JPH03142572A - Picture information processor - Google Patents

Abstract

PURPOSE:To economize a memory and to shorten time required for access by providing a processed result memory area to store the result of a processing, which is executed to density data, at the same address as each address to store the density data. CONSTITUTION:At the same address as each address to store the density data at each sampling point, a first flag F1 is provided to store contour data to be obtained from the density data, and a second flag F2 is provided to store area recognition data. Accordingly, at the same address to store the density data of each picture element, the contour data and the area recognition data can be stored and an arithmetic processing is easily executed.

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Field of Application (1) The present invention relates to an image information processing device. For example, it can be used in devices that apply image processing technology, such as clothing advertising design and X-ray analysis.

[Prior art] For example, image information input from a video camera, etc.
If you sample in a dimensional matrix and quantize the image density at each sampling point in tens to hundreds of gradations,
Digital representation of density data can be obtained. This concentration data can be stored in a memory IC or disk device, and
It can be easily processed by operating the CPU and DSP (digital signal processor) elements.

FIG. 14 shows a general image information processing device using a CRT as an image output device. In the same figure, when an image signal is input to the image input circuit 2 from a video camera 1 or the like,
The image input circuit 2 A/D converts the image signal and stores it in an image data storage memory 3 as digitally expressed density data.
The density data stored in the four-image data storage memory 3 is accessed by both the image processing circuit 4 and the display control circuit 5.

The image processing circuit 4 is composed of a CPU, a DSP element, etc., and can process the density data stored in the image data storage memory 3 in various ways. The display control circuit 5 takes out display data from the image data storage memory 3 and inputs it to the D/A conversion circuit 6.

The display data input to the D/A conversion circuit 6 is converted into D/A converter 6.
After /A conversion, the image output from the display control circuit 5 (
The image synchronization signal is input to C'RT7, where it is displayed as an image.

FIG. 15 shows a method of sampling image information in a two-dimensional matrix. Each sampling point is called a pixel. Here, sampling is performed at nXm (for example, 256X256> points, and each pixel (1,1) to (256,256
The density data of > is quantized with 8 bits (=25611 tones).

Conventionally, when the density data of each pixel is allocated to the image data storage memory 3, one pixel's density data (8 bits) is allocated to one address, as shown in FIG. That is, one pixel's density data (8 bits) is allocated to one address consecutively from address 4001H of the image data storage memory 3.

[Problems to be Solved by the Invention] By the way, when, for example, a region of an object is recognized from the density data stored in the image data storage memory 3 and the density value within that region is changed, the outline is determined from the density data of each pixel. The following procedure is generally adopted: 1) Recognize the area inside the contour, and 2) Change the density inside the area recognized in (2) above.

However, in the conventional method, the contour data obtained when the contour is extracted in (1) and the area recognition data obtained when the area inside the contour is recognized in (2) must be stored in separate memory areas. This has the disadvantage that the memory area must be increased and access takes time.

SUMMARY OF THE INVENTION It is therefore an object of the present invention to provide an image information processing apparatus that can solve these conventional drawbacks, save memory, and shorten the time required for access.

[Means for Solving the Problems] Therefore, the present invention provides image data in which image information is sampled in a two-dimensional matrix, and density data obtained by quantizing the image@density at each sampling point is stored for each address. In an image information processing device that includes a storage memory and performs arithmetic processing on the density data stored in the image data storage memory, each address is the same as each address for storing the density data at each sampling point of the image data storage memory. The present invention is characterized in that the address is provided with a processing result storage area for storing the results of processing performed on the density data.

[Function] A processing result storage area for storing the results of processing performed on the density data is provided at the same address as each address for storing the density data at each sampling point. The results of the processing performed can be stored at the same address as the density data. Therefore, memory can be saved and the time required for access can also be shortened.

[Example] Hereinafter, an example of the present invention will be described based on FIGS. 1 to 13.

FIG. 1 shows the bit configuration of the image data storage memory 3 of this embodiment, and FIG. 2 shows the main parts of the image information processing apparatus. In addition,
In FIG. 2, the image input circuit 2 in FIG.
The description of the display control circuit 5 and the display control circuit 5 is omitted.

As shown in these figures, in this example, the concentration data (here, 6 bits) at each sampling point mentioned above is
Each address 4001F stores .

A processing result storage area for storing the results of processing performed on the density data is allocated to the same address as 4002H~.

Here, the first flag F1 stores contour data obtained from density data, and the second flag F1 stores area recognition data.
The flag F2 is assigned to each flag.

Therefore, image processing using this will be explained in comparison with the conventional method.

Here, the model shown in FIG. 3 is set. This model shows the state in which humans are captured by vision. Here, the shaded portion A of the object is assumed to have a high density, and the surrounding ground portion B has a low density.

FIG. 4 shows a digital image data representation of the model shown in FIG. 3, that is, density data obtained by quantizing the image quality at each sampling point.

Therefore, let us consider a case in which a portion A of "'object'" is extracted from the data in FIG. 4 and the density value is changed. To implement this process, the following algorithm, which is commonly used in object extraction, is used here.

■Extraction of object outline ■Recognition of the area inside the contour ■Change the density of the area recognized in ■ above Hereinafter, these will be explained in order.

1. Extracting the outline of an object When extracting the outline of an object from digital image data, a method of scanning the image using a spatial difference filter 11 shown in FIG. 5 is generally used.

This converts the density data of four adjacent pixels to a.

If b, c, d, la dl+1b-cl -a
becomes.

FIG. 6 shows the results obtained by scanning the digital image data of FIG. 4 with the spatial difference filter 11 of FIG. FIG. 7 shows the binarized data of FIG. 6 using "15" as a threshold value. Here, the pixels marked with "1" are the outline of the object.

Since the final purpose of this processing is to convert the density of part A of the "object" in the image of FIG. 4, the data of FIG. 4 cannot be erased. Therefore, in the conventional method, a separate memory area must be provided for storing contour data, and the contour data of FIG. 6 or 7 must be stored there.

In this embodiment, since the configuration is as shown in FIG. 1, the contour data shown in FIG. 7 can be stored in the first flag F1. In other words, each pixel's density data (6 bits)
It is possible to store contour data for identifying whether each pixel is a contour at the same address where the pixel is stored.

■Recognition of the area inside the contour There are several ways to recognize the inside of the contour from the data shown in FIG. Here, a method based on the principle shown in FIG. 8 is used. When you cross an outline from the outside of an object and enter the inside of the object, you cross one outline. To get out, you have to cross another contour. In the case of a curved object as shown in FIG. 8, the object may pass through the inside of the object one more time.

At this time, the following relationship exists between the number of crossed contours and the determination of whether the position is inside or outside the object. That is,
As shown in FIG. 9, if the number of crossed contours is odd, it is inside, and if it is even, it is outside.

Therefore, a method for recognizing the inside of the outline in FIG. 7 will be explained using FIG. 10. First, the first row from the (1,1) pixel in FIG. 10 is scanned in the horizontal direction.

At this time, the number of crossed contours is counted, and if it is an odd number, it is stored in the area recognition memory shown in FIG. 11 as being inside the object. When the right end of the first row is reached, the count number is cleared to "0" and the next row is scanned. All "0"s are stored in the corresponding pixel columns.

Next, in the second line of scanning, the outline is recognized at point A in Figure 10 and counted as "1", and the corresponding pixel in Figure 11 is counted as "1".
Store. Next, mass points are also pixels on the contour, but if they are adjacent like this, the count number will not be increased.
Therefore, since the count number here is "1", the first
“1” is stored in the corresponding pixel in FIG. Similarly, the count number does not increase up to the point, but the count number is “
Therefore, the corresponding pixel in FIG. Since it remains as it is, we set it as "inner P" and store "1" all the way to the right end as shown in the second line of Fig. 11. However, when it reaches the two points on the right end of Fig. 10, Since the count number at the end of the second line scan is "1", it is determined that the contour crossed in this line is a tangent line, and scanning starts in the opposite direction this time. 11th until detection
Restore °゛0′° in the diagram.

Next, in the third line of scanning, the contour is detected at the point and the count becomes "1", and it is determined that "1tll+ in ""
"1°" is stored in the corresponding pixel in FIG. 11 until the second contour is detected at the point B. Below, in the same way, the 10th
Scan to the last line of the diagram and check the R inside the object? l End knowledge. At that point, the area recognition memory becomes as shown in FIG.

In the conventional method, a separate memory area must be provided for storing area recognition data, and the area recognition data shown in FIG. 12 must be stored there.

In this embodiment, since the configuration is shown in FIG. 1, the area recognition data shown in FIG. 12 can be stored in the second flag F2. In other words, since the second flag F2 can be used as a memory for area recognition, there is no need to prepare a separate memory area.

■Conversion of the area recognized in above ■When the above process ■ is performed, the second flag F2 contains the 12th flag.
The data shown in the figure is stored. In FIG. 12, pixels with a value of "1" are considered to be internal to the object. In other words, the pixel data itself stores the determination result of whether it is inside or outside the object.

Now, let us consider the process of converting the density of the original images shown in FIGS. 12 to 4. Here, a process is performed in which "10" is uniformly added to the density of the object in the image of FIG. This includes the 12th
This can be achieved by uniformly adding 10'' to the pixel data in FIG. 4 that correspond to the pixels with a value of 1 in the figure. The result obtained in this way is shown in FIG. 13.

Therefore, according to this embodiment, the first flag F1 for storing the contour data obtained from the density data and the first flag F1 for storing the area recognition data are placed at the same address as each address for storing the density data at each sampling point. Since two flags F2 are provided, the contour data and area recognition data can be stored at the same address where the density data of each pixel is stored.

This means that there is no need to provide a separate memory area as in the conventional case, and therefore memory can be saved. Also,
Since these data are stored at the same address,
It is easy to perform arithmetic processing, and the time required for access can be shortened.

Further, since these contour data and area recognition data are stored as binary data, subsequent processing is also easy.

In the above embodiment, the first flag F1 for storing contour data and the second flag F2 for storing area recognition data are provided at the same address as each address storing the density data of each pixel. Only one of them may be used.

Further, the data to be stored is not limited to the above-mentioned contour data and area recognition data, but any processed results may be used.

[Effects of the Invention] As described above, according to the present invention, at the same address as each address storing concentration data at each sampling point,
A processing result storage area is provided to store the results of the processing performed on the density data, so there is no need to provide a separate memory area to store the processing results, which saves memory. I can do something. ′! ! ,Ta,
Since the processing results are stored at the same address as each density data, calculation processing is easy and the time required for access can be shortened.

[Brief explanation of the drawing]

1 to 13 show an embodiment of the present invention, in which FIG. 1 shows a bit configuration of an image data storage memory, FIG. 2 shows a main part of an image information processing device, Figure 3 is a diagram showing an image data model, Figure 4 is a diagram showing a digitalized state of the model in Figure 3, Figure 5 is a diagram showing a spatial difference filter, and Figure 6 is a digital image of Figure 4. Figure 7 shows the results obtained when data is scanned with a spatial difference filter, Figure 7 shows the results of contour data obtained from the data in Figure 6, Figures 8 and 9 recognize the inside of the object. Figures 10 and 11 are diagrams explaining the method for recognizing the inside of an object, Figure 12 is a diagram showing the results of recognizing the inside of an object, and Figure 13 is a diagram showing the density conversion process. It is a figure showing a result. FIG. 14 is a block diagram showing a general image information processing device, FIG. 15 is a diagram showing a sampling method of an image @trt@, and FIG. 16 is a diagram showing a conventional image data storage memory. 3... Image data storage memory, Fl... First flag (processing result storage area), F2... Second flag (processing result storage area).

Claims (1)

[Claims] (1) An image data storage memory is provided that stores density data obtained by sampling image information in a two-dimensional matrix and quantizing the image density at each sampling point for each address. In an image information processing device that performs arithmetic processing on the density data, the result of the processing performed on the density data is stored at the same address as the address where the density data at each sampling point of the image data storage memory is stored. An image information processing device characterized by comprising a processing result storage area for storing.