JPH0481228B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a binarization method for binarizing a video signal from a camera.

[Conventional technology]

Recently, when performing visual inspections of workpieces, a video camera is used instead of human vision to image the workpiece, and the resulting video signal is digitally processed to enable visual inspection of the workpiece within the field of view of the camera. The size, location, number, etc. of scratches, dirt, etc. on the contained workpieces are identified.

An example of such a visual recognition device will be explained with reference to FIGS. 4 to 8. First, a line sensor camera (hereinafter simply referred to as a camera) 1 detects a broken line on a workpiece 2 illuminated with illumination (not shown). When scanning the imaging range (imaging field of view) 3 shown,
For example, if there is a scratch 4 that creates a shadow within the imaging range 3, the camera 1 will only detect the image level (signal level) of the pixel corresponding to the scratch 4, as shown in Figure 5, depending on its size and shape. Accordingly, the video signal _S1 , which has dropped sharply, is output.

In the visual recognition device 5 that receives this video signal _S1 , the binarization circuit 6 compares the video level of the video signal _S1 with a predetermined threshold level (binarization level), and determines the video level. If the video level is higher than the threshold level, it is "1", and conversely, if the video level is lower than the threshold level, it is "0". S ₂
By processing this in the next stage recognition logic 7, the size and position of the scratch 4 or the number of scratches 4 if there are multiple scratches 4 are recognized, and the recognition result is outputted to the outside as a signal _S3. There is.

Here, conventional examples of the binarization circuit 6 include a fixed threshold type (fixed binarization type) and a threshold integral type (floating binarization type).

In the case of a fixed threshold type, as shown in Fig. 6, the video signal _S1 arriving from the camera 1 connected to the connector 8 is input to the non-inverting input terminal of the comparator ₉ via the terminating resistor R1. be done.

A required fixed threshold level Src is applied to the inverting input terminal of this comparator 9, and the output side is connected to a pull-up resistor.
Connected to DC power supply +5V via R ₂ .

Therefore, the video signal _S1 is compared with the threshold level Src, and converted into a logic signal that becomes "0" only during the period when the video level of the video signal _S1 sharply drops, and the converted logic signal is sent to the inverter. 10 and output to the next stage recognition logic 7 as a logic signal _S2 .

The threshold level Src in this case is formed by dividing the voltage of the DC power supply E by a variable resistor VR ₁ , and is used by adjusting it according to the video level of the video signal S ₁ , but its waveform is As shown in FIG. 8, the voltage is constant and has a fixed threshold waveform.

On the other hand, in the case of the threshold integral formula, the seventh
As shown in the figure, the difference from the fixed threshold type shown _{in FIG} .
All other aspects are the same, and of course the operation of the comparator 9 is also the same.

The integrator circuit 11 is an integrator circuit with a relatively small time constant consisting of, for example, a resistor R ₃ and a capacitor C. This output voltage is divided (level adjusted) by a variable resistor VR ₁ to set the threshold level of the comparator 9. Used as Srm.

In this case, the waveform of the threshold level Srm changes in signal level following the video signal _S1 , as shown in FIG.

[Problem that the invention seeks to solve]

However, in such a conventional binarization circuit, the threshold level of the comparator 9 is either a fixed type or an integral type (floating type), which causes the following problems.

In other words, when the illuminance within the field of view is not constant,
If the peripheral light intensity of the lens, which is the optical system of camera 1, is drastically reduced, parabolic shedding occurs in the video signal _S1 as shown in Figure 8, but when this parabolic shedding occurs, both methods are effective. There is a problem that the fields of view ES ₁ and ES ₂ are narrower than the shooting range 3 of the camera 1.

In addition, in the case of a fixed type, there is a risk of false detection when the illuminance within the field of view changes, while in the case of an integral type, if the scratch 4 to be detected is large, the video signal S ₁ changes significantly and the threshold level
Srm also changes greatly accordingly, but such a large change not only makes it impossible to correctly recognize the size of the flaw 4, but also causes the problem that detection becomes unstable in general.

This invention attempts to solve such problems.

[Means for solving problems]

Therefore, in the video signal binarization method according to the present invention, a video signal output from a camera is integrated by an integrating circuit, and the integrated signal or a signal corresponding to the integrated signal is sequentially extracted via a sample and hold circuit. Then, the reference level based on this sequentially extracted signal and the video level of the video signal are compared by a comparator, and only when the video level becomes below the reference level, the signal extracted immediately before is compared. The video signal is held by the sample and hold circuit, and the video signal is delayed by a delay circuit by a time equivalent to the delay effect of the integration circuit and the sample and hold circuit, and the delayed video signal is extracted or held by the comparator. Binarize the signal at the signal level.

[Effect]

According to this method, while taking advantage of the advantages of a fixed threshold type that can perform stable detection and the advantages of a threshold integral type that is resistant to changes in illuminance, it is possible to reduce the effective field of view caused by parabolic shading, etc. Prevents narrowing.

〔Example〕

Embodiments of the present invention will be described below with reference to FIGS. 1 to 3 of the accompanying drawings.

FIG. 1 is a circuit diagram showing an embodiment of the present invention, and parts corresponding to those in FIGS. 6 and 7 are given the same reference numerals.

This embodiment is used as a binarization circuit 6 that constitutes a visual recognition device 5 shown in FIG. 4, and a delay circuit 12 is inserted between a connector 8 and a non-inverting input terminal of a comparator 9. At the same time, a sample hold circuit 13 and a peripheral circuit consisting of a comparator 14, an OR circuit 15, and a variable resistor VR ₂ are inserted between the integrating circuit 11 and the variable resistor VR ₁ , as shown in FIG. This is the same as the conventional threshold integral formula.

The video signal S ₁ inputted from the camera 1 via the connector 8 and the terminating resistor R ₁ is delayed by a predetermined time τ by the delay circuit 12 to produce a delayed video signal.
S ₁ ' and is applied to the non-inverting input terminal of the comparator 9.

The delay circuit 12 is composed of, for example, a delay line as shown in the figure, an ultrasonic delay element, a CCD, a surface acoustic wave element, etc., and its delay time τ is the delay with respect to the threshold level video signal S ₁ caused by the integration circuit 11. It is prescribed to be corrected.

On the other hand, the video signal _S1 is integrated by an integrating circuit 11 with a small time constant consisting of a resistor _R3 and a capacitor C, and the integrated signal is sent to a sample hold circuit 13.
is input.

Furthermore, the video signal _S1 is also input to the non-inverting input terminal of the comparator 14,
The sample and hold circuit 13 is connected to the inverting input terminal of
A reference level Sr formed by voltage-dividing the output signal _S4 of is inputted by a variable resistor _VR2 .

In addition, variable resistor VR ₂ is the threshold level.
It is connected in parallel with the variable resistor VR ₁ for adjusting Srx, and is used for adjusting the hold control of the sample and hold circuit 13.

The sample hold circuit 13 enters the sample (extraction) state when the S/H terminal is “1”, and
When the S/H terminal is “0”, it is in a hold (holding) state, and the switching is done by the output of the comparator 14 and the blanking signal (horizontal synchronization signal) from the camera 1.
It is carried out by BKL.

First, when the signal level of the video signal S ₁ becomes lower than the reference level Sr (S ₁ ≦Sr) formed by dividing the output signal S ₄ of the sample and hold circuit 13 by the variable resistor VR ₂ , the output of the comparator 14 falls from “1” to “0”, and this output is sent to the OR circuit 15.
By inputting the signal to the S/H terminal of the sample-and-hold circuit 13 via the signal, its operating state is switched from the sample state to the hold state.

In other words, the sample and hold circuit 13 enters the hold state only when the video level of the video signal S ₁ suddenly decreases and S ₁ ≦Sr, and holds the integral signal sampled (extracted) just before that. At the same time, the hold signal is output as the output signal S ₄ , and when S ₁ >Sr other than that, the integral signal that changes according to the video signal S ₁ is output as is as the output signal S ₄ .

Also, a blanking signal BLK as shown in Fig. 2 is output from camera 1, and this signal BLK is
Sample and hold circuit 13 via OR circuit 15
Each time an input signal is input to the S/H terminal of the sample/hold circuit 13, the sample/hold circuit 13 enters a sampling (extraction) state, and as a result, S ₁ ≦Sr just before the end of one scan, and one scan is completed while the integral signal is held. Even if the voltage is completed, sampling can be started from approximately 0V when the next scan starts.

Therefore, the output signal _S4 of the sample and hold circuit 13, which operates in this way, has the property of an integral signal that follows the video signal _S1 , but is resistant to large fluctuations in the video signal _S1 caused by large scratches, etc. This output signal S ₄ is connected to the variable resistor VR ₁
If the delayed video signal S ₁ ' is binarized by the comparator 9 using the threshold level Srx formed by dividing the voltage by It is possible to perform binarization processing that takes advantage of the fixed threshold type that allows stable detection.

Further, the effects of using the delayed video signal S ₁ ' obtained by delaying the video signal S ₁ by a predetermined time τ are as follows.

That is, from the waveform of the threshold level Srx, for example, as shown in _FIG.
It is preferable to remove large level drops and small level fluctuations seen in the waveform of S1, but due to the delay effect of the integrating circuit 11, the threshold level Srx before the end of scanning is _S1 with respect to the video level of the video signal _S1 . <Srx, so if left as is, the effective field of view will become narrower.

However, if the delayed video signal S ₁ ' obtained by delaying the video signal S ₁ by a predetermined time τ is used, there is a possibility that the delayed video signal S ₁ ' exceeds the threshold level Srx as shown in the figure. Even if parabolic shading occurs, it only occurs for a short time at the start, and the effective field of view ES ₃ is almost the same as the imaging range (scanning range) of the camera 1.

In the above embodiment, a signal corresponding to the integral signal obtained by dividing the integral signal into a required value by a variable resistor may be input to the sample and hold circuit 13.
In such a case, the variable resistor VR ₂ may be omitted and the signal S ₄ may be input directly to the inverting input terminal of the comparator 14. However, if you do this, it is of course necessary to change the level adjustment using variable resistor VR ₁ , but depending on the situation, variable resistor VR ₁
can be omitted and the output signal _S4 can be used as it is as Srx.

In addition, in the above embodiment, the line sensor camera 1
Although the present invention has been described with reference to the case of binarizing the video signal _S1 from the camera, it goes without saying that the present invention can also be applied to binarizing the video signal of an area sensor camera.

Furthermore, in the above embodiment, the comparator 1
4 described the example in which the video signal _S1 and the reference level Sr are compared, but the delayed video signal
S ₁ ' and the reference level Sr may be compared and the S/H of the sample hold circuit 13 may be switched based on the comparison result.

〔Effect of the invention〕

As explained above, according to this invention,
The video signal output from the camera is integrated by an integrating circuit, the integrated signal or a signal corresponding to the integrated signal is sequentially extracted via a sample hold circuit, and a reference level based on this sequentially extracted signal is set. The video level of the video signal is compared with the video level by a comparator, and only when the video level becomes below the reference level, the signal extracted immediately before is held in the sample hold circuit, while the video signal is is delayed by a delay circuit by a time equivalent to the delay effect of the integration circuit and the sample and hold circuit, and this delayed video signal is binarized by a comparator at the signal level of the signal extracted or held. , has the following effects.

(1) The effective field of view does not need to be narrowed even when detecting scratches in areas where the illuminance or amount of light within the field of view is not constant.

(2) No malfunctions even if the illuminance changes.

(3) Stable binarization can be performed without being affected by large scratches.

(4) Accurate threshold values can always be obtained.

Therefore, by using a visual recognition device equipped with a binarization circuit to which this method is applied, it is possible to always accurately recognize the size, position, and number of scratches, dirt, etc. on a workpiece.

[Brief explanation of the drawing]

FIG. 1 is a circuit diagram showing an embodiment of the present invention;
Figure 2 is a waveform diagram showing the blanking signal, Figure 3 is a waveform diagram showing the blanking signal.
The figure is a signal waveform diagram of each part to explain the effect of Figure 1. Figure 4 is a block diagram showing an example of the configuration of a visual recognition device. Figure 5 shows an example of the video signal _S1 in Figure 4. 6 is a circuit diagram showing a conventional example of a fixed threshold type binarization circuit, FIG. 7 is a circuit diagram showing a conventional example of a threshold integral type binarization circuit, and FIG. 8 is a waveform diagram. The figure is a waveform diagram for explaining the conventional problems. DESCRIPTION OF SYMBOLS 1... Line sensor camera, 2... Work, 3... Shooting range, 4... Scratch, 5... Visual recognition device, 6... Binarization circuit, 9, 14... Comparator, 11... Integrating circuit, 12... Delay circuit, 13... Sample and hold circuit.

Claims (1)

[Claims] 1 The video signal output from the camera is integrated by an integrating circuit, and the integrated signal or a signal corresponding to the integrated signal is sequentially extracted via a sample hold circuit, and the reference level is based on the sequentially extracted signal. and the video level of the video signal are compared by a comparator, and only when the video level becomes below the reference level, the signal extracted immediately before is held in the sample hold circuit; The signal is delayed by a delay circuit by a time equivalent to the delay effect of the integration circuit and the sample and hold circuit, and the delayed video signal is binarized by the signal level of the extracted or held signal by a comparator. A method for binarizing a video signal, characterized in that: