JPS62169279A - Picture data processing method - Google Patents

Abstract

PURPOSE:To obtain the exact number of objects by finding the number of high luminance areas inside a fringe in a picture data after a binarization. CONSTITUTION:A video signal outputted from a TV camera 11 is converted to a digital signal of '1' or '0' at a picture input part in a picture processor 100. The digital signal is inputted to a picture memory 14 through a switching circuit 13. When the input of one picture to the picture memory 14 is completed, the digital signal at the picture memory 14 is read out, and is inputted to a picture processor 15. The picture processor 15 recognizes the number of the objects by finding the number of the high luminance areas inside the fringe in the picture data after the binarization.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an image data processing method, and particularly to an image data processing method suitable for determining the number of objects having curved surfaces.

[Conventional technology]

Conventionally, an object to be processed is imaged, the image data obtained by the second imaging process is binarized, an island is found in the image data after the binarization process, and a label is attached to the island. Image data processing such as determining the number of islands and further determining the average area of the islands has been carried out.

For example, Japanese Patent Application Laid-open No. 58-37773 discloses a technique for labeling each station that exists as multiple patterns on one screen and checking whether or not islands are connected. There is.

However, it is not possible to determine whether the connected islands are really one island or whether they are recognized as connected islands because they are adjacent to each other, and no consideration has been given to this point. .

[Problem that the invention seeks to solve]

In the above-mentioned conventional technology, it is difficult to know whether what is recognized as a connected island is actually a single island, or whether it is recognized as a single island because multiple objects are close to each other. It is not possible to determine whether it is present or not. As a result, the actual number of objects in the image cannot be recognized by the number of islands obtained by image processing. Even if it is recognized, it will be inaccurate. Further, when calculating the average surface area, circumference, etc. of objects in an image, accurate values cannot be obtained if the actual number of objects is calculated from the number of islands.

An object of the present invention is to provide an image data processing method that can obtain an accurate number of objects from connected islands.

[Means for solving problems]

The above objective is achieved by determining the number of high brightness areas included in an island and separating connected islands based on that number.

In other words, when an image like the one shown in Figure 3 is binarized into white and black, high brightness areas (
A hole) is formed, as shown in Figure 4. When there are only one object, each island is independent and there is only one hole (high brightness area) due to specular reflection.

Further, when objects are adjacent to each other, the binarized images due to the shadows of the objects are connected n1 to apparently become one island. However, the holes caused by specular reflection are independent of each other and represent the characteristics before binarization.

Therefore, by counting the independent high-intensity areas (holes), the accurate number of islands can be determined, and the accurate number of objects can be recognized.

[Effect]

As mentioned above, by calculating the number of high brightness areas inside each station for the island obtained from the image data after binarization processing, even if multiple objects are close, the number of multiple objects can be calculated. be able to. This allows accurate numerical values of objects within the image to be obtained.

〔Example〕

An embodiment of the present invention will be described below.

FIG. 2 shows a configuration example of a video processing device to which the present invention is applied. A TV camera 11 is installed as an imaging device that converts an image of an object into an electrical video signal.
The image input section n of the video processing device 11 converts the video signal outputted from 1 into a digital signal of 0.degree.1" or 0.0".

The video signal converted into a digital signal at the image processing department is input to the image memory 14 via the switching circuit 13. When the input for one screen is completed, T is stored in the image memory 14.
The input from the V-camera 11 is blocked by a switching circuit and switched to allow access from the arithmetic processor 15.

The arithmetic processor 15 reads the value of @1° or "O" from the image memory 14 and labels (signs) it in the RAM memory 19 as shown in FIG.

The labeling operation is performed according to the procedure shown in the flowcharts of FIGS. 1A and 1B.

First, the data stored in the image memory 14 corresponds to the unlabeled state in FIG. They are arranged as follows.

The image processor 15 uses Y
The axis is fixed and the memory is read in the X-axis direction.

First, check whether the edge of the screen is "1", and if it is "1", write the X coordinate address into the change point table in the RAM memory 19, assuming that the edge has changed from "°O" to '1°. . In addition, if it is not ``1 degree,'' update the read address sequentially until ``1 m is found, and at the time it is found, write the X coordinate address value in the change point table (first
Processing of steps F101 to F107 in Figure A). After "°l" is discovered, similarly search for a point that changes from "1" to "0" and increase the readout X address of the image memory 14, and when a change point is detected, the address value is added to the change point table. (processing of FIO8 to F113 in Figures 1A and 1B).Next, it repeats in the order of "1" and then 10, and when the X coordinate final address of the image memory 14 is detected, the X final address is written. Insert the value into the change point table (F102.F
201. Processing of F108 and F201).

In this way, when one line is completed, the Y-axis address is set to 1.
Shift it step by step and record it in the change point table in the same way.

The change point table is assigned to each Y axis as shown in Figure 6, and storage addresses are assigned for each Y axis, such as a change from °0" to "1" and a change from 1° to "0° (hereinafter referred to as echo). has been decided.

When the determination of each change point table is completed, the arithmetic processor 15 examines the degree of connection of each change point table.

In other words, in Figure 5, the change point address of the Y1 axis is
, 2't', and the range of ° 1'' is from X5 to XLZ.Here, '1' is a new occurrence, so label 1 is between X5 and X12.About the Y2 axis When we examine the change point address, we find that the area from X4 to X8 is "11". Here, if we look at the overlap with the °1" section of the Yl axis, we can see that it overlaps in the section from X5 to
Since the 11° section from 0 to X13 overlaps, it is labeled as 1. Here, the section from X8 to 0'' from XlO is registered as a branch label, and a temporary label 2 is attached.

I went in the same way and created two branches on the Y5 axis.
” have merged, temporary label 2 is attached as official label 2 as a high brightness area (hole) within the island.

2, the overlapping state is changed to occurrence 1 as shown in Figure 7.
There are 4 types of branching, merging, and extinction, and 1 type of branching and merging.
” to label the internal hole if it occurs within the label.

In this manner, when labeling is completed, the total number of holes is determined.

For objects with curved surfaces, such as spherical objects, there is always one high brightness area (hole) due to specular reflection, so when calculating the average child species of objects within the screen, for example, By dividing the sum of the labeled areas of 11" and the sum of the areas of the holes by the number of holes, an accurate average area can be obtained without being affected by specular reflection.

According to this embodiment, the number of islands included in an image is determined by a simple program, and statistical processing of images, which has been difficult in the past, becomes possible.

In other words, by connecting islands, it becomes possible to calculate the average area of an image and the length of one side of the perimeter, which was previously impossible, and by multiplying pixels by weights, it becomes possible to find the area and length of an object. .

〔Effect of the invention〕

As explained above, according to the present invention, it is possible to obtain the accurate number of objects included in the connected islands from the connected islands, and it is possible to obtain the accurate number of objects in the image.

[Brief explanation of drawings]

1A and 1B are operation flow diagrams in an embodiment of the present invention, FIG. 2 is a hardware configuration diagram in an embodiment of the present invention, FIG. 3 is a diagram showing an example of an object in an image, and FIG. 4 is a diagram showing the object in Figure 3 that has been binarized, Figure f55 is a diagram that shows the image data after the binarization process has been labeled, and Figure 6 is the change point table configuration. The figure shown in FIG. 7 is a diagram for explaining the labeling state. 11...TV camera, 100...image processing device, 12...image input unit, 13...
...Switching circuit, 14... Image memory, 15...
... Arithmetic processor, 16 ... Memory %1
7 Fang 2 Fig. 3 Fig. 3 ------- ÷ 3 4-1---94 Fang 4 Fig. 5 Fig. 16 Yo Y+
Yz Fuyuo 111-cho 7ku7＼ ko line line ko + eye uin: previous line = + spinning parakeet ff' 1st line kotarora 4n

Claims (1)

[Claims] 1. In an image data processing method that images an object to be processed, processes the image data obtained by the imaging, and obtains an island corresponding to the object included in the image data, the image data obtained by the imaging is An image data processing method characterized in that the number of objects is recognized by digitizing the object and determining the number of high brightness areas inside the island in the binarized image data.