JPS60117967A - Picture processor - Google Patents

Abstract

PURPOSE:To perform the filter processing without so much delay in a data input mode by using an on-line 2-dimensional filter using a shift register. CONSTITUTION:The digital signals obtained by an A/D converter 3 are supplied successively to a shift register 8a. The output signals of shift registers 8a and 8b are supplied to shift registers the shift register 8b and 8c, respectively. Here the head 3-picture element data of registers 8a-8c are weighted by multipliers 10a- 10c, 10d-10f and 10g-10i and added by adders 9a-9c, respectively. The outputs of adders 9a-9c are added by an adder 9d. Finally 9 picture element data are weighted and added in all. These added data are divided by a divider 10 and averaged to be supplied to a frame memory 6. For the total time between the input of data and the end of filter processing, just a period of time equivalent to the input of data equivalent to nearly one scanning line is increased to the data input time.

Description

[Detailed Description of the Invention] [Technical Field of the Invention] The present invention relates to an image processing device that optically reads characters and figures by performing line scanning using a photoelectric converter, and further processes the image electronically. Regarding.

[Prior art]

An example of a conventional image processing device will be described with reference to FIG. In Figure 1, 1 is a photoelectric converter, 2 is an analog signal low-pass filter, 3 is an analog/digital converter, 4, 6
is a frame memory, and 5 is a two-dimensional filter.

Next, the operation will be explained. First, a document on which characters and figures are drawn is line-scanned and converted into an electric signal by a photoelectric converter l. This electrical signal is sent to a low-pass filter 2, where the noise is removed. Further, it is converted into a digital signal by an analog/digital converter 3. This digital signal is once input into the frame memory 4 for one screen, and when the input of data for one screen is completed, the data is read out, subjected to filter processing by the two-dimensional filter 5, and then input to the frame memory 6.

The processing contents of the two-dimensional filter 5 will be explained using FIG. In FIG. 2, A indicates the line scanning direction, and B indicates the document feeding direction. Assuming that data D(i,j) represents the density of the first pixel of the j-th scanning line, by performing a weighted average of D(i,j) and the surrounding data,
The filtered data can be determined. The weights of these data are appropriately determined depending on the purpose of image processing, such as sharpening a blurred image or enhancing contrast. For example, when determining new data by a simple average of data of D(i,j) and its surrounding eight pixels, all weights are set to 1, and the following calculation process is performed.

In this way, the two-dimensional filter uses the data D(i
, j), it is necessary to refer to the upper, lower, left, and right data.
When using a normal data input device that performs photoelectric conversion by line scanning, it is necessary to first input all data into a frame memory, read out the surrounding data together, perform filter processing, and input it into another frame memory. there were.

Therefore, one frame memory is required for one filter process, and there is a problem in that the memory capacity must be increased. Another disadvantage is that the processing time is long because filter processing is performed after all data has been input.

[Purpose of the invention]

The present invention has been made to improve the above-mentioned conventional problems, and an object of the present invention is to provide an image processing device that can shorten processing time without increasing memory capacity.

[Structure of the invention]

The image processing device of the present invention includes a means for converting an image into a digital signal representing the density of each pixel, and a means for sequentially inputting the digital signal and storing only n-1 (n is an integer of 2 or more) scanning lines. a first shift register that inputs and stores digital signals for n pixels; and a second shift register that inputs and stores digital signals for n pixels, and a
It is comprised of means for adding data for pixels and data for n pixels of the second shift register, and means for averaging the added data.

〔Example〕

An embodiment of the present invention will be described below with reference to FIG. In FIG. 3, 7 is a two-dimensional filter according to the present invention using a shift register, a so-called online two-dimensional filter (hereinafter referred to as such). In this figure, circuits with the same reference numerals as the circuits in FIG. 1 have the same functions as those circuits.

Next, the operation of this image processing device will be explained.

First, a document (not shown) is line-scanned, converted into an electrical signal by a photoelectric converter 1, noise removed by a low-pass filter 2, and then converted into a digital signal by an analog/digital converter 3. After this digital signal is filtered by an online two-dimensional filter using a shift register, it is input to the frame memory 6.

Next, the structure of the online two-dimensional filter 7 will be specifically explained using FIG. 4. 8a and 8b are shift registers capable of storing pixel data for one scanning line obtained by line scanning, and constitute the first shift register of the present invention. That is, in this example, a shift register is used that can store pixel data for a total of two scanning lines.

8c is a shift register capable of storing data for three pixels, and corresponds to the second shift register of the present invention.

9a to 9d are adders, and the shift register 8a
, 8b, the data for 3 pixels at the beginning of each scanning line and the data for 3 pixels stored in the shift register 8c, a total of 3×3=9 pixels are added. .

10a to 10i are multipliers for weighting the data for the nine pixels, and 11 is a divider for averaging the added data.

Next, the operation of this online two-dimensional filter 7 will be explained.

Digital signals obtained by the analog/digital converter 3 are sequentially input to a shift register 8a, an output signal of the shift register 8a is input to a shift register 8b, and an output signal of the shift register 8b is input to a shift register 8c. Here, the first three of shift registers 8a, 8b, 8c
The pixel data is sent to multipliers 10a to 10c, 10d, respectively.
~10f, 10g~10i, adder 9a
~9c is added. Further, the outputs of the adders 9a to 9c are added by an adder 9d, and a total of nine pixel data are weighted and added. This is divided by a divider 10, the average value is calculated, and the result is input into the frame memory 6.

As explained in the conventional example, when weighting each pixel equally, the multipliers 10a to 10i for weighting are unnecessary.

Furthermore, when inputting a document with characters or figures drawn on it, each pixel can be expressed as a binary value (that is, 1 bit). Therefore, the capacity of each of the shift registers 8a to 8C only needs to be 1 bit, and for the result of adding up the sum of 9 pixels, instead of using a divider, for example, if the sum is 0 to 4, it is "0", and if the sum is 5, ~
In the case of 9, it is also possible to use a comparator that outputs "1".

According to the present invention, the filtered output of pixel (i, j) is obtained at the time when pixel (i+1, j+1) is input, and a delay of one scanning line+one pixel always occurs. However, the total time from data input to the end of filter processing increases only by the time equivalent to approximately one scanning line of data input compared to the data input time, so this does not pose a problem.

FIG. 5 shows another embodiment in which a two-dimensional filter 7a having the same function as the two-dimensional filter 7 is connected in series in the embodiment shown in FIG. 4 to further improve the filtering effect. In this case, the filtered data of pixel (i, j) is obtained at the time when pixel (i+2, j+2) is input, but as described above, this does not pose a problem in terms of processing time.

In the above explanation, we have described the case where noise is removed two-dimensionally by taking a simple average of 3 x 3 pixels, but it is also effective in other cases such as 2 x 2 pixels, or 5 x 5 pixels, etc. be. It is also obvious that various filter effects can be obtained by weighting each pixel.

Furthermore, it is clear that the present invention can be applied even when each pixel is expressed by multiple bits of data. [Effects of the Invention] According to the present invention, online 2
The dimensional filter allows filter processing to be performed without much delay compared to when data is input. Therefore, the image processing time can be shortened without increasing the memory capacity compared to the conventional method.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an example of a conventional image processing device, FIG. 2 is an explanatory diagram for explaining filter processing in the conventional image processing device, and FIG. 3 is a block diagram showing an example of the present invention. 4 is a diagram showing an example of the structure of an online two-dimensional filter used in the image processing apparatus of the present invention, and FIG. 5 is a block diagram showing another embodiment of the present invention. 1...Photoelectric converter, 2...Low pass filter, 3...Analog/digital converter, 4, 6
...Frame memory, 5...Two-dimensional filter, 7,7a...Online two-dimensional filter, 8a~
8c...Shift register, 9a-9d...
- Adder, 10a to 10i... Multiplier, 11... Divider.

Claims (1)

[Claims] In an image processing device that optically reads characters and figures by performing line scanning using a photoelectric converter and further processes the image electronically, the image of the characters and figures is converted into a digital signal representing the density of each pixel. a means of converting it into
a first shift register into which the digital signals are sequentially input and stored for n-1 (n is an integer of 2 or more) scanning lines; and a second shift register into which digital signals for n pixels are input and stored.
a shift register; means for adding data for n pixels corresponding to the beginning of each scanning line in the first shift register to data for n pixels in the second shift register; and averaging the added data. An image processing device comprising means for.