JPH0594531A - Binarizing filter device - Google Patents

Abstract

(57) [Summary] [Objective] To obtain a binarization filter that removes noise components of image data with a simple configuration. [Structure] The input image data is passed through a filter having a hysteresis characteristic, and the output is binarized to obtain an output result in which both fixed noise components and random noise components are removed from minute noise components. ..

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention eliminates noise components contained in image data such as characters and images input from the outside by using a filter having a hysteresis characteristic. The present invention relates to a digital filter device.

[0002]

2. Description of the Related Art Conventionally, a method shown in FIG. 4 has been used as a method for removing noise components contained in input image data and binarizing them. This is because the fixed noise component 2 that appears at a specific position in the input image data 1 that is input due to the characteristics of the input device is stored in advance in the fixed noise component storage memory 3 provided in the signal processing system, and is input by the difference circuit 8. The input image data 1 containing the fixed noise component 2 and the stored reference fixed noise component 4 are obtained to obtain the output image data 5 not containing the fixed noise component 2 and binarize the output image data 5. is there.

In this apparatus, it is necessary to register the reference fixed noise component 4 in the fixed noise component storage memory 3 before use. This is to input an object to the image data capturing device 6 whose original output is exactly 0,
The input image data 1 obtained at this time is registered in the fixed noise component storage memory 3 as the reference fixed noise component 4. The reference fixed noise component 4 is the input image data 1 which is the output signal of the image data capturing device 6.
It is registered in the fixed noise component storage memory 3 while taking into consideration the position in which it appears.

The noise removing and binarizing operations are performed in the following procedure. The input image data 1 is input together with position information 7 indicating which position of the image data capturing device 6 the position corresponds to. The input position information 7 is given to the fixed noise component storage memory 3, and the fixed noise component storage memory 3 outputs the reference fixed noise component 4 of the corresponding position stored by the above operation to the difference circuit 8. The input image data 1 is also input to the difference circuit 8, and the difference circuit 8 takes the difference between the input image data 1 and the reference fixed noise component 4, and the resulting output image data 5 is binarized circuit 9 To enter. The binarization circuit 9 binarizes the given signal and obtains a final result 10.

Next, the operation of the prior art will be described with reference to FIG. For simplification of description, the image data capturing device 6 is a line sensor, and the length of the line sensor corresponds to eight data,
The output value is 0 to 7, and the fixed noise component 2 has a magnitude of 2 and appears as 0 0 +2 0 0 -2 0 0 in the operation direction. When an object whose output is originally 2 2 2 2 5 5 5 5 is input to the image data capturing device 6, a fixed noise component 2 is added to the input image data 1 and becomes 2 2 4 2 5 3 5 5. Takes a value.

This input image data is input to the difference circuit 8, and at the same time, the fixed noise component storage memory 3
The reference fixed noise data 4 of 0 0 +2 0 0 -2 0 0 stored in advance is also input to the difference circuit 8,
The difference circuit 8 takes the difference between the input image data 1 and the reference fixed noise component 4 to obtain output image data 5 of 2 2 2 2 5 5 5 5 which is binarized.
To obtain the final result of 10. If the input range of the truncation circuit is 0 to 7 and the output is 1 bit, the final result 10 is 0 0 0 0 1 1 1 1 when the fixed noise component 2 is removed, and 0 0 when the fixed noise component 2 is not removed. It becomes 1 0 1 0 1 1.

FIG. 6 shows the case where the following random noise is added to the above fixed noise component. 0 +2 0 0 0 0 0 -2 In this case, it cannot be removed only by taking the difference with the fixed noise component 2 stored in the fixed noise component storage memory 3, and the following noise is added to the output image data 5. Occurs. 2 (4) 2 2 5 5 5 (3) Also, the following noise is generated in the final result 10. 0 (1) 0 0 1 1 1 (0) The data in parentheses indicates data affected by random noise.

[0008]

The above-mentioned conventional apparatus has the following problems. The first point is that the fixed noise component storage memory 3 is required. The fixed noise component storage memory 3 needs to have the same size as the processing unit of the image data capturing device 6, and if the image data capturing device 6 is a line sensor, it has the same size as the number of data in the length direction. A storage area is required, and an area sensor requires a storage area having the same size as the number of data contained therein.

The second point is that it is necessary to accurately register the reference fixed noise component 4. If random noise 16 randomly generated as shown in FIG. 6 is mixed in while registering the reference fixed noise component 4, the reference fixed noise component 4 is generated.
If is mistakenly registered, the operation of obtaining the difference between the input image data 1 and the reference fixed noise component 4 will increase the noise component, which causes the output image data 5 to deteriorate.

The third point is that it has no effect on noise components other than the noise registered as the reference fixed noise component 4 as shown in FIG. When random noise 16 is added to the input image data 1, this noise is not stored in the fixed noise component storage memory 3,
Even if you take the difference, it will remain as it is.

[0011]

In order to solve the above problems, the present invention provides a binarizing filter device for binarizing image data such as characters and images inputted from the outside, by converting the input image data into On the other hand, a filter having a hysteresis characteristic for removing a noise component contained in the image data is provided to eliminate the influence of a minute noise component at the time of binarization.

Also, a latch for delaying the input image data containing a noise component by one processing time, and a size comparator for comparing the size of the input image data delayed by one processing time from the latch with the original input image data, The comparison data compared by the size comparator and a decoder for deriving a binary signal from the original image data are provided.

[0013]

According to the above configuration, the fixed noise component 2 can be removed from the minute noise without using the fixed noise component storage memory 3, and the random noise 16 can be removed. ..

[0014]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described in detail below with reference to the accompanying drawings. FIG. 1 is a block diagram of an embodiment of the present invention. This embodiment comprises an image data capturing device 6, a data holding latch 11, a size comparator 12 and a decoder 13. The input image data 1 output from the image data capturing device 6 is input to the latch 11, the size comparator 12 and the decoder 13. The previous input image data 14, which is the output of the latch 11, holds the input image data 1 immediately before the current input image data 1, and is obtained by delaying the input image data 1 by one process. This is also input to the magnitude comparator 12.

The comparison result 15 which is the output of the magnitude comparator 12
Is the input image data 1 and the previous input image data 1
4 is the result of comparing the magnitudes of 4 and 10 if the input image data 1 is larger, 01 if the input image data 1 is smaller, and 11 if the values of the input image data 1 and the previous input image data 14 are equal. Become. Comparison result 1
5 is input to the decoder 13.

The decoder 13 binarizes the input image data having a hysteresis characteristic based on the input image data 1 and the comparison result 15, and the final result 10 is obtained.
Output as.

The operation of FIG. 1 will be described below with reference to FIG. For simplification of description, the image data capturing device 6
Is a line sensor, the length of which is 8 pieces of data, and the output values are 0 to 7. The image data capturing device 6 captures the target and obtains the input image data 1. At this time, the input image data 1 originally obtained is 2 2 2 2 5 5 5 5, the fixed noise component 2 is 0 0 +2 0 0 -2 0 0, and the random noise 16 is 0 +2 0 0 0 0. If it is 0 0 -2, the value of the input image data 1 will be 2 4 4 2 5 3 5 3.

The previous input image data 14 which is the output of the latch 11 is obtained by delaying the input image data 1 by one process, and therefore becomes * 2 4 4 2 5 3 5. When the input image data 1 and the previous input image data 14 are compared by the size comparator 12, the comparison result 15 has the following value. * 10 11 01 10 01 10 01 The characteristics of the decoder are as shown in Fig. 3. In this example, since the first term of the input image data 1 is 2, the comparison result 15
The final result 10 is 0 regardless of the value of. In the second term, the input image data 1 is 4, and the comparison result is 10, so the final result 10 is 0. In the same manner, the final result 10 is 0 0 0 0 1 1 1 1 1. However, when simple binarization is performed without the hysteresis characteristic, the final result 10 is 0 0 1 1 0 1 0 1 0.

[0019]

Since the present invention is constructed as described above,
For minute noise, binarization can be performed by removing both fixed noise and randomly generated noise.

[Brief description of drawings]

FIG. 1 is a configuration diagram of an embodiment of the present invention.

FIG. 2 is an operation explanatory diagram of FIG.

FIG. 3 is a characteristic diagram of a decoder used in the present invention.

FIG. 4 is a configuration diagram of a conventional example.

FIG. 5 is an operation explanatory diagram of FIG. 4;

FIG. 6 is an operation explanatory diagram of FIG. 4;

[Explanation of symbols]

 11 Latch 12 Large / small comparator 13 Decoder

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Takeshi Yanome 22-22 Nagaike-cho, Abeno-ku, Osaka City Inside Sharp Corporation

Claims (2)

[Claims] 1. A binarizing filter device for binarizing image data such as characters and images inputted from the outside, wherein a noise component contained in the image data is removed from the inputted image data. A binarization filter device characterized in that a filter having a hysteresis characteristic is provided so as to eliminate the influence of a minute noise component at the time of binarization. 2. A binarizing filter device for binarizing image data such as characters and images inputted from the outside, a latch for delaying input image data containing a noise component by one processing time, and one of the latches. A size comparator for comparing the size of the input image data delayed by the processing time with the size of the original input image data; and a decoder for deriving a binary signal from the comparison data compared by the size comparator and the original image data. A binarizing filter device provided.