JPH02219183A - Recognizing device for position of object - Google Patents

Abstract

PURPOSE:To recognize a position of an object by turning on alternately each illuminating device and obtaining image data of a shadow of the object to be recognized from a difference of each image data, detecting a corner of a shadow part by executing a scan by a corner detection use pattern and deriving a position of the object to be recognized from the corner. CONSTITUTION:Each illuminating device 4, 5 placed so as to be opposed to each other in each diagonal upper part seen from an object to be recognized 2 is turned on alternately by a turn-on control means 6, and also, when these illuminating devices 4, 5 are turned on, respectively, an image signal from an image pickup device 3 is received and each image data is stored in image memories 8, 9. Subsequently, a difference of each image data stored in these image memories 8, 9 is derived by a difference image arithmetic means 11, 12 and image data of a shadow of the object to be recognized 2 is obtained, and with respect to the image data of this difference shadow, a corner detecting means 13 detects a corner of a shadow part by allowing a corner detection use pattern to scan. Next, from the corner detected by the corner detecting means 13, a recognizing means 13 recognizes a position of the object to be recognized 2. In such a way, the position of the object can be recognized without being influenced by a wiring pattern and a color, etc.

Description

Detailed Description of the Invention [Object of the Invention] (Industrial Application Field) The present invention is a method for detecting the position of an object, which is applied to inspecting the mounting state of chip components mounted on a board, such as dropping out or misalignment. Relating to a recognition device. (Conventional Technology) Electronic devices such as home appliances and OA (office automation) devices are becoming smaller, and along with this, the circuit boards used in electronic devices are also becoming smaller. For example, instead of circuit boards with lead wires, circuit boards that can mount small chipped resistors, capacitors, etc. are now being used. Currently, there is a demand for high-density mounting on circuit boards by miniaturizing such mounted components. When mounting such chip components on a circuit board, the chip components are automatically temporarily mounted on a circuit board on which a predetermined wiring pattern has been formed using a mounting device, and then soldering is carried out using a device, etc. Do it and finish. However, when mounted on a circuit board, they often fall off or become misaligned. If this soldering is detected later, a great deal of effort will have to be spent on its correction. For this reason, inspections such as dropping and misalignment of chip components on circuit boards are
Previously, soldering was done visually by workers. Therefore, the test results varied depending on the worker, resulting in low reliability. Therefore, technology has been developed to automatically inspect chip components for falling off or misalignment. There are basically three methods for such inspection: a light section method, a pattern comparison method, and an oblique light shadow method. Among these methods, the optical cutting method irradiates the cutting light onto the circuit board from an oblique direction, and when observing the circuit board from above, it is possible to obtain information that if there is a component, a step will occur in the cutting light. This method is used as a method. Further, the pattern comparison method is a method in which image data (shade level, color, etc.) of a board to be inspected is compared pixel by pixel with image data of a non-defective board stored in advance. Therefore, in the optical cutting method, it is necessary to scan the entire surface (or part) of the circuit board with the cutting light, and for example, in order to detect the rotation of a rectangular component, the irradiation must be performed two or more times at different locations per component. It has to be done and it takes time. Furthermore, the pattern comparison method requires a large number of pixel comparisons, which is also time consuming. Compared to the above two methods, the oblique shading method uses two opposing
The image data C of the shadows A and B of the chip component as shown in Fig. 7 is obtained from the difference between the two images when two lights are placed diagonally as seen from the chip component and the lighting of the lights is changed. , add this image data C in the X and Y directions to obtain each projection image data, and calculate its peak value,
E, F. Extremely high-speed inspection is possible by determining the edge positions of the chip component from G and determining the position of the chip component from these edge positions. However, in this oblique shading method, shadows A and B of chip components may be missing due to the state of the circuit board, that is, the presence or absence of wiring patterns, printed characters, etc. In this case, each peak value of the projected image data may be As shown in H, the edge becomes undetectable. Therefore, the position of the chip component cannot be detected, and the reliability of the test results is low. (Problems to be Solved by the Invention) As described above, although the oblique light shading method can inspect faster than the light sectioning method and the pattern comparison method, shadow information on chip components may be affected by the presence or absence of wiring patterns on the circuit board, printed characters, etc. Test results were sometimes missing, making test results unreliable. SUMMARY OF THE INVENTION Therefore, an object of the present invention is to provide a highly reliable object position recognition device that can recognize the position of an object without being affected by the wiring pattern, color, etc. on a circuit board. [Structure of the Invention] (Means for Solving the Problems) The present invention includes an imaging device disposed above an object to be recognized disposed on a substrate, and an image pickup device disposed above an object to be recognized disposed on a substrate, and an imaging device facing each diagonally upward when viewed from the object to be recognized. each of the lighting devices arranged in the same way, a lighting control means for turning on one of these lighting devices individually, and an image that receives an image signal from an imaging device when each lighting device is turned on and stores it as each image data. a memory, a difference image calculation means for calculating the difference between each image data stored in the image memory to obtain image data of the shadow of the object to be recognized, and a difference image calculation means for obtaining the image data of the shadow obtained by the difference image calculation means. The above object can be achieved by providing a corner detection means for detecting a corner of a shadow portion by scanning a corner detection pattern using the corner detection means, and a recognition means for recognizing the position of the object to be recognized from the corner detected by the corner detection means. This is a device for recognizing the position of an object. (Function) By providing such a means, each illumination device arranged diagonally upward and facing each other when viewed from the object to be recognized is turned on one by one by the lighting control means, and together with this, the illumination devices are turned on one by one by the lighting control means. When each device is turned on, it receives an image signal from the imaging device and stores each image data in the image memory. Then, the difference between the image data stored in the image memory is calculated by the difference image calculation means to obtain the image data of the shadow of the object to be recognized, and the corner detection means detects the corner with respect to the image data of the difference shadow. The corner of the shadow area is detected by scanning the pattern. Then, the recognition means recognizes the position of the object to be recognized from the corner detected by the corner detection means. (Example) Hereinafter, an example of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram of an object position recognition device. A chip component 2 is mounted on a circuit board 1. An imaging device 3 is disposed almost directly above the chip component 2, and illumination devices 4.5 are disposed at opposing positions obliquely above the chip component 2. One of these lighting devices 4 and 5 is turned on in response to switching signals a and b from a lighting switching circuit 6. The imaging device 3 captures an image including the chip component 2 and outputs the image signal, and this image signal is digitized by an A/D (analog/digital) conversion circuit 7 and stored in first and second image memories. Sent on 8.9. By the way, in this case,
A digital image signal when the lighting device 4 is turned on by the timing circuit 10 is stored in the first image memory 8,
Further, the digital image signal when the lighting device 5 is turned on is stored in a second image memory 91. A subtraction circuit 11 is connected to the output sides of these first and second image memories 8.9, and the difference between each image data stored in each image memory 8.9, that is, the image data of the shadow of the chip component 2 is determined. The signal is then binarized through the binarization circuit 12 and sent to the main arithmetic circuit 13. A pattern memory 14 storing a circular corner detection pattern P as shown in FIG. 2 and an AND operation circuit 15 are connected to this main operation circuit 13. It has a function of performing the following operations using the product calculation circuit 15. Note that the corner detection pattern P has a "1" level on the inside of the circle and a "0" level on the outside. Therefore, the main processing circuit 13 functions as a corner detection means for scanning the corner detection pattern P on the image data of the shadow of the chip component 2 to detect the corner of the shadow part. It has a function as a recognition means for recognizing the center position, inclination, etc. of the chip component 2 from the corner. The main arithmetic circuit 13 also has the function of sending a switching control signal to the lighting switching circuit 6 and an operation control signal to the timing circuit 10 to control the overall operation of the apparatus. Next, the operation of the apparatus configured as described above will be explained. When the imaging device 3 images the chip component 2 while only the lighting device 4 is turned on by the lighting switching circuit 6, the image signal output from the imaging device 3 is converted into a digital image signal by the A/D conversion circuit 7. The first image memory 8
The image data is stored as image data Dl as shown in FIG. In this image data Dl, 2' is an image of the chip component, and 2a is an image of its shadow. Next, when the imaging device 3 images the chip component 2 while the lighting device 4 is turned off by the lighting switching circuit 6 and the lighting device 5 is turned on, the image signal output from the imaging device 3 at this time is A/ Image data D as shown in FIG. 4 is converted into a digital image signal by the D conversion circuit 7 and stored in the second image memory 9.
2. In this image data D2, 2'
is an image of the chip component, and 2b is an image of its shadow. When each image data D1%D2 is stored in each of the image memories 8 and 9 in this way, the gray level of each pixel at the same coordinates in each image data stored in these image memories 8 and 9 is sequentially calculated by the subtraction circuit 11. sent to. Then, the subtracted output of the subtraction circuit 11 is binarized by a binarization circuit 12 using a predetermined threshold value and sent to the main arithmetic circuit 13. However, the portions of the same gray level in each image data become the "0" level, and the portions of mutually different gray levels, that is, the shadows 2a and 2b, become the "1" level, resulting in shadow image data as shown in FIG. It becomes D3. When the shadow image data D3 is thus obtained, the main processing circuit 13 executes the corner detection means. However, as shown in FIG. 6, this corner detection means sets area frames W1 and W2 in the portions including the corners of the shadows 2a and 2b, respectively, and sets the corners shown in FIG. 2 in these area frames W1 and W2, respectively. The detection pattern P is scanned vertically and horizontally. Here, the area frame Wl
To explain this in detail, first, a corner detection pattern P is set at the upper left corner of the area frame W1. Then, in this state, the logical product operation circuit 15 is activated, and the logical product of each gray level of each pixel of the corner detection pattern P and each pixel of the image data D3 of the same coordinate portion is calculated. As a result, corner detection pattern P and image data D3
If both levels are "1", the result of the logical product will be level "1", and all other level parts will be level "0". When this AND processing is completed, the main arithmetic circuit 13 counts and stores the number of "1" level signals using an internal counter. When this count ends, the corner detection means shifts the position of the corner detection pattern P by one pixel in the ” Count the number of levels. Thereafter, in the same manner, the corner detection pattern P is shifted one pixel at a time to calculate the count number of the level "1" within the area frame Wl. In addition, the count number of level "1" is calculated|required also in the area|region frame W2 similarly to the effect|action in the area|region frame W1. Next, the corner detection means is the area frame W! The position of the pattern P where the count number at each position of the corner detection pattern P is zero or less than a predetermined count number and is the most at the upper left end, that is, the smallest value of the coordinates (x, y) is detected. and,
If the corner detection pattern at this position is Pl, the contact position between this pattern P1 and the shadow image 2b is determined,
Further, from these contact positions, the corner position coordinates of the shadow image 2b, that is, the corner position coordinates of the chip component 2 (Xa+ya)
is required. On the other hand, in the area frame W2, the count number at each position of the corner detection pattern P is zero or less than the predetermined count number, and the lower right corner, that is, the coordinates (x, y)
The position of the pattern P that has the largest value is detected. Then, if the corner detection pattern at this position is Pl, the contact position between this pattern P2 and the shadow image 2a is determined, and from these contact positions, the corner position coordinates of the shadow image 2a, that is, the corner of the chip component 2 are determined. Position coordinates 0 (1
), yb) are obtained. Next, the main processing circuit 13 executes the recognition means. That is,
The recognition means calculates the center position (xc, yc) of the chip component 2 and the size l, w of the same component 2 from the above-mentioned corner position coordinates (Xa. 3/a) (xb, yb).
and the slope θ are calculated and determined. That is, the center position (xc, yc) of the chip component 2 is obtained by calculating xc m (xa + xb ) /2 yc = (ya + yb ) /2, and the size N + z is calculated by calculating p-z II cosθC wmz 0 s1nθC It can be found by calculation. Note that z − (x
a-xb)2+ (ya-yb)2. or,
The inclination θ of the chip component 2 is θ − tan'l (x a − x b) + (y a
-yb)l-θC. However, the main arithmetic circuit 13 compares the center position (xe, yc) of the chip component 2, the size 47, w, and the tilt θ of the chip component 2 with the design data, and determines whether the chip component 2 is mounted in the original position. Determine if there are any. Here, if the chip component 2 has fallen off, the shadows 2a and 2b will not occur, so the shadow 2
It is determined whether the chip component 2 has fallen off based on the presence or absence of a and 2b. Thus, in the above embodiment, each lighting device 4.5
are turned on one by one, image data D3 of the shadow of the chip component 2 is obtained from the difference between each image data at this time, and the corner detection pattern P is scanned against this image data D3 to detect the corner of the shadow part. Since the center position and inclination of the chip component 2 are determined from this corner, the shadow 2
a. By using only 2b, a shadow 2a is formed by the wiring pattern, printed characters, etc. formed on the circuit board 1. Even if 2b is not continuous and is missing, each corner can be detected and the position of the chip component 2 can be accurately recognized. Moreover, the amount of processing required to recognize the position of the chip component 2 is small, and the position of the chip component 2 can be recognized at high speed. Furthermore, in corner detection, since the corner detection pattern P is circular, each corner can be reliably detected even if the chip component 2 is tilted. Furthermore, in addition to recognizing the center position of the chip component 2 and its inclination, it is also possible to detect whether the chip component 2 has fallen off. It should be noted that the present invention is not limited to the above embodiments, and may be modified without departing from the spirit thereof. For example, the corner detection pattern is not limited to a circular shape, but may have a shape that contacts the shadows 2a and 2b, such as a semicircle or a quarter circle. In addition, the shadow part and corner detection pattern of the shadow image data D3 were set to the "1" level, but these
Processing may be performed at OJ level. The corner detection means may detect corners by other processing. [Effects of the Invention] As detailed above, according to the present invention, it is possible to provide a highly reliable object position recognition device that can recognize the position of an object without being affected by the wiring pattern or color on the circuit board. .

[Brief explanation of the drawing]

1 to 6 are diagrams for explaining an embodiment of the object position recognition device according to the present invention, in which FIG. 1 is a configuration diagram, FIG. 2 is a schematic diagram of a corner detection pattern, Figures 3 and 4 are schematic diagrams of image data when light is irradiated by each illumination device, Figure 5 is a schematic diagram showing image data of shadows, and Figure 6 is for explaining the function of position recognition. and FIG. 7 are diagrams for explaining the prior art. DESCRIPTION OF SYMBOLS 1... Circuit board, 2... Chip component, 3... Imaging device, 4.5... Lighting device, 6... Lighting switching circuit,
8, 9... Image memory, 11... Subtraction circuit, 12.
...Binarization circuit, 13... Main calculation circuit, 14... Pattern memory, 15... Logical product calculation circuit. Applicant's representative: Patent attorney Takehiko Suzue Figure 2 Figure 3 Figure 4

Claims (1)

[Claims] An imaging device disposed above an object to be recognized disposed on a substrate, each illumination device disposed diagonally upward and facing each other when viewed from the object to be recognized, and one of these illumination devices individually. a lighting control means for lighting up each of the lighting devices; an image memory for receiving an image signal from the imaging device and storing it as each image data when each of the lighting devices is turned on; and a difference between each image data stored in the image memory. a difference image calculation means for obtaining image data of the shadow of the object to be recognized by calculating the difference image; An apparatus for recognizing the position of an object, comprising: corner detection means for detecting a corner; and recognition means for recognizing the position of the object to be recognized from the corner detected by the corner detection means.