JPS59114407A - Area measurement system of object - Google Patents

Abstract

PURPOSE:To enable the measurement of area quickly by counting pixels of an object to measure the area thereof after the identification of the object. CONSTITUTION:An image read out from a video camera PDD is binary coded with a binary coding circuit 7 and written into a picture memory 2 with a memory controller 8. Processor 1 executes a pretreatment according to an image processing program of a program memory 3. Then, the processor 1 executes the process of measuring area under the control of the image processing program. Then, the area of the desired object is received beforehand with the processor 1 through an interface circuit 5 and a bus 6 from an external robot controller and compared with the areas of objects stored in a storage area 4b of a data memory 4 to identify the object of the area matching it or most similar thereto.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a method for measuring the area of an object used for object identification, and particularly relates to a method for measuring the area of an object for determining the area of the object from an image obtained by an imaging means. The development of robots has been remarkable, and they are equipped with various sensors, and the outputs of the sensors are used to perform various controls. In particular, industrial robots equipped with visual functions can recognize objects and selectively perform work on them, greatly increasing the functionality of the robot. FIG. 1 shows a configuration diagram of an industrial robot having such a visual function, and objects A1 and A2 within the imaging range of the video camera PDD.
2 shows a diagram in which the robot's hand HD performs a task.

The images from the video camera PDD are processed and the objects A1 and A are
2, identifies the position of the object, positions the robot's node HD on the selected object A1 or A2, and performs work such as picking up the object or fitting it with another object. In this way, objects are identified from the image of the video camera PDD by a combination of shape recognition and area recognition. To identify the area of the object, as shown in Figure 2, an image within the imaging range is captured using a video camera PDD.
Binarize and store in screen memory. At this time, during binarization, 10# pixels are stored for the background color, and 1" are stored for pixels in the part of the object. Next, the screen memory is read and scanned to calculate the number of pixels @1" on each scanning line. By counting, multiple areas are measured, and the objects A1 and A2 are identified by this area. At this time, as shown in FIG. 2, if objects A1 and A2 are separated and both objects do not exist on the same scanning line,
Measuring the area of individual objects is easy.

However, as shown in FIG. 6, if there are multiple objects AI and A2 on the same scanning line, or if there is a hole A3N
exists, the number of pixels P1# on one scanning line cannot be simply counted, and the problem arises that the area cannot be measured simply by scanning the data stored in the image memory. . That is, it is necessary to detect whether pixel "1#" on one scanning line belongs to object A1 or A2, or object A6 or A3N, which requires extremely complicated control. was there.

SUMMARY OF THE INVENTION Therefore, an object of the present invention is to provide an object area measuring method that can easily measure the area of an object having holes or when a plurality of objects coexist within a screen.

Hereinafter, the present invention will be explained in detail using examples.

FIG. 4, FIG. 5, and FIG. 6 are explanatory diagrams of the operation of the present invention.
FIG. 4(8) shows a diagram in which two objects A1 and A2 exist between scanning line 1 and Sn -dl, and the area of each of objects A1 and A2 is determined.

When the video camera PDD images an object like the one shown in FIG. 4, and the obtained image signal is binarized and stored in the screen memory, it becomes as shown in FIG. 4 (b). That is, object parts A1, A2
Only one is set to "1", and the others are set to "0".

Here, in the screen memory, n (for example, 8) pits are assigned to one pixel, so the pixels of the object part are 'ooo
oooooi”, other pixels are “ooooooooo”
It becomes.

Next, the image processing section reads the contents of the screen memory, detects the outline, and designates the object number. That is, the image processing section scans the screen memory, sequentially reads the stored contents of each pixel of each scanning line Sβ to Sα, and compares the contents of adjacent pixels to detect a change point. Then, using the detected change point as a contour, the stored content of that pixel is rewritten from "1" to the object number. In this example, the object number must be "2" or higher, so the fifth
As shown in the figure, the contour pixels of object A1 are "2"("0000
001.0'), the contour pixel of object A2 is l'-3J
("00.000011"). The above is a pre-processing step, which makes it possible to identify objects in pixels.

Next, an area measurement step is performed. To do this, count the number of 11'' and object numbers on each scan line Sβ to Sα on the screen memory. That is, the image processing section reads each pixel of the screen memory and determines the stored contents. If the stored content is 10'', it is the background color and is not counted. Memory contents are 10.”
, if it is other than "1", it is an object number and is processed as follows. First, the image processing section prepares count value storage areas a1 to an for each object number in advance, and also prepares the object identification register shown in FIG. 6. For example, when the image processing unit scans the Smth screen memory and identifies the object number at position X4, the object identification register is changed from "0" to object number "2".
(Figure 6 @). At the same time, "1" is added to the contents of the storage area a1 corresponding to the object A1.The image processing unit then reads "1" at the position x5 in the screen memory, and detects the object corresponding to this "1". Therefore, it refers to the object identification register and detects that it belongs to object A1,
Add "1#" to the contents of the storage area a1. In this way, each time 1# is read from the screen memory, "1" is added to the contents of the storage area corresponding to the object identification register. When the content of the position of is read out, the object number is "2" and it is the outline of the sb object A1, so the content of the object identification register is set to "0" and the content of the storage area a1 is set to @1 in Fig. 6 (C1). # to add. Furthermore, each location of screen memory x10
The contents of ~X17 are read sequentially, but they are all "0"
Therefore, the image processing section does not perform any operation on the object identification register or storage area. Next, the object number "6" is read at the xlB location in the screen memory, so the contents of the object identification register are replaced with the object number "6" (see Figure 6).
Add "1" to the contents of storage area a2 corresponding to object A2. From then on, screen memory positions tX19 to X27 "
"1" refers to the object number in the object identification register and sequentially adds "1" to the contents of the storage area a2. Since the object number "ro" is read out at position X28 of the screen memory, the contents of the object identification register are set to roJ and the storage area a
Add ``1#'' to the contents of 2. After that, the screen memory location χ
Since all values up to α are "Ω", the contents of the object identification register range from 0 to 1, and no change occurs in the contents of the storage area.

When the area measurement (counting of pixels in the object portion) for one scanning line Sm is completed in this manner, area measurement is performed for the next scanning line Sm11 in the same manner.

When similar area measurements are performed for all scans @si to Sα, the number of pixels (i.e., area) of the corresponding object is stored in each count storage area a1 to an.

Next, an example applied to the object A3 having the hole A3-N & shown in FIG. 6 will be described with reference to FIGS. 7 and 8. After capturing an image of object A3 with the video camera PDD and storing it in the screen memory, after performing the above-mentioned preprocessing, the image shown in Fig.
" is the object number "-6" of hole A3N on the outline of hole A3N.
" is set. The operation at a certain scanning line Sα is similar to that described above: object number "2" is read at position ) "1" is added to the contents of a6, and thereafter, each time "1-1" is read from the screen memory, "1" is added to the storage area a3 with reference to the contents of the object identification register.

When the contents of location X10 of the screen memory are read, it is "object number 1-3", so in FIG. 1" is added to the contents. Next, "o" is read from screen memory locations X11 to X15, but since the contents of the object identification register are not rOJ, the latest written object numbers 1-6" are read. The content of storage area a4 of object (hole) 'A3N' is determined as ''1''.
to add. Object number "3" at position X16 in screen memory
is read out, so set "6" in the object identification register to "0".
(Replacement (Fig. 8 0), contents of storage area a4
1# Add. Next, at location X17 in the screen memory,
When "1" is read out, the contents of the object identification register are referenced and °'i" is added to the contents of the storage area a6. Thereafter, the object identification register is referenced for one scanning line Sα in the same manner.
II I IP addition to the contents of the corresponding storage area is repeated, object number "2" is read out at position X22, and "2" in the object identification register is replaced with "0". If this is repeated for each scanning line 82 to Sn in the same way as in the previous example, the area of object A3 will be stored in storage area a and the area of object (hole) A3N will be stored in storage area a4. Become.

In this way, the area can be easily measured based on the contents of the object identification register.

In the example of FIG. 7, the "O" inside the hole A3N may be replaced with "11" by preprocessing.

FIG. 9 is a block diagram of one embodiment of the present invention, in which 1 is a processor (processing device) that executes image processing operations according to an image processing program to be described later, and 2 is a screen memory. 8 bits are allocated to each pixel. 6 is a program memory, which is composed of read-only memory (ROM) and stores the image processing program, and 4 is a data memory, which can be accessed randomly. It consists of a memory (RAM) and includes a count value storage area 4b and an object identification register 4a, 5 is an interface circuit, which is used for communicating with the outside, and 6 is a data/address bus. 7 is a binarization circuit, which binarizes the image signal of the video camera PDD ("1N or 0"), 8 is a memory controller, and the output of the binarization circuit 7 is changed to 8 bits and written to the screen memory 2.

Next, the operation of the configuration of the embodiment shown in FIG.
is written as shown in FIG. 4(B).

Processor 1 first executes preprocessing according to the image processing program in program memory B. That is, the screen memory 2 is scanned, each pixel on the scanning line is sequentially read out via the bus 6,
The contour is extracted by comparing adjacent pixels, and the pixels of the contour portion on the screen memory 2 are replaced with object numbers. In this way, the contents of the screen memory 2 become as shown in FIG. 5 (or FIG. 7). The processor 1 then stores the storage area 4 of the data memory 4.
Storage areas a1 to an corresponding to the object numbers are set in b.

When the preprocessing step is completed, the processor 1 executes an area measurement step under the control of the image processing program. That is, each pixel on each scanning line of the screen memory 2 is read out via the bus 6, and it is determined whether the read memory content is "0" or other than "0". If the read memory content is "0", the next pixel is read and the same determination is made.

If the read pixel is not "0", it is determined whether it is an object number or not. In this embodiment, since "2" or more is used as the object number, if it is other than "1", it is written into the object identification register 4a via the bus 6 as the object number.

At the same time, "1" is added to the contents of the storage area 4b corresponding to the object number. If the read pixel is "1", the object number has already been written in the object identification register 4a, so the processor 1 reads the contents of the object identification register 4a of the data memory 4 via the bus 6 and sets the object number. is recognized and "1" is added to the contents of the corresponding storage area 4b. When the same object number is read again, the processor 1 looks at the contents of the object identification register 4a in the data memory 4 and erases it by replacing this ζ body number with 10, since it has been written once. , the content of the corresponding storage area 4b is “1”
to add.

In this way, if the processor 1 performs this process on all the scanning lines of the screen memory 2, the number of pixels (ie, the area) of each object will be obtained in the storage area 4b of the data memory 4.

In order to identify a desired object using this area, the area of the desired object is received in advance by the processor 1 from an external robot control device via the interface circuit 5 and the bus 6, and is then stored in the data memory 4. It is sufficient to compare the area of each object stored in the area 4b and detect the object that matches or has the closest area. Furthermore, to obtain the coordinates of this object,
The processor 1 calculates the center coordinates of the detected object in the screen memory, converts it into real coordinates on the plane on which the object is placed, and sends the coordinates to the robot control device via the bus 6 and interface circuit 5. As a result, the probot control device performs operations on the object by controlling the positioning of the hand HD (FIG. 1) at this coordinate position.

As explained above, according to the present invention, after storing the image signal from the imaging means in the screen memory, the control section measures the area of the object by counting the pixels of the object from the stored memory contents. In this method, the screen memory is rewritten by giving an object number to the pixel corresponding to the outline of the object, and when counting pixels, when the object number is read out, it is set in the object identification register.
Since the pixels of an object are added to the pixel count value of the corresponding object by referring to the contents of the object identification register, even if multiple objects are present on the readout scanning line, it is possible to easily distinguish them and count the pixels. This has the advantage that objects can be identified simply by reading out the screen memory without having to separately store the object's outline position address, and there is no need to compare addresses each time a pixel is read. This has excellent practical effects in that the memory capacity can be reduced accordingly, complex image processing programs can be simplified, and area measurement can be performed in a short time.

Although the present invention has been explained using one example, the present invention is not limited to the above-mentioned example, and various modifications can be made in accordance with the gist of the present invention, and these are excluded from the scope of the present invention. isn't it.

[Brief explanation of drawings]

Fig. 1 is a configuration diagram of an industrial robot with a visual function, Figs. 21 and 3 are explanations of the principle of measuring the area of an object, Fig. 8 is an illustration of another embodiment of the present invention, and Fig. 9 is a diagram of the invention. FIG. 1 is a block diagram of an embodiment of the present invention. In the figure, PDD...video camera (imaging means)
, A1, A2, A3...Object, 1...Processing device, 2...2...Screen memory, 4...Data memory, 4a...Object identification Register, 4b...
... Count value storage area. Agent Patent attorney Minoru Tsuji and 2 others v l Figure 3 Figure 2 Figure 4 (,4) "91□- (B)

Claims (1)

[Claims] It has an imaging means for taking an image of an object, a screen memory for storing an image signal from the imaging means, and a control section for reading and scanning the screen memory and counting the stored pixels of the object. In an object area measurement method in which the area of the object is measured by a number, the control unit extracts the outline of the object from the contents of the screen memory, stores an object number in a pixel corresponding to the outline of the screen memory, and When the object number is read out by the readout scan, the object number is set in the object identification register, and the control unit ii! 2. A method for measuring the area of an object, characterized in that each time the accumulated data is read out, the object identification register is referred to and added to the pixel count value of the corresponding object.