JPH0766438B2 - Image storage - Google Patents

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image storage device, and more particularly to an image storage device used as a preprocessing device for image processing.

(Prior Art) In a conventional image storage device of this type, addressing simply corresponding to scan line pixels of a camera is adopted. That is, as shown in FIG. 9 and FIG. 10 as examples of addressing in the case of a multi-valued image and a binary image, the addresses of pixels in the horizontal direction i and the vertical direction j are sequentially arranged, and In principle, the same addressing was used in each case.

(Problems to be Solved by the Invention) In an image processing system using such a conventional image storage device, it is time-consuming to calculate addressing in recognition of a chip around a circular object or recognition of characters arranged on the circumference. I was dying. By the way, the recognition process based on the number of black and white pixels in the window does not increase the number of windows so much.
Moreover, when the windows are made by hardware, the number of windows is limited and the shape cannot be freely formed.

(Means for Solving the Problems) Means provided by the present invention for solving the above problems are an image for reading image data of a predetermined object written in a video memory based on an address of an output of an image processing apparatus. In the storage device, an address conversion memory table, an offset value generating means, and an offset address converter are provided, and the address conversion memory table converts the address into another address according to conversion data stored in advance and outputs the converted address. The offset value generating means sets, when an offset value is a value indicating how much the image represented by the image data written in the video memory deviates from a fixed position on a predetermined screen, Center of horizontal and vertical fixed position histograms obtained for standard images Are stored in advance, horizontal and vertical histograms relating to the image represented by the image data are created as individual image histograms, and the difference between the center of the fixed position histogram and the center of the individual image histograms is calculated in the horizontal and vertical directions. Direction, and outputs the difference in the horizontal direction and the vertical direction as the offset value, the offset address converter converts the address of the output of the address change memory table by the offset value, and the video memory uses the offset An image storage device, wherein the output of an address converter is input as a read address to read the image data.

(Example) Next, this invention is demonstrated with reference to drawings.

FIG. 1 is a block diagram of the first embodiment of the present invention, which is an example of an image storage device having a function of giving an offset at the time of output and converting an address. The video signal 2 from the camera 1 is written into the video memory 4 converted into a digital signal by the A / D converter 3 as multi-valued video data by the sequential addressing 30 from the synchronizing circuit 5. The addressing of the video memory 4 in this writing is the same as in FIG. On the other hand, at the same time when the image is taken into the video memory 4, a horizontal and vertical histogram creating circuit 6 in the mask creates a histogram.
This histogram data is calculated by the CPU 7 and the positional deviation of the object from the normal position is calculated and set in the offset address converter 8. On the other hand, it is assumed that necessary address data is set in the address conversion memory table 9 from the CPU 7. When data is accessed from the CPU 7 to the video memory 4 in this state, first, the address 10 from the CPU 7 is converted according to the contents of the memory table 9 for address conversion to generate the address 11, and further by the offset address converter 8. An address 12 is generated by multiplying / dividing and adding / subtracting the address 11 according to the setting, and the data 13 of the video memory 4 is read to the CPU in accordance with this address.

An example of data read processing according to this example will be described in detail with reference to FIG. The object illustrated in this figure is the mouth of the bottle. The detection of the chip of the mouth is performed as follows. First, the dotted line portion 14 shown in FIG. 2 is the circumferential position of the mouth portion in the normal case. Now, when the actual position of the mouth is the solid line part 15,
Find the displacement of the position. In order to determine this position, the horizontal and vertical direction histogram creating circuit 6 in the mask creates horizontal histograms 17 and vertical histograms 18 created from the image data in the mask 16 in the respective directions, that is, horizontal offsets 19.
Calculate vertical offset 20. By setting this value in the offset address converter 8, it becomes possible to access the address 11 from the address 11 so that the apparent top position is in a normal position.

Further, as shown in FIG. 3, when the distance 21 from the center to the edge of the mouth is measured in each direction, the occurrence of the chip 22 can be detected because the distance 21 becomes smaller in the part where the chip 22 occurs. . However, it is time-consuming to measure the distance from the center all the time. Therefore, the measurement is performed from the inner circle 23 to the outer circle 24 of the dotted line, following the arrow 25 from the inner side to the outer side.

For this measurement, address conversion data as shown in FIG. 4 is set in the address conversion memory table 9. That is, in the address conversion memory table 9, the radial arrow from the center is at the address 11 when viewed from the address 10.
26a …… z is arranged in a sequence, and C
Access from the PU should be made sequentially.

Although the address area 27 is not used in this embodiment due to the addressing from the address 10, for example, as shown in FIG. 5, the data on the circumference having a point on the line segment 28 as a starting point other than the radial portion is sequentially arranged. It is also possible to put in. In this way, different addressing may be applied from the address 10 to the same pixel on the same screen, which allows the peripheral matching to be performed immediately.

When the output to the video memory has the memory table for address conversion and the offset converter as in the present embodiment, there are the following advantages. First, various processes can be performed by preparing a plurality of address conversion tables for an image once collected. Further, by preparing a plurality of addresses for the same pixel, it is possible to perform a plurality of processes for the same screen. Further, in general, the image shift amount, that is, the offset value is obtained after image acquisition,
By setting this in the offset converter after image acquisition, the CPU or image processing circuit can always access the object so that it is in the normal position.

FIG. 6 is a block diagram showing a second embodiment of the present invention. This embodiment has a table for address conversion on the input side of the video memory, and is used for those having standard conversion and position accuracy. In this embodiment, the video signal 2 of the camera 1 is converted into a digital signal by the A / D converter 3 and stored in the video memory 4. At this time, the sequential addressing 30 from the synchronization circuit 5
Is converted by the address conversion memory table 9,
The address 31 is given to the video memory 4, and the video data is written in the video memory 4 in accordance with this addressing. On the other hand, the data access from the CPU, the image processing circuit, or the D / A converter 32 is performed by normal addressing. As described above, even if the image captured by the camera 1 is distorted as shown in FIG. 7 (a), the image obtained by converting the data in the video memory 4 into analog data becomes normal (see FIG. b)).

(Effects of the Invention) As described above, in the present invention, the address conversion memory table allows continuous processing of video data on an arbitrary curve or arbitrary area, and image data to be processed can be rearranged serially. Therefore, there is an effect that the processing by the later hardware can be simplified. Further, by inserting an offset converter, it is possible to correct the positional deviation, and it is possible to realize a state in which the target images are always lined up at the same address conveniently and sequentially when viewed from the image processing apparatus in the next stage. .

According to the present invention, as shown in FIG. 8 (a), the characters arranged in an arc shape on the disk plate are as shown in FIG. 8 (b) when viewed from the OCR and D / A monitor in the next stage. Since the characters are arranged normally and the positioning is performed, the recognition speed and the recognition rate are improved, and the visual recognition becomes easy.

In the present invention, it is necessary to calculate and set the addressing in the address conversion table in advance, but generally, the contents of this address conversion table are set once except for the offset, and the same processing is performed on the same object. Since there is no need to reconfigure, high speed processing is possible. Further, when performing a plurality of processes on a plurality of objects, the present invention is applied to the output side, the address conversion memory table is used as a plurality of banks, and all the settings are performed in advance according to the process. It is sufficient to switch the memory table for address conversion to be used as well as the setting value of the offset converter.

As described in detail above, according to the present invention, it is possible to provide an image storage device that has a video memory and can flexibly perform complicated image processing.

[Brief description of drawings]

FIG. 1 is a block diagram of the first embodiment of the present invention, and FIG. 2 is a conceptual diagram showing position normalization by an offset converter when the embodiment of FIG. Third
The figure is a schematic diagram showing the detection principle of the chip at the top of the bottle,
FIG. 4 is a conceptual diagram showing address translation by the address translation memory table of the embodiment shown in FIG. 1, and FIG. 5 is a conceptual diagram showing a case where circumferential address translation is performed in addition to FIG.
FIG. 6 is a block diagram of a second embodiment of the present invention having an address conversion memory table on the input side, and FIGS. 7A and 7B are diagrams showing an example of conversion of an input image according to the embodiment of FIG. 8 (a) and 8 (b) are diagrams showing examples of images when the present invention is applied to OCR, and FIG. 9 is a schematic diagram showing addressing in the case of a multi-valued image in a conventional image storage device,
FIG. 10 is a schematic diagram showing addressing in the case of a binary image in a conventional image storage device. 1 ... Camera, 2 ... Video signal, 3 ... A / D converter, 4 ... Video memory, 5 ... Synchronization circuit, 6 ... Horizontal / vertical histogram creation circuit in mask, 7 ... CPU, 8 ...
... Offset address converter, 9 ... Address conversion memory table, 10 to 12 ... Address, 13 ... Data, 14 ... Normal position on the circumference of the roof (dotted line), 15 ... Actual position of the circumference (solid line part), 17 ... horizontal histogram, 18 ... vertical histogram, 19 ... horizontal offset,
20 ... Vertical offset, 21 ... Distance, 22 ... Missing, 23 ...
… Inner circle, 24 …… Outer circle, 25,26 …… Arrow, 27 …… Address area, 28 …… Circumferential origin, 29 …… Available address area, 30 …… Sequential addressing, 31 ……
Write address after conversion, 32 …… D / A converter.

Claims (1)

[Claims] 1. An image storage device for reading out image data of a predetermined object written in a video memory based on an output address of an image processing device, comprising an address conversion memory table, an offset value generating means, and an offset address converter, The address conversion memory table converts the address into another address in accordance with the conversion data stored in advance and outputs the converted address, and the offset value generating means outputs a predetermined image represented by the image data written in the video memory. When the offset value is a value indicating how far from the fixed position on the screen, the centers of the horizontal and vertical fixed position histograms obtained for the standard image of the predetermined object at the fixed position are stored in advance. The horizontal and vertical directions of the image represented by the image data Is created as an individual image histogram, the difference between the center of the fixed position histogram and the center of the individual image histogram is obtained in the horizontal and vertical directions, and the difference in the horizontal and vertical directions is output as the offset value. The offset address converter converts the address of the output of the address change memory table by the offset value, and the video memory inputs the output of the offset address converter as a read address and reads the image data. Image storage device.