JPH0421193A - Device for recognizing object - Google Patents

Abstract

PURPOSE:To rapidly find out a maximum diameter by setting up plural straight lines passing the center point of a circumscribed rectangle of a binary image and finding out a maximum mutual distance of peripheral points maximizing distances between respective peripheral points and these straight lines. CONSTITUTION:When a circumscribed rectangle extracting part 4 finds out the circumscribed rectangle 11 of a binary image 10, a maximum diameter extracting part 5 finds out the coordinates (5, 5) of the center point 9. A straight line 12a passing the point 9 with 0 deg. inclination theta and parallel with the X axis is set up. Then, the coordinates (2, 0), (3, 0) of positively maximum peripheral points and the coordinates (2, 9), (3, 9) of negatively maximum peripheral points are found out through the straight line 12a regarded as zero. Then, a straight line 12b is set up and coordinates (2, 0), (10, 8) are similarly found out. Then, a straight line 12c is set up and coordinates (0,3), (0,4), (0,5), (0,6) and coordinates (10,7), (10,8) are found out. Then, a straight line 12d is set up and the coordinates (6, 2), (8, 4) and coordinates (2, 9) are found out. Then, the extracting part 5 determines the maximum diameter. In this case, the maximum diameter is obtained from points (2, 0) and (10, 8).

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Application in Industry A) The present invention relates to an object recognition device that performs digital processing on image data of a recognition target obtained by an optical method to determine a recognition target.

[Prior Art] This type of conventional object recognition device first captures an image of a recognition target using a CCD camera or the like, captures the obtained image data and stores it in an image memory, and then uses prepared software or , the recognition target object was determined by performing binarization processing using hardware, determining the address of each pixel as a binary image of the recognition target object, and further determining a feature quantity such as the maximum diameter.

In order to find the maximum diameter, this conventional object recognition device calculates distances for all combinations of coordinates to all surrounding pixels. Hereinafter, a method for determining this maximum diameter will be explained using FIG. 6.

In Fig. 6, the binary image of the recognition target is represented by the shaded area as an aggregation of pixels (hereinafter referred to as points) that can be represented by the X coordinate indicated by 0 to lO and the Y coordinate indicated by 0 to 9. Suppose it has a shape.

First, let point (7) be the starting coordinate, and this point (2, O)
and other peripheral points, that is, (2, O) and (2, 1), (2, 0) and (2, 2), (2,
o) and (0,3), (2゜0)
and (0,4), ... (2,O)
The two-point processing of (4,1)(2,O) and (3,O) is calculated so as to go around the outer circumferential point sequence. Then, among these, the coordinate values of the two points with the maximum distance between the two points and their distance, in this case the coordinate values (2, O) and (10,
8) and store the distance @9J1.

Next, set the point (8) as the starting coordinate, and this point (2,1)
In the same way as described above, calculate the distance between two points and other outer peripheral points by going around the outer peripheral point sequence, and then memorize the coordinate values and distances of the two points where the distance between the two points is maximum. do.

Below, we will move the start coordinates in order and calculate 2 for each case.
The distance between the points is determined, and the coordinate values of the two points with the maximum distance between the two points and the distance therebetween are stored.

When the distances between two points of all the outer peripheral point sequences have been calculated and stored in this way, the one with the maximum distance is set as the maximum diameter of the binary image.

In this method, when the number of peripheral point sequences is n, the number of times the distance is calculated is given by the following formula.

In the case of the image illustrated in Figure 6, n=36, so 63
You will have to calculate it 0 times.

(Problems to be Solved by the Invention) As described above, in the conventional object recognition device, when the number of outer peripheral points is n, distance calculations must be performed nC twice by extracting and combining two outer peripheral points from among them. I had to. Therefore, if the shape of the binary image is relatively simple or the shape of the binary image is small, the maximum diameter can be determined in a small enough processing time to cause no problem. However, the shape of the binary image is complex,
If the size is large, the amount of calculation becomes enormous, which causes a problem in that too much processing time is required, resulting in a delay in recognition.

The present invention was made in order to solve the above-mentioned problems, and an object of the present invention is to obtain an object recognition device that can quickly calculate a feature amount such as the maximum diameter of an object to be recognized.

(Means for Solving the Problems) An object recognition device according to the present invention includes means for calculating a circumscribed rectangle of a binary image from an address of run length data when binarized, and a central part of the calculated circumscribed rectangle. Set multiple straight lines that pass through the point and are offset by a predetermined angle from each other, calculate the distance from these straight lines to each peripheral point of the binary image, and check if this distance is positive with the straight line as the border. Means is provided for determining, for each straight line, the outer circumferential point that is the largest on the side and the outer circumferential point that is the largest on the negative side, calculating the mutual distance of these outer circumferential points, and setting the largest one as the maximum diameter. It is something that

(Function) In this invention, a plurality of straight lines passing through the central point of the circumscribed rectangle of the binary image are set, and the distances from these straight lines to each peripheral point are calculated to find the peripheral point at which the distance is the maximum. , and calculate the mutual distance between these peripheral points and determine the largest one as the maximum diameter. Therefore, even if the shape of the recognition target is complex or the binary image is large, it can be recognized at high speed. The maximum diameter can be found by

(Embodiment) FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention.

In the figure, an image capturing section (1) is an imaging means such as a CCD camera, which images an object to be recognized, obtains a video signal thereof, and further converts this imaging signal into a D/A and outputs it.

The binarization unit (2) binarizes the digital data to obtain a binary image of the recognition target object.

In addition, the run length address data creation section (3) includes 2
The connection of pixels of the recognized object that has been converted into a value is taken as run-length address data, and the start and end addresses are calculated. A circumscribing rectangle extraction unit (4) calculates a circumscribing rectangle of the binary image based on the run-length address data. The maximum diameter extraction unit (5) uses the circumscribed rectangle of the binary image and the run length address data to find the maximum diameter of the binary image. The recognition unit (6) determines the recognition target object using the maximum diameter as a feature quantity.

Among the elements constituting this embodiment, the image capture unit (1)
, the binarization unit (2), the run-length address data creation unit (3), the circumscribed rectangle extraction unit (4), and the recognition unit (6), the details of which have already been proposed or known, or Since it can be easily inferred from digital processing, its explanation will be omitted, and the specific processing of the maximum diameter extraction section (5), which is the most distinctive part of the present invention, will be explained using FIGS. 2 to 5. Explain in detail.

First, as shown in Figure 2, a binary image (
10) is the X coordinate indicated by 0 to lO and the Y coordinate indicated by 0 to 9.
It has a shape surrounded by thick lines as a set of points that can be represented by coordinates. If the circumscribed rectangle extraction unit (4) finds the circumscribed rectangle (11) of this binary image (10), the maximum diameter extraction unit (5) calculates the locus i (5, 5) of the central point (9). seek.

Next, as shown in Figure 3(a), point (9) in the center
) and parallel to the X-axis with an inclination θ=0° (12a
) (corresponding to the fact that pixels are regarded as points, a straight pixel row is regarded as a straight line). Next, a binary image (10
) Drop a perpendicular line to the straight line (12a) from the line (12a), find the length of each perpendicular on the positive and negative sides of the straight line (12a), and find the coordinates of the peripheral point that is maximum on the positive side. (2, O)
, (3, O) and the coordinates (2
, 9), and (39) are determined and stored.

Next, as shown in Figure 3(b), a straight line (12b) rotated around point (9) with an inclination θ = 45° is set, and in the same manner as described above, the binary image (10) is Drop a perpendicular line to the straight line (12b) from the outer circumference point sequence, find its length on the positive side and negative side, and find the outer circumference coordinate (2, O) that is maximum on the positive side and maximum on the negative side. Locus 1 (10
, 8) and memorize it.

Next, as shown in Figure 3 (C), set a straight line (12c) rotated around point (9) with an inclination θ = 90°, and do the same thing as above to find the maximum value on the positive side. The coordinates of the outer peripheral point are (0, 3), (0, 4), (05), (o, s)
And the location of the peripheral point that is maximum on the negative side! (10,7),
(10, 8) and store it.

Next, as shown in Fig. 3(d), set a straight line (12d) rotated around point (9) with an inclination θ of 135°, and do the same as described above to obtain the maximum value on the positive side. The coordinates (6, 2), (8, 4) of the outer circumferential point where the values are, and the coordinates (2,9) of the outer circumferential point that are maximum on the negative side are determined and stored.

The points obtained in the above manner become a set of candidate points that give the maximum diameter of the recognition target. FIG. 4 shows a list of these candidate points.

Next, the maximum diameter extraction section (5) sequentially selects two candidate points in FIG. 4, calculates the distance between the two points, and determines the maximum distance among them as the maximum diameter. In this case, the points giving the maximum diameter are (2,O) and (10,8).

Through the above calculation operations, it is possible to reduce the time required for calculating important feature quantities in object recognition to one test compared to conventional methods, and the calculation speed is correspondingly faster.

Note that in the above embodiment, the feature quantity of the maximum diameter was obtained as in the conventional apparatus, but according to this method, the feature quantity of the Feer diameter can also be obtained at the same time. The Feret diameter refers to the width in the X direction and the width in the Y direction of the circumscribed rectangle (11) in FIG. 2.

Therefore, in this embodiment, when determining the Feret diameter of the recognition target, it is possible to specify an inclination and determine the Feret diameter corresponding to the inclination.

In other words, a straight line (
12d) The maximum length in the positive and negative directions of the perpendicular line drawn to (12d) becomes the Feret diameter perpendicular to the straight line (12d). Also,
Set a straight line (13) perpendicular to the straight line (12d), and determine the maximum length in the positive and negative directions of the perpendicular line drawn from the outer peripheral point sequence of the binary image to the straight line (13), or the horizontal direction to the straight line (12d). The Feret diameter is (14).

Furthermore, in the above example, four straight lines passing through the center point of the circumscribed rectangle and offset by 45 degrees from each other were set, and the outer peripheral point at which the distance from these straight lines was the maximum was found. If the shape of the recognition target is simple, the angle is 60° to each other.
All you have to do is set three straight lines that are offset from each other and perform the same calculation. Furthermore, if the recognition target is extremely complex, you can set six straight lines that are offset by 30 degrees from each other and perform the same calculation. do it.

〔Effect of the invention〕

As is clear from the above description, according to the present invention,
This method has the advantage that the maximum diameter, which is an important feature quantity in object recognition, can be determined at high speed, and the feature quantity called Feret diameter can also be obtained at the same time.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention.
Figure 2 is a diagram showing an example of a binary image for determining the maximum diameter in the same example, and Figures 3 (a) to (d) are explanations for explaining a specific method for determining the maximum diameter in the same example. Figure 4 is a list of candidate points extracted to determine the maximum diameter in the same example, Figure 5 is an explanatory diagram of the Fuerre diameter that is simultaneously determined in the process of determining the maximum diameter in the same example, and Figure 6 FIG. 2 is an explanatory diagram for explaining a method for determining the maximum diameter employed in a conventional object recognition device. 1): Image capture unit 2): Binarization unit 3): Run length address data creation unit 4), circumscribed rectangle extraction unit 5, maximum diameter extraction unit (6: u identification unit (9: central pixel ( 10): Binary image (11): Circumscribed rectangle (12a) to (12d): Straight line (14), (15) - Feret diameter In each figure,
The same reference numerals indicate the same or equivalent parts.

Claims (1)

[Claims] A means for capturing image data of a recognition target, a means for binarizing this image data, a means for calculating an address of run length data when binarized, and a means for calculating an address of run length data from this address.
A means for calculating a circumscribed rectangle of a value image, a plurality of straight lines passing through a point in the center of this circumscribed rectangle and deviated from each other by a predetermined angle, and calculating a distance from these straight lines to each peripheral point of the binary image. In addition to calculating the distance, for each straight line, find the outer circumferential point where the distance is maximum on the positive side and the outer circumferential point where the distance is maximum on the negative side with the straight line as a boundary, and calculate the mutual relationship between these outer circumferential points. An object recognition device comprising: means for calculating distances and determining the maximum distance as the maximum diameter of the binary image; and means for determining the recognition target based on the maximum diameter.