JPH066597A - Noise elimination device - Google Patents

Abstract

PURPOSE:To detect and eliminate plural designated noise patterns by outputting an opposite output on a correspondent noise detection circuit so as to invert a black pixel while being read from a scanner at present when at least one coincident output is outputted among the noise detection circuits 1-3. CONSTITUTION:Three line registers of noise detection circuits 1-3 always revise binary serial data read by a scanner in the unit of lines to latch up-to-date picture data. A comparator compares a noise pattern in an NXN matrix to be eliminated and set in a pattern register with a pattern in each picture data matrix. When any pattern is coincident with any of the patterns, data in a relevant line register to a black picture element position of the pattern register are inverted (that is, noise elimination) and output data 1-3 are ANDed to obtain output serial data. Thus, plural noise patterns of an optional shape designated from read binary serial data are eliminated in real time.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a noise eliminating device for eliminating noise in an image read by an image reading device such as a scanner device.

[0002]

2. Description of the Related Art In recent years, with the spread of tools such as workstations, DTPs, personal computers, and the like, OA is advancing, and a scanner device or the like is used as an image data input means for OA equipment, which is noise-free and of high quality. The data is desired. However, conventionally, when a low-quality original is used (such as repeated copies), or when the image data other than the image data to be read (that is, the background of the original) is composed of multiple colors, the image quality is significantly reduced due to the effect of noise. Was there.

[0003]

However, in the conventional method, when reading binary image data with an image reading device such as a scanner, a low quality original is used, or the image data other than the read target image data is composed of a plurality of colors. In this case, the image quality was significantly reduced due to the influence of noise.

[0004]

In order to solve this problem, the present invention solves this problem by a group of line registers for holding binary serial data read from an image reading device such as a scanner on a line-by-line basis and N × N to be removed. A pattern register group that stores a plurality of noise patterns composed of a matrix (N is a natural number), a comparator group that compares the N × N dot area of the line register group with each pattern register of the pattern register group, and a comparator group For all the pattern registers that have output the coincidence output, the line register N × corresponding to the position of the black pixel in the pattern register
By having a data inversion circuit that inverts black pixels in the N dot area, it is possible to remove a plurality of noise patterns set in the pattern register group of an arbitrary shape from the read data in real time.

[0005]

According to the present invention, a plurality of two-dimensional noise patterns of arbitrary shapes can be simultaneously removed in real time from binary serial data read by a scanner or the like that scans a document line by line and reads image data. Is.

[0006]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 is an overall block diagram of a noise eliminator according to an embodiment of the present invention.

FIG. 2 is a detailed block diagram of the noise detection circuit 1 of the overall block diagram of FIG. The noise detection circuits 2 and 3 have the same circuit configuration as the noise detection circuit 1.

In this embodiment, as shown in the left diagram of FIG.
The pattern of the removal noise is composed of a 3 × 3 matrix,
The bit data names corresponding to each bit position are K, L, M, N, O, P, R, S, and T, respectively.

The binary serial data read from the scanner device is sequentially read downward line by line in the direction indicated by a from the upper left point of the document as shown in FIG. Further, as shown in FIG. 3, the portion of the read image data subjected to noise removal processing is a 3 × 3 matrix (A, B,
C, D, E, F, G, H, J) image data matrix. There are three removal noise patterns.

In FIG. 1, 1 is a noise detection circuit for detecting and removing a first noise pattern from binary serial data read from a scanner. Similarly, noise detection circuits 1, 2 and 3 detect second and third noise patterns. Each of the noise detection circuits 2 and 3 to be removed, and when any one of the noise detection circuits 1 to 3 outputs a coincident output, an inverted output is output to the corresponding noise detection circuit, and the current read image It is a data inversion circuit that inverts black pixels (that is, noise) of a data matrix (3 × 3) to remove noise.

The operation of this embodiment shown in FIG. 1 will be described. First, binary serial data obtained by reading a document in units of one line by a scanner device or the like is simultaneously input to the noise detection circuits 1 to 3, respectively.

Each noise detection circuit has a line register of 3 lines, and every time 1 dot of new line data is input, the content of the oldest line register is updated by 1 dot, and the 3 lines of the latest continuous document are always displayed. Holds image data (only 3 lines capture serial data from the original start point).

Here, as shown in FIG. 3, the three lines b, c, and d are held in the noise detection circuits 1 to 3, and A to B in FIG.
Image data matrix of J and pattern register 1 of FIG.
The noise patterns set to ~ 3 are compared and one of them is
If any one of them matches, the data of the image data matrix of FIG. 3 at the black pixel position (K to T) of the corresponding pattern register is inverted (that is, noise is removed) and output as noise-removed serial data. In FIG. 1, output data 1 to 3 are outputs from which noise patterns individually set by the noise removing circuits 1 to 3 have been removed. By ANDing the output data 1 to 3, the noise removing circuits 1 to 3 are output.
It is possible to obtain the output serial data from which all the noise patterns set by are removed.

Now, the noise detection circuit 1 of FIG. 1 will be described. 2 shows the noise detection circuit 1. In FIG. 2, 1 is a first line shift register that holds serial data input from the scanner in line units and shifts the serial data to the next line shift register. 1 (stores the most recently read line data), 2 holds the serial data input from the line shift register 1 on a line-by-line basis, and shifts the serial data to the line shift register of the next stage Line shift register 2 (line shift register 1
Hold the line data one line before), 3 is a third line shift register 3 that holds the serial data input from the line shift register 2 in line units and shifts the serial data to the next line shift register. (Holds line data one line before in the line shift register 2) 4, latch for holding shift data output from the final stage of the line shift register 3, 5 stores noise pattern to be removed from the read image data A pattern register 6 is a line shift register 1, 2, 3 for holding the read continuous 3 lines of image data.
Nine comparators (comparators A to A) for comparing each bit data of the 3 × 3 data matrix (A to J bits in the line registers 1 to 3) composed of the data in It is a comparator group consisting of comparators J).

First, in FIG. 2, binary serial data read from a scanner device or the like is input to the line shift register 1 and is shifted one after another every time the next serial data arrives, and the final stage data is the line shift register. Two
Is output to. Similarly, the output data of the line shift register 2 is input to the line shift register 3 and is shifted one after another each time the next serial data arrives at each input end.

The line shift register 1 holds the previous line currently being read by the scanner, and replaces the line data currently being read by the scanner by shifting the serial line data in 1 dot units. The line shift register 2 performs the same operation for the line data one line before the line shift register 1. That is, the line shift register 2 holds the data of two lines before, which is currently being read by the scanner, and replaces the previous line data by a shift operation in units of one dot of serial line data. Similarly, the line register 3 replaces the data three lines before read by the scanner with the data two lines before.

The case where the scanner is actually scanning the original will be described. As shown in FIG. 3, the scanner sequentially reads downward line by line in the direction indicated by a from the upper left point of the original, and is composed of three consecutive lines b, c, and d from the read image data. × 3 matrix (A,
B, C, D, E, F, G, H, J) image data matrices are subjected to noise removal processing. If it is assumed that the process is currently being performed at the positions of the image data matrix shown by A to J in FIG. 2, the line data of the line shift register 1 is replaced with the line data of the next line f. The line shift register 2 has line data c replaced by line data d, and the line shift register 1 has line data b replaced by line data c. That is, each bit data of the processed image data matrix shown by A to J in FIG. 3 corresponds to A to J in the line shift registers 1 to 3 in FIG. When the A to J bits in the line shift registers 1 to 3 and the corresponding bits of the K to T bits in the pattern register are compared in the comparator group and all the comparators A to J in the comparator group output a match,
The image data matrix being processed is determined to be noise, the coincidence output 1 is output, the inverted output 1 of the data inversion circuit 4 in FIG. The data inversion circuit inverts to white pixels to remove noise. The serial binary data from which noise has been removed is sequentially output from the latch. As shown in FIG. 4, if a removal noise pattern is set in the pattern register,
The D and F data of the image data matrix are inverted to become white pixels, and noise is removed. Next, the image data matrix of FIG. 3 shifts to the right by one dot and the next image data matrix process is performed. After that, when the processing is completed up to the right end of the original, the same processing as above is performed on the image data of all the originals from the left end of the original one dot below.

The noise detection circuits 2 and 3 shown in FIG. 1 also perform noise removal on the noise pattern that is similarly set.

In this embodiment, the image data matrix 3 ×
Although 3 dots are used, if the number of line shift registers is N (N is a natural number) and the pattern register is N × N dots, the size of the removal noise can be freely expanded.

Needless to say, it is possible to increase the number of removable noise patterns by increasing the number of noise detection circuits shown in FIG.

[0021]

As described above, a group of line registers for holding binary serial data read from an image reading device such as a scanner in units of lines and an N × N matrix to be removed (N is an arbitrary number) Pattern register group that stores a plurality of noise patterns composed of natural numbers), a comparator group that compares the N × N dot area of the line register group with each pattern register of the pattern register group, and a coincident output in the comparator group With respect to all the pattern registers that output the above, by having a data inversion circuit that inverts the black pixel in the N × N dot area of the line register corresponding to the position of the black pixel in the pattern register,
It is possible to realize a noise eliminator that can remove a plurality of noise patterns set by a pattern register group of an arbitrary shape from read data in real time.

[Brief description of drawings]

FIG. 1 is an overall block diagram of a noise eliminator according to an embodiment of the present invention.

FIG. 2 is a block diagram of a noise detection circuit 1 in FIG.

FIG. 3 is an explanatory diagram of an image data matrix on a document from which noise is removed according to this embodiment.

FIG. 4 is a diagram showing an example of setting a noise pattern for removal.

[Explanation of symbols]

 1 Noise Detection Circuit 1 2 Noise Detection Circuit 2 3 Noise Detection Circuit 3 4 Data Inversion Circuit 5 Line Register 1 6 Line Register 2 7 Line Register 3 8 Latch 9 Pattern Register 10 Comparator Group

Claims (1)

[Claims] 1. A scanner device for scanning a document in line units to read image data, and a line register group for holding read binary serial data in line units and an N × N matrix to be removed (N is a natural number). ), A pattern register group for storing a plurality of noise patterns, a comparator group for comparing the N × N dot area of the line register group with each pattern register of the pattern register group, and A noise eliminator having a data inversion circuit for inverting black pixels in the N × N dot area of the line register corresponding to the positions of black pixels in the pattern register for all the pattern registers that have output coincidence. apparatus.