JPH0490078A - Centroid detector - Google Patents

Abstract

PURPOSE:To reduce an error generated by the setting of a threshold and to improve detecting accuracy by obtaining plural corrected centroids based upon data obtained by binarizing image data by using plural different thresholds. CONSTITUTION:In centroid operation to be executed by a CPU 1, image data in an image memory 8 are successively read out as a line along the (x) axis and inputted to a binarizing circuit 7. The circuit 7 can set up two kinds of thresholds H, L based upon the control of the CPU 1. The operation using the thresholds H, L is executed once in each image stored in the memory 8. Thus, plural centroid data are obtained based upon the data obtained by binarizing image data by plural different thresholds and more accurately corrected centroids can be obtained based upon the centroid data. Consequently, an error to be generated due to binarization can be reduced in image centroid computing processing and rapid processing can be attained.

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Field of Application 1] The present invention relates to a center of gravity detection device, and particularly to a center of gravity detection device that treats two-dimensional image data as binarized data and detects its center of gravity.

[Prior Art] Conventionally, in two-image recognition processing, a technique is known in which the position of the center of gravity of two-dimensional image data is detected as a reference point of a certain image pattern.

In this type of processing, for example, in order to position the printed circuit board at a predetermined position on a processing machine, a plurality of circular or other printed patterns are formed on the printed circuit board, and the center of gravity of these printed patterns is There is a known use in which the reference position is detected as a reference position, and this position is positioned at a predetermined position on a processing machine.

It is known that the position of the center of gravity of a two-dimensional image pattern, that is, a figure, can be expressed using the moment of inertia, considering image density or brightness as mass.

When considering xy orthogonal coordinates that include a certain figure, the coordinates (xG, yG) of the center of gravity are expressed as follows.

xG=Xa+om/S+H(1)yG
= Ymom /S ・-(
2) However, S: Area of the figure The image data is treated as binary because it is often used as the center point and the calculation cost is low.

In this case, the moments of inertia regarding the x and y axes can be handled as the sum of array indices (two-dimensional addresses) of pixels whose image data takes logic 1.

[Problems to be Solved by the Invention] Binarization of image data may be performed at any time, such as before storing it in the image memory, after reading it from the image memory, or before calculating the center of gravity.

However, this binarization causes an error of 1/2 pixel. For example, suppose here that the density value is quantized into values from 0 to 225, and as shown by diagonal lines in FIG. 4(a),
Assume that an image of an object exists on a CCD matrix.

The image in FIG. 4(a) directly represents the shape of an object that should be treated as a single processing target in a chip mounter, etc.6 However, there are variations in its density distribution, and for example,
When binarized using the threshold value 244, binarized image data as shown in FIG. 4(b) is obtained;
96, the distribution is such that binarized image data as shown in FIG. 4(C) is obtained. The histogram in FIG. 5 shows the concentration distribution on the straight line indicated by the arrow LL in FIG. 4(a).

Here, the center of gravity of the object in Fig. 4(a) is
It must be detected at the position indicated by the mark in (a), but with the threshold value in Figure 4 (b), the center of gravity is at the triangular position shifted by 1/8 pixel to the right, and as shown in Figure 4 (C ) is detected at a position shifted by 1/8 pixel to the left.

An object of the present invention is to reduce errors caused by binarization and to enable high-speed processing in image gravity center calculation processing.

[Means for Solving the Problems] In order to solve the above problems, the present invention treats two-dimensional image data as binarized data, and uses a center of gravity detection device that detects the center of gravity of the two-dimensional image data as binary data. a first processing means that calculates the center of gravity of the image data based on the data binarized using the threshold; A configuration consisting of a second processing means for determining the center of gravity of the two-dimensional image data was adopted.

[Operation J] According to the above configuration, a plurality of centroid data is obtained based on data obtained by binarizing the image data using a plurality of different threshold values, and a more accurately corrected centroid is determined based on this data. I can do it.

[Example] Hereinafter, the present invention will be described in detail based on the example shown in the drawings.

FIG. 1(a) shows the structure of a gravity center detection device according to the present invention. In FIG. 1(a), reference numeral 8 denotes an image memory, CC
Multivalued image data read from an image sensor such as D is stored. For example, the data is 8 bits (0 to 2
55).

Access to the image memory 8 is controlled by an address generation circuit 9. In the address generation circuit 9, the two-dimensional address ~
Performs one-dimensional address conversion processing. First, when using a CPU that performs slow multiplication operations, the burden on the CPU can be reduced by performing two-dimensional address operations using a hardware operation circuit, and high-speed processing can be expected.

Here, the start address latch/memory 5H15, which will be described later, is
L, end address latch/memory 6H16L right and CP
It is assumed that the address generation circuit 9 is used when the image memory 8 is accessed by the UI.6 The center of gravity calculation is performed by the CPU 1 consisting of a microprocessor, etc.
This is done by A control program (described later) for the CPU 1 is stored in the ROM 2.

In the centroid calculation, the image pig in the image memory 8 is
As shown in FIG. 7B, the signals are sequentially read out as lines along the X-axis and input to the binarization circuit 7.

The binarization circuit 7 is capable of setting two types of thresholds, H and L, under the control of the CPU 1. Here, the threshold value H is, for example, 255, and L is 96 (8
bit quantization). The calculations using the threshold values H and L are performed once for each image in the image memory 8.

The output of the binarization circuit 7 is supplied to the data latch and FI in each binarization session for thresholds H and L.
Starting address latch/memory 5 days consisting of FO memory,
to 5L, end address latch/memory 6H1 to 6L 5Actually, from the beginning of each line, the address where the image data first changes from 0 to 1 is latched into the start address latch/memory 5H1 to 5L latch. ,
In that case, the address where the image data 1 has changed to O is further latched in the end address latch/memory 6H16L, and then the latch data is stored in the memory portions of the start address latch/memory 5H1 to 5L and the end address latch/memory 6H1 to 6L. is transferred.

If there is no valid data in the line, after the line processing is completed, the memory values of the start address latches/memories 5H to 5L and the end address latches/memories 6H1 to 6L are set to 0. In this way, by scanning the image data in the image memory 8 every time the threshold value HJ of the binarization circuit 7 is determined, the start address latch/memory 5th to 5L and the end address latch/memory 6H5 to 6L are scanned.
The start and end positions of the image data for each line are stored within.

After the above processing is completed, the start and end positions of the image data for each line in the start address latch/memories 5H1 to 5L and end address latches/memories 6H1 to 6L obtained for each different threshold value H, L are calculated. Threshold H
, L, and then average (eg, arithmetic average) the corrected center of gravity position.

Start address latch/memory 5H, 5L, J and end address latch/memory 6H.

The selection of 6L to 6L is performed by the CPU 1 via the selector 3.

FIG. 2 shows a control program for the CPU 1 for performing the above-mentioned gravity center calculation.

In step Sl, the threshold value of the binarization circuit 7 is set to H,
The binarization circuit 7 binarizes the l line data.

The binarized line data is stored in a buffer (not shown) or the like. Processing from step S2 onwards is performed on this line data.

Data processing for one screen is performed by a loop of one line processing of steps 82 to S6j, i and step SlO.

First, in step s2, it is determined whether or not there is valid data on the line, and if there is no valid data, step 510 is performed.
The end of the I-line scan is determined.

If one line scanning is completed in step 510, the process advances to step S7, and if one line scanning is not completed, step S2 is repeated.

If valid data exists in step S2, the leading data is latched in the leading address latch/memory 5H in step S3. In this case, the process moves to step S4, and the end of valid data (inversion to 1-0) is monitored through a loop of steps S4 and S5.

If the line data ends in step S5, the process moves to step S6, and the start address latched in step S3 and the end address latched in step S4 are stored in the memory portion of the start address latch/memory 5H1 end address latch/memory 6H. 6 In step S7, the end of one screen is detected, and if it is the end of one screen, the process moves to step S8, otherwise the process from step Sl is repeated.

If the end of one screen is detected in step S7, the process moves to step S8, and the total sum of each line and the first moments of the X and y axes are calculated from the start and end addresses of the data6.Then, in step S9, Perform calculations using equations (1) and (2) to find the center of gravity (x HG, y H
Find G).

The processing in steps Sll to S20 is performed in steps Sll to S20, respectively.
Here, the start address latch/memory 5H1, the end address latch/memory 6H are replaced with the start address latch/memory SL, the end address latch/memory 6L, and the center of gravity (x LG,
yLG) is performed. Since the processing contents are exactly the same, the explanation will be omitted here.

In step S21, the center of gravity position (x HG, y H
G)! 3 and the arithmetic mean of (x IG, y LG) (
(xHG+ xLGl/2, (xHG+ yLG)
/2) is calculated as the corrected center of gravity (xG, yG).

When performing the above processing, the center of gravity of the image data in Figure 4 is found at the position of the X mark in Figure 3 (here, the center of gravity (x LG, y LG) when the threshold [L = 96 (indicated by a triangle). The center of gravity determined by the above method is more accurate and closer to the center of gravity shown in FIG. 4(b) as shown.

Therefore, according to the above embodiment, the accuracy of detecting the center of gravity can be greatly improved compared to the conventional method. Also, the effective area of the image is treated as a range surrounded by the start/end addresses of the line data.

Sequential processing in real time for each line is possible, and real-time and high-speed center of gravity detection processing is possible. Furthermore, as shown above, the effective area of the image is
The end address can be thought of as the contour line of the image, but since the center of gravity position is detected assuming that everything within this contour line is valid image data, there may be defects in the scanned image itself, foreign objects, etc. This has an excellent effect in that even if a defect occurs in the image data during the scanning process, the original position of the center of gravity of the target image pattern can be accurately detected.

In this example, the effective area of the image is
The range surrounded by the end address is assumed, and there are no cases where the image data is hollow. However, the method of calculating the center of gravity using different thresholds and averaging them is the method that refers to the hollow parts of the image as well. But it can be used. In this case,
In each threshold calculation, not only the start/end addresses of image data but also the addresses of all valid (1) data are added and divided by the area.

Moreover, although the example using two threshold values H and L has been shown above, the number of threshold values may be three or more. Also,
Correction using the center of gravity obtained from a plurality of threshold values may be performed by a method other than arithmetic averaging. By increasing the threshold, the calculation time increases, but the error of the obtained center of gravity becomes smaller.

[Effects of the Invention] As is clear from the above, according to the present invention, in a center of gravity detection device that treats two-dimensional image data as binarized data and detects its center of gravity, the image data is processed by a plurality of different thresholds. a first processing means that calculates the center of gravity of each image data based on the data binarized using the second processing means; and a two-dimensional image input from a plurality of center of gravity data of the image data obtained for each threshold value by this processing means. Since it adopts a configuration consisting of a second processing means for determining the center of gravity of the data,
In order to obtain multiple corrected centroids based on data obtained by binarizing image data using multiple different threshold values, errors caused by threshold settings are reduced, and detection accuracy is increased (
It has a great effect of improving.

[Brief explanation of the drawing]

FIG. 1(a) is a block diagram showing the configuration of a center of gravity detecting device adopting the present invention, FIG. 1(b) is an explanatory diagram showing the image data scanning direction of the center of gravity detecting device adopting the present invention, and FIG. 2
The figure is a flowchart showing the control procedure of the device in FIG. 1, FIG. 3 is an explanatory diagram showing the center of gravity detection process of the center of gravity detection device adopting the present invention, and FIGS. 4(a) to (c) are conventional An explanatory diagram showing the technical problems, FIG. 5 is a graph of the image density distribution in FIG. ...Start address latch/memory 6--End address latch/memory 7--Binarization circuit 8-Image memory 9--Address generation circuit

Claims (1)

[Claims] 1) A center of gravity detection device that treats two-dimensional image data as binarized data and detects its center of gravity based on data obtained by binarizing the image data using a plurality of different threshold values. A first processing means for determining the center of gravity of each image data; and a second processing means for determining the center of gravity of the input two-dimensional image data from a plurality of center of gravity data of the image data determined for each threshold value by this processing means. A center of gravity detection device comprising: 2) The second processing means sets the arithmetic mean of the plurality of centroid data of the image data obtained for each threshold value by the first processing means as the centroid of the input two-dimensional image data. The center of gravity detection device according to claim 1. 3) Extraction means for inputting image data in units of a predetermined scanning line and extracting the first and last coordinates of effective image data corresponding to the effective image pattern in the scanning line; Data in a predetermined format is obtained and accumulated from the first and last coordinates of the obtained valid image data, and after scanning is completed, the entire first and last coordinates of the valid image data obtained for each scanning line are calculated based on the cumulative result of the data in the predetermined format. Claim 1 or 2 further comprising means for detecting the coordinates of the center of gravity of the outline pattern of the corresponding effective image pattern as corresponding to the coordinates of the center of gravity of the image pattern. Center of gravity detection device.