JP2783513B2 - Electronic component mounting equipment - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an electronic component mounting apparatus, and more particularly to an electronic component mounting apparatus for acquiring an image of a component having predetermined portions regularly arranged at predetermined intervals and positioning the component by recognizing the image. It is about.

[0002]

2. Description of the Related Art Conventionally, in a chip mounter, a CCD is used.
The positioning of components and boards mounted via image recognition processing such as components taken from a camera or the like or marks on the board is performed.

In particular, TLC (touchless centering)
In the mounting by the method, the image of the sucked part is taken in with a camera attached to the TLC head, binarized, and the data such as the shape, the center position of the part, the rotation angle, etc. are measured from the data pattern, and this data is measured. The components are mounted based on.

[0004]

In the above prior art, 2
Since a value image is targeted, there are disadvantages such as poor detection accuracy at a portion where a large number of fine pins (leads) such as an IC project, and weakness against noise.

An object of the present invention is to solve the above-mentioned problems,
It is an object of the present invention to provide an electronic component mounting apparatus capable of mounting components having fine parts such as C with high accuracy and being resistant to noise.

[0006]

SUMMARY OF THE INVENTION In order to solve the above-mentioned problems, an electronic component for acquiring an image of a component having predetermined portions regularly arranged at predetermined intervals and positioning the component by image recognition. In the mounting device, an image signal of a predetermined portion of the component is extracted, and the extracted image signal is input to a complex filter having a bandpass characteristic and a Hilbert transform characteristic corresponding to a spatial frequency of a predetermined portion of the component, and A configuration is employed in which the component is detected as a representative position of a predetermined portion of the component based on the filter output, and the component is positioned based on the detected position of the predetermined portion of the component.

[0007]

According to the above arrangement, a representative position of a predetermined portion of a component is detected by subjecting the image signal of the component to complex number filtering, and the component can be positioned based on the detected representative position of the predetermined portion. .

[0008]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, the present invention will be described in detail based on embodiments shown in the drawings.

FIG. 1 shows the overall structure of a substrate transfer machine 1 employing the present invention.

In the drawing, reference numeral 1 denotes a substrate transporter, and reference numeral 2
Is a printed circuit board on which components (such as IC chips) are mounted,
Reference numeral 3 denotes an xy robot, and reference numeral 4 denotes an electronic component supply device to be mounted.

Reference numeral 5 denotes an IC mounted on the printed circuit board 2 by the chip mounter head, reference numeral 6 denotes a suction nozzle of the chip mounter head, reference numeral 7 denotes a TLC unit for acquiring image data picked up by the chip mounter head, and reference numeral 8
An image processing apparatus for recognizing a center position, a rotation angle, and the like of a chip from image data acquired by the TLC unit 7.

FIG. 2 shows the structure of a TLC unit of a chip mounter employing the present invention. The TLC unit recognizes a component gripped by the chip mounter head through image processing.

In FIG. 2, reference numeral 13 denotes a suction port for sucking parts, which communicates with a negative pressure chamber defined by two upper and lower glass stages 14, 15. The control of the suction negative pressure is performed via the suction elbow 11.

In order to control the rotational position of the component sucked by the suction port 13, a negative pressure chamber is formed at the head of the cylindrical rotating unit 10. The rotating unit 10 is rotatably supported by a chassis 18, and its rotating position is controlled by a motor 12 via a gear 19.

An image of the component sucked by the TLC camera 9 is taken through the transparent glass stages 14 and 15. The lens 16 of the TLC camera 17 is arranged toward the glass stages 14 and 15.

In this embodiment, an illumination unit 9 which can be moved in the direction of the arrow is provided above the rotating unit 10 to illuminate the component sucked by the suction port 13 when photographing by the TLC camera 17. The illumination unit 9 is moved onto the rotating unit 10 as necessary via a driving unit (not shown).

Next, the operation of the above configuration will be described with reference to FIG. FIG. 5 is a flowchart showing the operation of the chip mounter.

The IC 5 sucked and picked up by the IC supply device 4 is carried by the xy robot 3 to the suction port 13 of the TLC unit 7 (step S1), and the TLC camera 1
7 is performed (step S2).

The image signal of the IC imaged as shown in FIG. 3 is sampled / quantized and converted into a digital density signal, and a window is set for each of the upper, lower, left and right pin rows (step S3). By extracting the image signal of each pin row on the left and right, applying a complex filter corresponding to a fixed spatial frequency between the pins in the window, and finding the zero cross point of the phase, a portion representing the position of the pin ( Here, the center position of the pin is extracted (step S4).

Here, a process of detecting a portion representing the position of a pin (hereinafter referred to as a pin position) will be described. In the pin position detection processing, an image (density) signal output from the TLC camera 17 is input to a complex filter constituted by a digital filter (which may be constituted by an analog filter), and a phase is calculated from the filter output. By taking

FIG. 4 shows, from the top, the shape of an IC pin row,
A real part and an imaginary part (lower two stages) of image data after passing through a complex filter having Hilbert transform characteristics with respect to density data of an image of the same pin row acquired from the TLC camera 17 are shown.

The complex filter used here assumes that the brightness (density) of a certain pixel is affected by the brightness of the neighboring pixels, and approximates the degree of influence of the neighboring pixels by a Gaussian curve. This is a filter that obtains the degree of change in brightness by taking the second derivative of the total value.

For example, in FIG. 6, the brightness of pixel A is affected by the brightness of surrounding pixels B0 to B3 and C0 to C3. At this time, it is considered that the degree of influence of the pixels B0 to B3 at the same distance is B0 = B1 = B2 = B3, and the degree of influence of the pixels C0 to C3 at the same distance is C0 = C1 = C2 = C3. However, it is clear that the influence of the pixel (B0 = B1 = B2 = B3) is greater than the influence of the pixel C0 (= C1 = C2 = C3).

As described above, it can be seen that the degree of influence of a certain pixel from a neighboring pixel can be expressed by a function of the physical distance between the pixel and the neighboring pixel. In the case of the complex filter, the square of the distance may be considered as a variance of the statistic.
Although a two-dimensional image has been described with reference to FIG. 6, in this embodiment, since only the position detection in the arrangement direction of the pins of the IC is targeted, the complex number filter is configured by a one-dimensional filter. The filter multiplier of the real part of the one-dimensional complex filter is

[0025]

(Equation 1)

The filter multiplier of the imaginary part is

[0027]

(Equation 2)

It can be shown that Here, σ is the standard deviation of the Gaussian distribution, ω is the spatial frequency of the pin image of the IC, x
Indicates a distance (corresponding to a time on a digital filter or a filter tap position).

The complex number filter having the outputs of the real part and the imaginary part can be regarded as a smoothing filter, and is effective for suppressing a noise component. For example, in FIG. 6, if pixel A is noise, the density of surrounding pixels should be completely independent of the detected value of pixel A, and therefore, when passing through the real part and imaginary part of the complex filter, the pixel A The density of surrounding pixels is taken into account as the density value, and noise components are suppressed.

After passing through a complex filter as described above,
The image signal is as shown in FIG. The pin position can be obtained from the filter outputs (R, I) of the real part and the imaginary part.

Assuming that the pin position is the center of the pin shown at the top, the phase of the real part output (R) and the phase of the imaginary part output (I) are 90%.
The real part output (R) and the imaginary part output (R) correspond to the pin positions to be detected, respectively, and the real part output (R) and the imaginary part output (I) correspond to the pin positions to be detected. The pin position can be obtained using the output (I).

The imaginary part output values are sampled at the sampling timings of X1, X2,... Xn, and the real part outputs R1, R2,.
.. Rn and the imaginary part outputs I1, I2,.

[0033]

(Equation 3)

It can be shown that The fifth stage in FIG. 7 shows sampling timings X1, X2,.
.., θn corresponding to n.

The sampling timings X1, X2,... Xn
, .Theta.n correspond to the primary linear line in the complex number filter as shown in the lowermost stage of FIG. 7, and therefore, on this linear line, .theta.1, .theta.2 corresponding to the timings X1, X2, X3. , Θ3, the result obtained by linear interpolation from the phase information corresponds to the pin position.

In this manner, the pin position can be easily and accurately detected using the outputs of the real part and the imaginary part of the complex filter. In this case, as shown at the bottom of FIG. 7, the X coordinate of the pin position can be obtained with high accuracy by simple linear approximation and simple calculation.

It is needless to say that the pin position can also be detected by performing processing for detecting the position of the peak or the zero cross using either the output of the real part or the imaginary part. For example, the edge of the pin can be detected by detecting the zero-cross point of the real part output using the zero-cross method that can be executed very easily, and the center of the pin can be detected by detecting the zero-cross point of the imaginary part output.

In this way, a complex number filter is applied,
Take the second derivative of the image signal and find its solution, ie 2
If the zero cross point of the next derivative curve is obtained, this position can be detected as a pin position.

Thereafter, the center of each pin row is calculated by averaging the obtained pin position data (step S4 in FIG. 5) for each pin row (step S5), and the center of the four pin rows on the four sides is calculated. From the point, the center position and the rotation angle of the IC are measured (step S6), and based on the data, the xy robot 3
(Step S7), and then mounted on the substrate 2 whose position has been measured in advance (Step S7).
8).

According to the above-described embodiment, a band-pass filter and a complex filter having second-order differential characteristics are applied by utilizing the fact that the pin interval of the component is a constant frequency, and the phase position such as a peak or zero cross is applied. , The position of the IC pin and other parts can be detected with high accuracy. Component detection accuracy can be improved up to sub-pixel accuracy. In the conventional binarization method, the detection accuracy is about ± 1/4 pixel. However, according to the present invention, the detection accuracy can be improved to ± 1/10 pixel. Further, by using the real part and imaginary part outputs of the complex filter, there is an advantage that highly accurate position detection can be performed by phase interpolation with good linearity, and the noise component is strong.

Further, even for the noise component, the edge portion of the component can be accurately detected even if the noise component is mixed due to the smoothing characteristic of the complex number filter.

Further, according to the above configuration, since the hardware of the filter portion is easily realized, high-speed processing can be realized simply and inexpensively.

In the above, a complex number filter has been exemplified as a filter for detecting a pin position.
aplacian of Gaussian) filter, LOS (Laplacian of Sink) filter, or DOS (Difference of
A filter using a SINK function such as a (Sink) filter may be used.

Although an example using a complex number filter has been described above, it goes without saying that the same processing as in FIG. 7 can be performed by using two real number filters whose output phases are different from each other by 90 °.

[0045]

As is apparent from the above description, according to the present invention, an electronic component which acquires an image of a component having predetermined portions regularly arranged at predetermined intervals and positions the component by image recognition. In the mounting device, an image signal of a predetermined portion of the component is extracted, and the extracted image signal is input to a complex filter having a bandpass characteristic and a Hilbert transform characteristic corresponding to a spatial frequency of a predetermined portion of the component, and Detecting the position as a representative position of the predetermined part of the component based on the filter output, and adopting a configuration for positioning the component based on the detected position of the predetermined part of the component, the component can be positioned at high speed and with high accuracy. Can do it. At this time, since a complex filter having a bandpass characteristic and a Hilbert transform characteristic corresponding to the spatial frequency of a predetermined part of the component is used, the phase interpolation with good linearity is performed by using the real part and imaginary part outputs of the complex filter. This enables highly accurate position detection, and enables high-accuracy detection of parts such as IC pins and edge parts, and eliminates the need to set thresholds.
Due to the smoothing characteristic of the filter, the filter is extremely resistant to noise components, and has excellent advantages such as accurate detection of components and accurate mounting of components based thereon.

[Brief description of the drawings]

FIG. 1 is a perspective view showing the appearance of an electronic component mounting apparatus according to the present invention.

FIG. 2 is a perspective view of a TLC unit of FIG. 1;

FIG. 3 is a photographed gray scale image of an IC.

FIG. 4 is an explanatory diagram illustrating an example of a pin position detection process.

FIG. 5 is an operation flowchart of the apparatus.

FIG. 6 is an explanatory diagram illustrating an effect of a complex number filter.

FIG. 7 is an explanatory diagram showing a real part and imaginary part output process of a complex filter.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Substrate transfer machine 2 Printed circuit board 3 xy robot 4 Electronic component supply device 5 IC 6 Suction nozzle 7 TLC part 8 Image processing device 9 Lighting 10 Rotating part 11 Suction elbow 12 Motor 13 TLC stage suction port 14 Above glass stage 15 Below glass stage 16 Lens 17 TLC Camera 18 Chassis 19 Gear 21 Pin Row Data

Claims (5)

(57) [Claims] 1. An electronic component mounting apparatus for acquiring an image of a component having predetermined portions regularly arranged at predetermined intervals and positioning the component by image recognition, wherein an image signal of the predetermined portion of the component is output. Extracting, and inputting the extracted image signal to a complex filter having a bandpass characteristic and a Hilbert transform characteristic corresponding to a spatial frequency of a predetermined portion of the component, as a representative position of the predetermined portion of the component based on the output of the complex filter. An electronic component mounting device, comprising: detecting and positioning the component based on the detected position of a predetermined portion of the component. 2. The electronic component mounting apparatus according to claim 1, wherein a predetermined waveform point of a real part output of the complex number filter is detected as a representative position of a predetermined portion of the component. 3. The electronic component mounting apparatus according to claim 1, wherein a predetermined waveform point of an imaginary part output of the complex filter is detected as a representative position of a predetermined portion of the component. 4. A real part output and an imaginary part output of the complex number filter are sampled by a predetermined clock, and phase information of the real part output or the imaginary part output is obtained from the obtained real part and imaginary part sampling data. The electronic component mounting apparatus according to claim 1, wherein a predetermined waveform point of a real part output or an imaginary part output is detected as a representative position of a predetermined part of the component by linear interpolation from the phase information. 5. An electronic component mounting apparatus for acquiring an image of a component having predetermined portions regularly arranged at predetermined intervals and positioning the component by image recognition, wherein an image signal of the predetermined portion of the component is provided. Extracting the extracted image signals and inputting them to two real filters having band-pass characteristics and Hilbert transform characteristics corresponding to spatial frequencies of a predetermined portion of the component and having output phases different by 90 °; An electronic component mounting apparatus, wherein the electronic component mounting apparatus detects a position of the component as a representative position of the component based on the output, and performs positioning of the component based on the detected position of the component.