JPS596383B2 - Diagnostic method for surface inspection equipment - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a self-diagnosis method for testing whether a surface inspection device is normal or abnormal.

Surface flaws on an object to be inspected, such as a rolled steel plate, are of course undesirable, and depending on the degree, they can make the product defective.

Each flaw is caused by a specific cause, and unless it is discovered early and the cause of the flaw is removed, a large number of products with the same type of flaw will occur, all of which will become defective, resulting in a large economic loss. This may result in financial loss. Therefore, various surface flaw inspection devices have been developed and put into practice. The applicant is Japanese Patent Application No. 51-86387, Japanese Patent Publication No. 57-35788.
One such example is the inspection device proposed in Japanese Patent Application No. 51-86387. By the way, such surface inspection equipment is installed at strip manufacturing sites such as strip mills, and is often operated in adverse environments such as high temperature, high humidity, and dust, so it is necessary to constantly monitor whether the equipment is operating normally. This is desirable.

Although there are various causes of device failure, the imaging system is relatively common. In other words, when detecting flaws, a method is used in which the surface of the object to be inspected is imaged with a television camera and the image signal from the camera is processed.If the object to be inspected is a steel plate being rolled, etc., a high-brightness strobe is used to detect the steel plate. However, high-brightness strobes have a relatively short lifespan. If the strobe reaches the end of its life and its brightness decreases, the desired flaw detection will, of course, become impossible.
The same applies if the sensitivity of the television camera decreases or malfunctions. In addition, when processing the image signal, the image signal includes a synchronization signal and an image signal of objects around the steel plate, so in order to prevent false detection by excluding these,
The side edges of the steel plate are detected, and an inspection range setting signal is used that sets the detection range to the surface area within both edges.
When ear waves occur, the reflection of light becomes abnormal, making it difficult to distinguish from the surroundings and making it impossible to detect edges. In order to deal with such a situation, the previously proposed device stores the inspection range signal obtained from the previous edge detection and uses this when the edge cannot be detected, but if the edge cannot be detected continuously. When this occurs, the inspection range signal becomes old and does not match the current situation, and an edge of the steel plate falls within the inspection range, resulting in a situation where the edge is erroneously detected as a flaw.
The present invention aims to promptly detect such a situation and issue a warning that the surface inspection is not being performed normally.
Its features include a light source that instantaneously illuminates the surface of a moving object to be inspected, a television camera that captures a still image of the surface of the object to be inspected, and a camera that determines the edges of the object from the output image signal of the camera. A circuit that outputs an inspection range setting signal from the image signal, a circuit that detects a sudden amplitude change in the image signal and outputs a change point signal, and a circuit that detects flaws on the surface of the object to be inspected from the inspection range setting signal and the change point signal. In the method for diagnosing a surface inspection apparatus comprising a circuit, an alarm is issued for an abnormality in the surface inspection apparatus when the number of times of imaging in which no abrasions on an object to be inspected are detected exceeds a certain number. Next, this will be explained in detail with reference to the drawings. FIG. 1 shows an outline of a surface inspection device already proposed by the applicant.

Reference numeral 1 denotes a high-resolution television camera, which captures a still image of the steel plate 6a when a strobe 6b illuminates the object to be inspected, which is moving at high speed, in this example, a steel plate 6a being rolled.

2 is a change point signal output circuit, 3 is an inspection range setting signal output circuit, 4 is an AND gate, and 5 is a flaw signal output circuit, which process the image signal Sg from the camera 1 to detect flaws.

An overview of flaw detection will be explained with reference to the waveform diagram in FIG. 2. Image information from the television camera 1 S8! tFigure 2 Sg
As shown in FIG. 2, it consists of a horizontal synchronizing signal Syn, an image signal S1 of the surface of the object to be inspected, an image signal S2 of its surroundings, and a flaw signal N included as the case may be. Since it is the flaw signal N that is desired, the signal Sg is first differentiated by the circuit 2, and then full-wave rectified to obtain the change point signal b shown in FIG. 2b. This signal includes signals indicating all points where the amplitude of the image signal suddenly changes, such as the synchronization signal Synl, the signal S3 which rises and falls at the edge of the object to be inspected, and the flaw signal N. Therefore, in order to distinguish the flaw signal from other signals, the edge signal is extracted from the image signal Sg in the circuit 3, and the inspection range of the pulse width T shown in Fig. 2c is narrower than the interval between the edge signals on both the left and right sides. Create a setting signal and add it to AND gate 4. If the output signal b of the circuit 2 is passed through this gate 4, the second
A flaw detection signal N can be obtained as shown in FIG. Of course, the image signal on the surface of the steel plate undergoes minute changes due to various causes, and therefore the output signal of the changing point signal output circuit 2 also includes a large number of minute pulses (not shown). Therefore, in order to output only the flaw signal, a flaw signal output circuit 5 consisting of a threshold determining means such as a Schmitt circuit is used, and a flaw signal N is obtained when the flaw signal is above a certain threshold level TH. As shown in Figure 2 Sg, the edge detection necessary for setting the inspection range is performed by using a method that shows that the level of the image signal on the surface of the object to be inspected is significantly different from the surrounding area, and this is larger than the level change due to flaws or synchronization signals. The period T, which is narrower than the width W, can be determined by signal processing such as slightly delaying the width detection (edge detection) signal and the image signal.

FIG. 3 shows an example, in which 3a is a distributor, 3b and 3c are delay circuits with the same delay time τ, 3d is a differential amplifier, and 3
e is a Schmitt circuit, and 3f is a flip-flop.
The image signal Sg from the television camera 1 is directly applied by the distributor 3a to one input terminal of the differential amplifier 3d via the delay circuit 3b, and to the other input terminal of the differential amplifier 3d on the other hand. FIG. 4 Sg shows the image signal Sg,
FIGS. 4E and 4F show the waveforms of the delayed signal e and the direct signal f of the signal Sg applied to the amplifier 3d. The signal e is a signal obtained by simply delaying the signal Sg by a fixed time τ, but since the signal input circuit is connected to the delay line 3c with one end open, the influence of the reflected wave appears after 2τ, and the signal f is as shown in the figure. As shown in the figure, there are two stages of rising and falling parts. When the amplifier 3d takes the difference between these signals e (5f), the result is a signal g with the waveform shown in Figure 4g.If this is thresholded in the Schmidt circuit 3e by setting the threshold level to the level Th shown in the figure, The signal H shown in FIG. 4h is output from the Schmitt circuit.
l, h2 are obtained. These signals Hl and h2 correspond to both edges of the object to be inspected as shown in the figure, but the leftmost signal h1 is delayed from the left edge by a delay time τ, and therefore the pulse interval of the signals Hl and h2 indicating both edges is is the width W of the incense body to be tested
narrower by τ. If the flip-flop 3f is set or reset using these signals Hl and h2, the fourth
A signal c indicating the inspection range T shown in FIG. c is obtained. In this way, the inspection range setting signal is obtained by detecting the edges of the object to be inspected, but as mentioned above, if the edges of the steel plate to be inspected are wavy, detection may not be possible. Edges cannot be detected if they come in close contact with side guides or the like that guide them.

To cope with such a case, a memory is used to set the inspection range, and when an edge cannot be detected, the previous inspection range read from the memory is used. A circuit with such a function is shown in Figure 5. In FIG. 5, 14 is a detection line setting circuit, which starts from the horizontal synchronization signal a predetermined number of times after the vertical synchronization signal VD in order to use the horizontal line portion of a predetermined height in one screen for deriving the inspection range setting signal. A signal Su having a width of one horizontal synchronizing signal is output, which continues until the next horizontal synchronizing signal.

6 is a flip-flop, 7 is a clock generator, 8a, 8
b is memory, 9a, 9b are up/down counters, 10
Reference numerals a to 10f are AND gates, 11a to 11c are NAND gates, 12 is an OR gate, and 13 is an edge detection circuit consisting of 3a to 3e in FIG.

The operation will be explained with reference to the waveform diagram in FIG.
is output, and the AND gate 10a is opened by this signal Su, and the signals H1 and h2 from the edge detection circuit 13 are output.
The flip-flop 6 is set and reset through the .

The columns Su, h, etc. in FIG. 6 show the waveforms of the signals Su, h, etc. The same applies to other signals. Therefore, the Q output of this flip-flop is between the left and right edges on the detection line (
Actually, it is turned on (shorter than this by τ), and this turns on the OR gate 12 and the AND gate 1, which is opened except during the vertical blanking period when the vertical blanking signal BL is applied.
It is output as an inspection range setting signal c through 0e.
This part corresponds to the circuit shown in Figure 3, and the flip-flop 6
corresponds to flip-flop 3f in FIG. The output signal Su of the detection line setting circuit 14 is also sent to the counters 9a and 9.
clock CLK from clock generator 7.
The counter for counting is used as an up counter, and is further input to memories 8a and 8b via the NAND game opening 1a.

Since the horizontal synchronizing signal HD is applied to the other input terminal of the Nando game port 1a, when the horizontal synchronizing signal appears, that is, at the starting edge of the signal Su, a signal n is generated which becomes 0, and the memory 8a, 8
Clear b. Q output and Q of flip-flop 6
The output is applied to one input terminal of AND gates 10b and 10c, and the output of AND gate 10a is applied to the other input terminal of these AND gates.
0b outputs one edge signal 1 of the object to be inspected, and the AND gate 10c outputs the other edge signal j. Here, one and the other edges are, if the television camera scans from left to right, one is the left edge that appears first, the other is the right edge that appears later, and these Although the signal has a time difference τ from the correct edge as shown in FIG. 4, the time difference is ignored here for simplicity. These signals 1 and j are stored in memories 8a and 8b.
and store the contents of counters 9a and 9b in memories 8a and 8.
Set to b. The contents of the memories 8a, 8b are also set to the counters 9a, 9b when a horizontal synchronizing signal is input to the load terminals of the counters 9a, 9b. Therefore, when considering the operation at the current detection line, first, the horizontal synchronization signal HD
appears, a signal Su is generated and the NAND gate 11a
The output clears memories 8a and 8b and counters 9a and 9.
The up counter setting of b is performed, and the contents of memories 8a and 8b (currently O) are loaded into counters 9a and 9b.

The loading ends at the fall of the horizontal synchronization signal HD,
Counters 9a and 9b start counting the clock CLK. Then, when one edge signal h1 (the left one in this example) is input, the AND gate 0a sets the flip-flop 6 through the signal h1, opens the AND gate 10b, and generates the signal 1, which is the content of the counter 9a. is set in the memory 8a. The contents of the counter 9a indicate the time or length from the horizontal synchronizing signal, that is, from the start of horizontal scanning until the left edge of the object to be inspected appears. Thereafter, a right edge signal H2 is output from the edge detection circuit 13, which passes through an AND gate 10a to reset the flip-flop 6, and becomes a signal j through an AND gate 10c, which is opened at this time, to transfer the contents of the counter 9b to the memory 8b. Set. The content of this counter 9b is the time or length from the horizontal synchronizing signal, that is, from the start of horizontal scanning until the right edge signal appears. Therefore, the difference between the contents of the memories 8a and 8b indicates the distance between the left and right edges, that is, the width. In this way, width detection is performed on the horizontal detection line at a predetermined height of one screen, and the detected width is stored in the memory.
From the next horizontal scanning line onward, the output signal Su of the detection line setting circuit is O, and as a result, the counters 9a and 9b are set to down counters.

Then, when the horizontal synchronizing signal HJ) appears, the memory 8a
, 8b are loaded into the counters 9a, 9b, and clock counting (subtraction) is started at the falling edge of the signal HD. In this example, since the contents of the memory 8a are smaller than the contents of the memory 8b, the contents of the counter 9a first become zero, and a signal k (this signal is referred to as the follower signal) is generated from the follower terminal F, and this signal is sent to the AND gate. The signal is applied to one input terminal of the clock signal 10d, and is inverted by the inverter 11b and applied to the counter enable terminal E, thereby stopping clock counting. Since one counter 9b continues counting (subtraction) again, the signal l at the follower terminal F is O, which is inverted by the inverter 11c to become 1, and is applied to the other input terminal of the AND gate 10d. Therefore, the gate 1
0d produces 1 output, and this 2 is output as signal c through OR gate 12 and AND gate 10e. After that, the counter 9b also continues to subtract and the content becomes O.
The signal l at the follower terminal F becomes 1, which is inverted by the inverter and applied to the enable terminal E to stop counting.
At the same time, the AND gate 10d is closed and the signal c is turned off. Therefore, the signal c output by these counters 9a, 9b, etc. is the above-mentioned inspection range setting signal,
It is used together with the inspection range setting signals from the edge measurement systems of circuits 13, 10a, and 6, and enables inspection range setting even when there is no signal from the three actual measurement systems. This operation of outputting the inspection range setting signal from the storage system is repeated every time the horizontal synchronizing signal HD is input, and when the detection line setting circuit 14 generates an output, the stored contents are updated every 4q.

By the way, with surface inspection equipment, if the side edges of the steel plate are wavy as mentioned above, the side edges may appear dark from the television camera and become indistinguishable from the surroundings, making it possible to detect edges. As a countermeasure in such a case, the method shown in Fig. 5 is to memorize the edges, but if the edge cannot be detected continuously on each screen, the contents of memories 8a and 8b will not be updated and will become very old. I end up.

Eventually, the inspection range deviates from the inner steel plate surface of the left and right edges, and the edges enter the inspection range, causing a situation such as being mistakenly recognized as a flaw. The present invention prevents this by counting the number of times that edge detection is not performed, and when the number exceeds a certain number, for example 4, an alarm is issued. Furthermore, even if edges are not detected, there is no abnormality in the inspection device as long as the inspection range does not deviate from the predetermined range on the surface of the steel plate. Considering this point, if edges are not detected and the flaw detection circuit
When the number of times that (Fig. 1) produces an output exceeds a predetermined number,
It suffices to treat this as an abnormality and issue an alarm, which is done in the embodiment of the present invention. The threshold level of the flaw detection circuit 5 is set to a fairly low level in response to the fact that the flaw signal is about an order of magnitude lower than the edge signal of 200 to 500 mV, and therefore the inspection range is shifted. If the edge signal is input to the flaw detection circuit 5, it will definitely be detected as a flaw. Therefore, since the flaw detection circuit 5 continuously outputs an output in such a case, it is possible to detect a shift in the inspection range by counting the number of outputs and issuing an alarm when the number exceeds a predetermined value. FIG. 7 shows an example of a circuit for performing such an operation. Reference numeral 20 denotes an edge detection circuit, which corresponds to the circuit 13 in FIG. 5 and outputs left and right edge signals H1 and h2.

21 is a position analysis circuit 1 for the left edge signal h1; 22
23 is a memory, 23 is a pulse generation circuit, and 24, 25, 26 are position measurement circuits for the left edge signal H2, a memory, and a pulse generation circuit.

Further, 27 and 28 are flip-flops which are set by the vertical synchronizing signal VD and left and right edge signals H1 and h2, and 29 and 30 are the timing signal Sf which is turned on between the output of the flip-flop and one field; This is an AND gate to which the horizontal synchronizing signal HPn for starting output of the inspection line signal is input. A follower one-terminal signal k of the counter 9a is obtained, a follower one-terminal signal l of the counter 9b is obtained from the pulse generator 26, and a detection range setting signal c is obtained from the inhibit gate 31 corresponding to the AND gate 10d in FIG. is output. This circuit generates the signal 1 when an edge is detected on the image detection line.
j updates the memory. Therefore, the update of this memory is taken out by the OR gate 32, and the flip-flop 33, which is set by the vertical synchronizing signal VD, is reset. Therefore, if an edge is detected and the memory is updated, the Q output of the flip-flop 33 becomes O, but if the opposite is true, the Q output remains 1. On the other hand, when the abnormality detection circuit 34 corresponding to the flaw detection systems 2 to 5 in FIG. 1 produces an output, the flip-flop 35, which is reset by the vertical synchronizing signal D, is set. The outputs of these flip-flops 33 and 35 are led to an AND gate 36, and the output of the AND gate 36 is counted by a counter 37. Therefore, the count value of the counter 37 indicates the number of screens in which no edge is detected and the abnormality detection circuit 34 produces an output. This counter outputs a carry signal when the counted value reaches a predetermined value, thereby issuing an alarm ALM to the operator. As explained in detail above, according to the present invention, the self-diagnosis of the surface inspection device, especially checking whether the edge signal for setting the inspection range is detected normally or not, and whether the inspection range is normal or not. The reliability of surface inspection results can be increased.

[Brief explanation of drawings]

Figures 1, 3, and 5 are block diagrams showing the configuration of the previously proposed surface inspection device, Figures 2, 4, and 6 are waveform diagrams for explaining its operation, and Figure 7 is the invention of the present invention. FIG. In the drawing, 6a is an object to be inspected, 6b is a light source, 1 is a television camera, 3 is an inspection range setting signal output circuit, 2 is a change point signal output circuit, and 5 is a flaw detection circuit.

Claims (1)

[Claims] 1. A light source that momentarily illuminates the surface of a moving object to be inspected, a television camera that captures a still image of the surface of the object to be inspected, and an edge of the object to be inspected that is determined from the output image signal of the camera. A circuit that outputs an inspection range setting signal from the edge, a circuit that detects a sudden amplitude change in the image signal and outputs a change point signal, and detects flaws on the surface of the object to be inspected from the inspection range setting signal and the change point signal. A diagnosing method for a surface inspection apparatus comprising a circuit for diagnosing a surface inspection apparatus, the method comprising the step of: warning that the surface inspection apparatus is abnormal when the number of times of imaging in which an edge of an object to be inspected is not detected exceeds a certain number. 2. Claim 1, characterized in that when the edge of the object to be inspected is not detected and there is an output from the flaw detection circuit, and the number of times of imaging exceeds a certain value, an abnormality of the surface inspection device is alerted.
Diagnosis method of surface inspection device described in section.