JPS58142248A - Tester - Google Patents

Abstract

PURPOSE:To reliably detect a minute defect, by a method wherein a delay differentiation of a photographing signal is conducted two times, an obtained signal is binary-coded at a different reference level, and a signal generated due to a poor product or a defect is counted. CONSTITUTION:A signal (a) from a camera 1 is amplified to produce a signal (b), and a delay circuit 3 adds a delayed signal (c) to a differential amplifier 4. If a first delay differentiation is conducted through finding of a difference between a signal (c) and a signal (b), a signal wave- form of a good product is inclined to some extent, and in case a contamination or a defect is present, it forms a wave-form which deformed in the middle of inclination. A signal (d) is added to a differential amplifier 6 and a delay circuit 5, a delayed signal (e) is added to a differential amplifier 6, and a signal (f) through finding of a difference between the two signals. Then, through comparison of it with a negative set level (g), in the case of a good product, a binary-coded signal (h) produces 1 pulse, and in the case of a poor product, it produces 2 pulses or more. An output of a comparing circuit 7 is applied to an FF 9 to obtain a signal (i). A theoretical product computation is conducted on the signal (i) and the signal (h) to obtain a pulse output. An output pulse (j) is counted to compare it with a set value A. Meanwhile, a signal (l) binary-coded by means of a positive set level (K) produces 2 pulses in the case of a good product, and 3 pulses or more in the case of a poor product, and an output (q) can be obtained only in the case of 3 pulses or more.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention is capable of stably and highly accurately inspecting the presence or absence of defects on an object to be inspected by appropriately processing an image signal obtained by imaging the object to be inspected with a television camera or the like. The present invention relates to a defect inspection device that can be used.

Generally, in order to automatically inspect defects in industrial products such as tablets, the object to be inspected is imaged by an imaging device such as a television camera, and the image signal obtained is binarized by comparing it with an appropriate threshold level. By appropriately processing this binary signal, the object to be inspected is inspected for defects, etc., but the inspection accuracy in this case depends on the accuracy of the binarization process y. For this reason, various methods have been proposed for binarization processing.

18A is an explanatory diagram for explaining such a binarization processing method.

That is, as shown in FIG.
By repeating the process in the horizontal and vertical directions, WR is obtained, and an image signal as shown in the figure is obtained. In this case, since there are five stains D on the object OB as shown in Figure (A),
In the image pickup signal in FIG. 2(B), a gentle depression as shown by the arrow is generated.

An example of a binarization method that binarizes the image signal obtained by converting the image signal is disclosed in Japanese Patent Publication No. Sho 54-3638. Unlike most conventional systems that use a fixed level method in which the threshold level for binarization is kept constant, this method uses a floating level method that changes the threshold level according to the brightness level of the image signal. However, it is clear that this method is not limited to detecting gradual depressions as described above. Note that FIG. 1 e → shows a binary signal based on a floating level. On the other hand, JP1
A method as shown in Japanese Patent No. 4854-34886 is also known. This extracts the delayed signal shown by the dotted line from the imaging signal shown by the solid line in the main part of the same figure.
(The same figure is obtained by taking the O difference between the cough slow heron signal and the imaging signal)
Obtain a differential signal such as
It is binarized by comparing it with vSN.

Therefore, in a normal case, one pulse is obtained at each of the rising and falling edges of the imaging signal as shown in (E) and (F) in the same figure, but if there is a defect, two or more pulses will be obtained, and the defect can be detected by this. do. However, even with this method, it is not possible to detect defects caused by faint stains as shown by the arrows in the figure.

The present invention has been made in view of the above, and it is an object of the present invention to provide a defect inspection device that can reliably and highly accurately detect minute signals.

The feature of this invention is to emphasize the defective part by performing the delayed differential operation twice on the imaging signal obtained by scanning the object to be inspected by the imaging means, and to compare the signals obtained by the operation with different reference levels. Binarize with
The operation of extracting only the signals caused by defects or defects from the converted signal 0 and counting the number of each is performed for the entire imaging screen. It's possible to judge, which is why I gave it a 5.

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

Iris 2gA is a block diagram showing an embodiment of the present invention, and FIG. 3 is a waveform diagram showing waveforms of various parts in FIG. In FIG. 2, l is an imaging device such as a television (TV) camera, 2 is an amplification circuit, 3 and 5 are delay circuits, 4 and 6 are differential amplification circuits, 7 and 8 are comparison circuits, and 9 - 11 are seven rib-flops (hereinafter also simply abbreviated as FP) which are reset by the horizontal synchronization signal of the imaging device 1, 12 and 13 are counting circuits which are each reset by the vertical synchronization signal of the imaging device 10, and 14 and 15 are respectively reset by the vertical synchronization signal of the imaging device 10. Digital Comparison-Route, 16
1 is a determination circuit formed by, for example, a microcomputer, 17 is a discharge section for the object to be inspected OB, ANI-3 is an AND circuit, and OR is an OR circuit.

In addition, 1 to n, p and q in the same figure indicate each part O signal,
These correspond to the symbols attached to each waveform in FIG. 3. Also,
FIG. 3(A) shows the case where the object to be inspected is a good product, (B) shows the case where there is a lightly soiled part, and FIG. 3(C) shows the case where there is a defective part.

Television (T) obtained by imaging the inspection object OB of the tablet part
V) The signal a from the camera 1 is amplified by the amplifier 2 and becomes a signal wave as shown in 0), (b) and (c) in FIG. terminal and delay circuit 3. The delay circuit 3 applies a signal C, which is the imaging signal obtained through the amplifier 2 and is delayed by a predetermined time, to the other terminal of the differential amplifier 4 . And differential amplifier 4
, perform the first delay differentiation by taking the difference between the delayed signal C and the undelayed signal S (third
(See waveform d in the figure). As a result, the signal waveform has a slope when the product is good (see figure ←), and when there is dirt (see figure middle)) or defective (see figure (c)).
) have waveforms that are deformed in the middle of the slope.

The signal d thus delayed and differentiated is applied to one terminal of the differential amplifier 6 and the delay circuit 5. The delay circuit 5 is
A signal C, which is the output d of the differential amplifier 4 and is delayed by a predetermined time, is applied to the other differential amplifier 6. Then, by taking the difference between these two signals O in the differential amplifier 6, the second Oj
1) The application signal shown by f in (a) to (c) of the same figure is obtained. Due to this two-time delay differential operation, different parts are generated in the signal waveforms of good products and defective products, so in order to extract those parts, predetermined comparison levels g and k are set in comparator circuits 7 and 8, respectively. By comparing, the signal is binarized. Note that g is set to a negative level, and k is set to a positive level. It is compared with the negative O setting level g and the binarized signal is sent to the third WJ (
As is clear from comparing A) with the same WI trunk) and (→), it is clear that for good products, it will be 1 pulse, and for non-defective products, it will be 2 or more pulses.The output of comparison circuit 70 is further fed to FF9. A different signal 1 is obtained depending on whether the product is good or not.This is because the FF9 operates at the falling edge of the pulse. Therefore, by performing a logical AND operation on the armpit signal L and the output signal from the comparison circuit RI using the AND circuit ANI, the output terminal will not receive an output if it is a non-defective product, and it will indicate that it is not a non-defective product. In this case, a pulse output of 1 pulse or more can be obtained (
(See signal j in Figure 3). Incidentally, since the FF 9 is reset by the horizontal synchronizing signal, the above-mentioned operation is performed every horizontal scan. Therefore, it is of course possible to use this result to determine the acceptability of the object to be inspected at °C, but even if it is a good item, defective signals will appear on several horizontal scanning lines, that is, pulse j will appear on the output side of the AND circuit ANIO. There may be cases where it is output. Therefore, in this embodiment, the counting circuit 1
2, the output pulses of the AND circuit ANI on one screen are counted, and the counted value is compared with a predetermined set value A in the digital comparator circuit 14, thereby ensuring greater accuracy. In other words, even in the case of a non-defective product, the defective signal j may be output, but the number of outputs is considered to be extremely small compared to the case of a defective product, so by selecting the set value A to an appropriate value, This will enable them to make appropriate judgments.

On the other hand, the signal t binarized by the positive setting level k is also processed in the same way as above, but in this case, for a good product, 2
If the pulse is a defective product, it will be 3 or more pulses.Therefore, by providing 2R flip-flops such as FF10 and FF110 for the signal t, only when the number of output pulses is 3 or more, the output from the AND path AN3, q is obtained.Therefore, except for the fact that one extra stage of flip-flops is provided, its function is exactly the same as above, so the details will be omitted.In addition, Fpt
For the oo output m, the AND circuit AN20 output n, the FFll0 output p, and the output q of the AND circuit AN3, the third
See the corresponding waveforms in the figure, the output pulse q thus obtained is transmitted through a counting circuit 13 similar to the counting circuit 12 described above.
is counted for one screen, and further compared with a reference value B in a digital ratio IIR circuit 15. In addition, in the case of a good product, it will never exceed 3 pulses, so the standard value B
can generally be made smaller than the reference value.

The outputs of digital comparison circuits 14 and 15 are OR circuits OR
is applied to the decision circuit 16 via the decision circuit 16.
Then, it is determined that the tablet OB is defective based on the output of at least one of the comparison circuits 14 and 15, and a signal is given to the discharging section 17 to discharge the tablet OB, which is the object to be inspected, to the defective product storage section.

As described above, according to the present invention, the image signal obtained by photographing with a television camera or the like is differentiated twice, and the result is binarized by comparing it at the positive and negative setting levels, respectively.
The binarized signal is subjected to a predetermined operation to extract only the resulting signal and count it in °C, which eliminates minute defects such as yellow stains on white that could not be detected with conventional methods. Certain objects can be detected reliably.

Note that the present invention can also be applied to industrial products such as tablets, caramel, and chocolate as described above. It can also be applied to visual inspection of II products such as mandarin oranges and apples.

[Brief explanation of drawings]

Fig. 1 is an explanatory diagram for explaining the method of binarizing the object to be inspected, Fig. 2 is a block diagram showing an embodiment of the present invention, and Fig. 3 is a waveform diagram showing waveforms of various parts in Fig. 2. . Description of symbols 1... Imaging device, 2... Wide narrow path, 3
, 5... Delay (9) path, 4, 6... Differential amplification circuit, 7, 8... Ratio 1 [jialNr,
9～11・・・Flip-flop, 12°13・
... Counting circuit, 14, 15 ... Digital comparison circuit, 16 ... Judgment path, 17 ...
...Discharge Department, OB...Inspected object agent Patent attorney Akio Namiki Agent Patent attorney Kiyoshi Matsuzaki 111 Figure 1

Claims (1)

[Claims] The object to be inspected is made horizontal by the imaging means! & an eleventh differential amplification means for calculating the difference between an imaging signal obtained by direct scanning and a transmissive imaging signal obtained by uniformly enhancing the #imaging signal, and an output signal of the amplification means and a scissor output signal. a 1st and 20th differential amplifying means for calculating the difference between the electricity-transmitting output signal and the electricity-transmitting output signal;
a first and a twentieth system that binarizes the O output from the differential amplification means by comparing it with mutually different reference levels;
a comparing means, and a first and second comparing means for extracting only a signal 0 caused by a defective portion on the object for each horizontal scan by performing predetermined processing on each output from the first and second comparing means, respectively. 1 and 20 extraction means, and first and second counting means for respectively counting the number of output signals from the first and second extraction means over the entire horizontal scanning period, This is an inspection device that determines the quality of the object to be inspected based on the count output.