JPS6322241B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a shape detection device that accurately detects the shape of a convex object from an optically sectioned image.

In automatic shape measurement and inspection using the optical section method, the optical section image obtained by the optical system is detected by an imager such as a TV camera, and the optical section line is extracted from the detected two-dimensional gray image. Input data needs to be compressed to facilitate further processing.

In the light cutting method illustrated in FIG. 1, a slit-shaped light beam is projected from a slit projector 1 onto an object to be measured 2, which is detected by an image detector 3, and the light is cut from the detected two-dimensional grayscale image. As a conventional technique for extracting lines, there is a method of obtaining a light cutting line by extracting the brightest point for each vertical scanning line of a detected image. Fig. 2 is a diagram for explaining this, and Fig. 2a is a light section image, that is, 2
The bright parts of the dimensional gradation image are expressed as black.
FIG. 2b shows the video signal of one arbitrary vertical scanning line i in FIG. 2a. The ordinate Zi of the point where the brightness V of this video signal is the brightest is detected, and this is taken as the output value of the scanning line i. By performing this for all scanning lines in the vertical direction, the optical section line shown in FIG. 2c, that is, the waveform data can be obtained.

Such conventional technology is effective for detecting the shape of an object whose shape is almost constant in the y-axis direction, as illustrated in FIG. When detecting a bright spot due to abnormal reflection in a part that is not directly hit by the slit light, the brightest point in that part may be mistakenly detected as a light cutting line.

As an example, FIG. 3 shows a configuration of an optical cutting method for detecting the shape of a soldered part of an electrical component mounted on a printed wiring board as the object to be measured. In FIG. 3, the slit projector is composed of a lamp 4, a condenser lens 5, a slit 6, and a projecting lens 7, and the image detector is composed of an imaging lens 8 and an image pickup device 9. The slit-shaped light is projected onto the soldering part 10, which is the object to be measured, from above through the lens 7, and is detected diagonally from the side.
FIG. 4 is an enlarged view of the periphery of the object to be measured, and the slit light projected from above forms a slit bright line 11 on the soldering part 10. By the way, at the point 12 on the soldering surface, the light from the lens 7 is specularly reflected and enters the lens 8 due to the angle of inclination of the surface at that part. Most of the light from the lens 7 is focused on the slit bright line 11, but some of it is scattered, so that the point 12 is observed brightly. Figure 5a shows an example of the optical section image.
In such a case, if the optical section line is extracted using the conventional method, an abnormal optical section line will be obtained as shown in FIG. 5b.

As a conventional technique, here we have shown the problems with the simplest method of extracting the brightest position, but this is different from other methods, that is, when the video signal V is set to a preset value in each scan of image detection.
Compared with V ₁ and V ₂ (V ₁ > V ₂ ), when the maximum value Vmax of the video signal V in one scan is smaller than V ₂ , Z=0,
Z is the Z position where V=Vmax when it is between V ₁ and V ₂ , Z is the Z position where V=V ₁ first in one scan when it is greater than V 1, and _Z is the Z position where V=V ₁ is the last when V is greater than V ₁ . A patent application filed in 1973 to extract the optical cutting line Z(x) by setting the Z position at _Z2 and setting Z=( _Z1 + _Z2 )/2.
The 146427 method also has the same problem.

An object of the present invention is to eliminate the above-mentioned drawbacks of the prior art, and to extract the correct light section line from a light section image including optical disturbance for a convex object, and to obtain the shape of the convex object. An object of the present invention is to provide a shape detection device capable of accurately detecting a shape.

Abnormal reflection, which is exemplified as a problem with the prior art, occurs on the surface of the object to be measured. Therefore, the abnormal reflection part exists below the slit emission line. Taking advantage of this,
In the present invention, optical cutting lines are selectively extracted. That is, each video signal V when scanning vertically from the image detector from top to bottom is set to a preset value V ₁
In comparison with , the optical section line is first extracted only for the section where V≧V ₁ , and even if a section where V≧V ₁ occurs thereafter, no optical section line is extracted. That is, the present invention provides a slit-shaped light band,
A slit projector that projects light onto a convex object from directly above, and an imaging system that two-dimensionally images a slit bright line image on the convex object from an oblique direction with respect to the slit-shaped light band. A device and a two-dimensional image obtained from the imaging device are input, and the one-dimensional image showing the same horizontal coordinates of the convex object is examined from the highest side of the convex object, and the slit is Video signal level V corresponding to the bright line image
detects the section A in which a portion where the value is greater than a predetermined set value V ₂ has finished occurring for the first time, outputs the section A signal, and sets the video signal level V to a predetermined set value V ₂
If the detection interval setting means does not output the interval signal, the video signal level V is set lower than the set value V ₂ for the interval in which the interval signal is detected by the detection interval setting means. When the set value V ₃ is reached, the optical section line is extracted as a height coordinate value Z signal indicating the peak of the video signal, and when the set value V ₃ is not reached, a signal without the optical section line is extracted. When the video signal level V reaches the set value V ₁ set slightly lower than the saturation level for the section in which the section signal is detected by the maximum value position detecting means and the detection section setting means, the above-mentioned The video signal is initially set to the above setting value V ₁
Height coordinate value Z indicating the center position between the coordinate value Z ₁ that became , and the coordinate value Z ₂ that finally became the above set value V ₁
A center position detecting means for extracting an optical cutting line as a signal, and an output signal of the maximum value position detecting means and an output of the center position detecting means depending on whether the video signal level V exceeds the set value _V1 . and a selection means for selecting and outputting a signal, the detection section setting means, the maximum value position detection means, the center position detection means, and the selection means determine the convex shape based on the two-dimensional image obtained from the imaging device. This is a shape detection device characterized in that it processes the object in the lateral direction and detects the shape of the convex object using a light section line.

FIG. 6 shows the overall configuration of an embodiment of the present invention.
This consists of a light cutting method shape detection section consisting of a slit projector 1 and an image detector 3, and a light cutting line extraction circuit 13 which inputs the detected video signal and outputs a light cutting line signal. The optical cutting line signal extracted in step 13 is output to a display device for shape measurement or to a signal processing circuit for automatic inspection.

FIG. 7 shows the basic configuration of an embodiment of the optical cutting line extraction circuit 13 of FIG. Optical cutting line extraction circuit 13
is the video signal V, the set values V ₁ , V ₂ , V ₃ (V ₁ > V ₂ ≧
V ₃ ), and a trigger signal horizontal synchronization signal Hd for each scan of the image detector 3, and is composed of a maximum value position detection circuit 14, a center position detection circuit 15, a detection interval setting circuit 16, and a selection circuit 17. Here, the imaging method of the image detector 3 is one in which horizontal scanning is repeated from the top to the bottom of the screen as in a normal TV camera, as shown in Fig. 8a, and one in which horizontal scanning is repeated from the top to the bottom of the screen, as shown in Fig. 8b.
One idea is to repeat vertical scanning from left to right on the screen, like a TV camera rotated 90 degrees.
First, an embodiment using the method shown in FIG. 8b will be described.

Next, the function of the optical cutting line extraction circuit 13 consists of a peak position detection method implemented by circuits 14, 15, and 17, and a detection interval setting process implemented by circuit 16. The former corresponds to the content of patent application No. 146427-1972. That is, when the maximum value Vmax of the video signal falls within the interval [V ₁ , V ₃ ] as shown in FIG. 9a, the maximum value position detection circuit 14 determines its Z position Z. Also, as shown in Figure 9b, Vmax is V ₁
When it is larger, the center position detection circuit 15 first determines the Z position Z ₁ where V≧V ₁ and finally the Z position Z ₂ where V ≧ V ₁ , and further calculates Z = (Z ₁ + Z ₂ )/ 2
Find its center position as . If V becomes _V1 or more in the center position detection circuit 15 during one scan, the selection circuit 17 selects the center position detection circuit 1.
If V does not exceed _V1 , the selection circuit 17 outputs the output of the maximum value position detection circuit 14 as the value of the optical cutting line. Further, if V does not exceed _V3 during one scan, the selection circuit 17 outputs zero. The detection interval setting is as above
Detection is performed only for A up to the end of the first mountain where V≧V ₂ as shown in FIGS. 10a and 10b. However, as shown in FIG. 10c, if V≧V ₂ does not hold, section A is the entire area. As a result, an input video signal as shown in FIG. 5a is converted into an optical section line as shown in FIG. 11.

The first example is a more detailed version of the embodiment shown in FIG.
This is Figure 2. The details of the operation will be explained in more detail using FIG. 12.

The maximum position detection circuit 14 has a peak hold 1
8, a comparator 19, a one-shot 20, an AND 21, a register 22, a comparator 24, and a flip-flop 25.

The center position detection circuit 15 includes a comparator 26,
Flip flop 27, one shot 28, 3
1, AND29, 32, _Z1 register 30, _Z2 register 33, adder 34, and flip-flop 38.

The detection interval setting circuit 16 includes a comparator 35,
It consists of a flip-flop 36.

There is also a counter 2 to know the Z coordinate at each moment.
3 and a selection circuit 37.

In the maximum position detection circuit 14, the video signal V enters a peak hold 18, and the peak value of V up to a certain moment during one scan is held. Comparator 19 compares this held peak value with V at that moment, and outputs "1" (ON) to one shot 20 when V is equal to or greater than the held peak value. On the other hand, V is compared with V ₃ in the comparator 24, and when V≧V ₃ , the comparator 24 outputs “1”. Furthermore, in the detection interval setting circuit 16, the comparator 35
Compares V ₂ and outputs "1" when V<V ₂ . The flip-flop 36 is reset by the trigger signal Hd, and the output changes from "1" to "0" (OFF) when V first shifts from ≧V ₂ to V<V ₂ . The AND 21 takes the AND of the outputs of the one shot 20, the comparator 24, and the flip-flop 36, so that in the interval A (Fig. 10a) and V
≧V ₃ and “1” is output at the moment of peak. This is used as a load signal for the register 22, and the instantaneous value of the counter 23, which is reset by the trigger signal Hd, is sent to the register 22.
to be memorized. As a result, at the end of one scan, the register 22 holds the Z coordinate Z of the maximum value in the section where V≧V ₂ for the first time.

In the center position detection circuit 15, the video signal V is V ₁
A comparator 26 compares the output of the flip-flop 27 with the set side of the flip-flop 27. Therefore, the one shot 28 detects the first moment in one scan when V≧V ₁ . By ANDing this with the signal indicating interval A (FIG. 10b) from the flip-flop 36 and using the output as the load signal for the register 30, the _Z1 register 30 receives V≧ in interval A. The Z position Z ₁ that first becomes V ₁ is held. On the other hand, the one shot 31 detects the moment when V≧V ₁ is no longer satisfied, and by ANDing this with the interval A and using the output as the load signal for the register 33, the Z ₂ register 33 is loaded. Section A
Finally, the Z position Z ₂ where V≧V ₁ does not hold is held. The adder 34 performs the calculation Z=(Z ₁ +Z ₂ )/2 from the hold values of the Z ₁ register 30 and the Z ₂ register 33.

The flip-flop 25 uses Hd as a reset signal and the output of the comparator 24 as a set signal, and is used to detect whether a case of V≧V ₃ exists during one scan. Flip-flop 38 uses Hd as a reset signal, and AND29
The output of 1 is used as a set signal, and is used to detect whether the case shown in FIG. 10b has occurred.
Therefore, the selection circuit 37 uses these as condition signals, and when the flip-flop 38 is set, the output value of the adder 34 is used as the condition signal.
When remains reset and flip-flop 25 is set, it outputs the hold value of register 22, and otherwise outputs zero.

Next, as another example, an example of the imaging method shown in FIG. 8a will be shown. Here, the detection section in the height direction and the z-axis is limited to (Za, Zb) (Fig. 8 a)
reference). A first embodiment of the optical cutting line extraction circuit 13 is described below.
Shown in Figure 3. As shown in Figure 8a, the input signals are a horizontally scanned video signal V, Hd indicating the starting point of each horizontal scan, a clock corresponding to the sampling interval in the horizontal direction and the x-axis direction in each horizontal scan, and a level V. ₁ , V ₂ , V ₃ and the z-axis detection sections Za and Zb shall be provided by a digital switch. The optical cutting line extraction circuit 13 includes these digital switches 39, 40, 41,
46, 47, comparator 42, 43, 44, 4
5, Z coordinate generation circuit 48, X coordinate generation circuit 49,
A/D converter 50, gate 51, data selector 52, read/write control circuit 53, 54, 5
5, 56, random access memory 57, 58, 59 for storing the values of Zm ₁ , Zm ₂ , Vmax;
It is composed of a random access memory 60 for storing the Zm written flag, a random access memory 61 for storing the first peak detected flag, and an adder 62.

The video signal V is converted by the Z coordinate generation circuit 48 into a digital signal by the A/D converter 50 through the gate 51 only during the interval (Za, Zb). The comparator 44 compares the video signal with V ₃ in the digital switch 39 and detects V≧V ₃ . Comparator 43 similarly detects V≧V ₁ . Comparator 42 similarly detects V<V ₂ . On the other hand, the x coordinate is generated by the x coordinate generating circuit 49 by counting clocks, and this is referred to as the address of the random access memories 57, 58, 59, 60, and 61. Therefore, the comparator 45 compares the video signal V with Vmax stored in the memory 59,
If V>Vmax, V is updated as Vmax in the memory 59, and the content Vmax of the memory 59 becomes Vmax at the x-coordinate at that time. Note that each memory is cleared prior to extracting the optical cutting line. The memories 60 and 61 also store the detection section A in FIG.
_This sets In order to ignore it, once V≧V ₂ becomes V<V ₂ , the first peak detected flag is stored, and extraction of the optical section line for the x-coordinate is thereafter stopped.

The read/write control circuits 54, 55, and 56 respectively control the video signal V, the set values _V1 , _V3, and the memory 5.
According to the comparison result with the maximum video signal Vmax which is sequentially updated and stored starting from Za in 9, the following operation is performed.

1 If video signal < Vmax, 54, 55, 56
Does not work together.

2 Video signal > Vmax, and video signal <
If V ₁ and the first peak has not been detected at 61, 54, 55, and 56 operate, and write the z coordinate at that time to 57, 58, and the value of the video signal at that time to 59, respectively.

3. If the video signal=Vmax and the first peak has not been detected in 61, only 54 operates and writes the z coordinate at that time in 58.

4 If the video signal ≧ V ₁ and the first peak has not been detected in 61, 54 and 56 operate, write the z coordinate at that time to 58, and write V ₁ to 59 via the data selector 52. .

5 54,5 if the first peak has been detected at 61
5 and 56 do not work together.

During these operations, 60, 53, and 61 operate as follows and store the first peak detected flag in the memory 61. In other words, the memory 60 uses the read/write control circuit 55 for the section (Za, Zb).
When a Zml write command is issued, a flag 1 is set in the memory contents corresponding to the X coordinate. Therefore, the contents of the memory 60 are "0" for each X coordinate until Zm ₁ is detected for the first time, become "1" after detection, and "1" is stored thereafter. The read/write control circuit 53 ANDs the contents of the memory 60 and the comparator 42 . Therefore, when Zm ₁ is detected once and thereafter the video signal <V ₂ is established, it becomes "1".
Therefore, the contents of the memory 60 are "1" after Zm ₁ is detected once and the video signal < V ₂ is established for the first time.
is memorized. That is, if the video signal <V ₂ is satisfied after the first peak is detected, the flag "1" is set, and it is remembered that the first peak has been detected at that x coordinate.

When the above operation is executed from Z coordinate Za to Zb, the next Z coordinate data corresponding to the x coordinate is written into the memories 57 and 58.

For the x-coordinates where V≧V ₁ does not hold between 1 Za and Zb, the Z-coordinates that indicate the peak value within the Z-coordinate interval where V≧V ₃ are written in the memories 57 and 58. Note that when V≧V ₂ as shown in FIG. 10a, only the first section A is set as the detection section.

2 As shown in Figure 10b, V≧V ₁ between Za and Zb
If V≧V ₁ , the start end Z ₁ of the section is written into the memory 57 and the end end Z ₂ is written into the memory 58. Note that when V≧V ₂ as shown in FIG. 10b, only the first section A is set as the detection section. Therefore, after the processing at Z=Zb is completed,
The gate 51 is closed by the command 48, and the x coordinate is incremented to 54, 55, 58, 57.
When a read command is issued from 58 to 62,
The memory contents of 57 are output, 62 outputs their average value, and the optical section line extracted from 62 is output as Z coordinate string data.

Although the image detector was not mentioned in the above embodiments, it can be used for TV cameras, two-dimensional solid-state image sensors, combinations of one-dimensional solid-state image sensors and optical scanning methods, photomultiplier tubes, etc., and optical receivers. Any combination of target scanning methods may be used.

In the above embodiment, V ₁ > V ₂ > V ₃ , but V ₂ =
V ₃ is also fine.

In the above embodiments, when V≧V ₁ , the center position is detected using V ₁ as the threshold, and when V ₃ ≦V<V ₁ , _the maximum position is detected. V ₃ may only detect the maximum position. In these cases, the configuration of the optical cutting/extracting circuit shown in the embodiment becomes simple, using only a part of it.

In the embodiment shown in FIG. 13, Za and Zb are set with a digital switch and used as fixed values, but they are made variable values and Z ₁ and Z ₂ are set according to each x coordinate.
It may also be a method of changing.

In the embodiment shown in FIG. 12, Za and Zb are not particularly given, but as in FIG.
Zb may also be applied.

According to the present invention, it is possible to extract a light section line of an object whose shape changes significantly without being affected by abnormal reflection from the object. This is particularly effective in detecting the shape of shiny objects such as soldered parts.

[Brief explanation of the drawing]

Fig. 1 is an explanatory diagram of the optical cutting method, Fig. 2 is an explanatory diagram of the conventional technique of the optical cutting line extraction method, Fig. 3 is an explanatory diagram of the printed board soldered part detection method using the optical cutting method, and Fig. 4
Figure 5 is a diagram explaining the state of abnormal reflection, Figure 5 is a diagram explaining an example of a malfunction of the prior art due to abnormal reflection, Figure 6 is an overall configuration diagram of an embodiment of the present invention, and Figure 7 is a diagram of the present invention. FIG. 8 is a functional block diagram of an embodiment of the optical cutting line extraction circuit.
Figure 9 is a diagram illustrating the functions of maximum position detection and center position detection, and Figure 10 is a diagram illustrating the function of detecting the maximum position or center position only from the bright area that occurs at the beginning of the present invention. FIG. 11 is a diagram illustrating an example of normal operation according to the method of the present invention, FIG. 12 is a detailed configuration diagram of an embodiment of the present invention, and FIG. 13 is another embodiment of the present invention. FIG. 1... Slit projector, 2... Image detector, 13... Light cutting line extraction circuit.

Claims (1)

[Claims] 1. A slit projector that projects a slit-shaped band of light onto a convex object from directly above, and a slit bright line that projects a slit-shaped band of light onto the convex object from an oblique direction with respect to the slit-shaped band of light. An imaging device that captures an image two-dimensionally, and a two-dimensional image obtained from the imaging device are input, and the most of the convex object is determined for a primary image showing the same horizontal coordinates of the convex object. The search is performed starting from the highest side, and the section A in which the video signal level V corresponding to the slit bright line image is higher than the predetermined set value _V2 is detected for the first time, and the section A signal is output. If the signal level V does not reach the predetermined set value _V2 ,
The video signal level V is set to the set value for the detection interval setting means in which no interval signal is output and the interval in which the interval signal is detected by the detection interval setting means.
When the set value _V3 , which is set lower than _V2 , is reached, the height coordinate value Z indicating the peak of the video signal is
maximum value position detection means for extracting a light cutting line as a signal and extracting a signal without a light cutting line if the set value V ₃ is not reached; and for the section in which the section signal is detected by the detection section setting means; When the video signal level V reaches the set value V ₁ that is set slightly lower than the saturation level, the video signal first reaches the coordinate value Z ₁ at the set value V ₁ and finally reaches the above setting. a center position detecting means for extracting an optical cutting line as a height coordinate value Z signal indicating the center position between the coordinate value _Z2 which has become the value V1 _, and whether or not the video signal level V exceeds the set value _V1 . a selection means for selecting and outputting the output signal of the maximum value position detection means and the output signal of the center position detection means; and the selection means process the convex object in the lateral direction based on the two-dimensional image obtained from the imaging device, and detect the shape of the convex object using a light section line. Shape detection device.