG. 11 is a flowchart for initialization processing. is a flowchart of interrupt processing lNTl,
Figure 12 is a flowchart of measurement processing 1, Figure 13 is a flowchart of judgment output processing, and Figure 14 is interrupt processing ■N.
To flowchart 1 of T1. Explanation of symbols 4... One-dimensional linear sensor camera, 7... Serial data processing device, 16... DMA controller, 2o...
・V! DEO-RAM, 21-CPU, 30-...
Image memory, 3I... Window X coordinate data storage area, 35... Y coordinate arrangement data storage area, 38...
Windwork area, 40...Timer area agent
Patent attorney Akio Namiki Agent Patent attorney Kiyoshi Matsuzaki MIA Figure 11711J 118 Ro 9m Earthquake Figure 50 Figure j 112 Figure II 13WJ 114a

Claims (1)

[Claims] 1) For an article that is moved relative to an imaging device along a certain direction (hereinafter referred to as the Y direction), a window area is set within the field of view in which the article is desired. a serial data processing device for obtaining the imaging output by the imaging device in the window area of the article, and comparing this with a certain reference value to determine the quality of the article; Scanning one line along a direction intersecting the Y direction (hereinafter referred to as the X direction), sequentially decomposing the image of the article into pixels, and converting it into a time series binary signal. , a one-dimensional imaging device as the imaging device outputting as serial data, a first storage means for storing in advance the X-coordinate interval of the window area, and a second storage means for also storing in advance the Y-coordinate interval of the window area. storage means; monitors changes in the relative position in the Y direction between the article and the one-dimensional imaging device, and the Y coordinate of the imaging device is stored in the window area stored in the second storage means; When it is detected that the window area is in the Y coordinate interval, it learns the X coordinate interval of the window area from the first storage means;
Among the serial output data from the one-dimensional imaging device,
means for extracting and examining data within the X-coordinate interval of the window area, counting data that takes a predetermined logical value for each scan, and sequentially adding the data to create an imaging output for the window area; A serial data processing device characterized by: 2) In the serial data processing device according to claim 1, the window area is made up of a plurality of window areas, and the creating means creates an imaging output for each set window area. A serial data processing device characterized by: