JPH0766998A - Device and method for noise elimination - Google Patents

Abstract

PURPOSE:To eliminate only isolated noise representing strong impulse property by predicting a threshold level based on preceding prediction data corresponding to past input data so as to compare a threshold level and a difference between current input data and predicted data. CONSTITUTION:A picture signal fed to an input terminal 11 is block-processed by a block processing unit 3 and fed to a coefficient computing element 13 via a signal value prediction device 4 and a data memory 9. The signal prediction device 4 and the coefficient computing element 13 obtain prediction data with respect to current input data based on past input data and a threshold level prediction device 5 is used to predict a threshold level based on the past input data and past prediction data. Then a comparator 7 compares a difference between the current input data and the present prediction data with a threshold level and a selection device 6 selects either the current input data or corresponding prediction data based on the result of comparison. Thus, only isolated noise with strong impulse property is eliminated.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a noise removing apparatus and a noise removing method for removing noise from a signal in which noise is mixed, and particularly to removing noise from a digital signal containing noise having strong impulse characteristics.

[0002]

2. Description of the Related Art Conventionally, as a noise removing method for removing noise from a digital signal in which noise is mixed, for example, a moving average method and a median method are known.

In the above-described moving average noise removal method, the average value of signals (hereinafter referred to as neighboring signals) that are temporally or spatially close to the signal to be processed (the signal from which noise is desired to be removed) is calculated as follows. It is like performing an operation of replacing with a signal related to processing.

The noise removal method based on the median is such that the median value (median) when the neighboring signals are arranged in the order of ascending or descending and the signal related to the processing is replaced. is there.

[0005]

By the way, for example, noise (error) that occurs during recording / reproduction of a recording / reproducing apparatus that handles digital signals has a strong impulse property (random noise), and this noise signal is isolated. It has the following properties. The term "isolated" means that there is no other noise signal in the neighboring signal.

However, when noise removal processing using the above-described conventional noise removal method is performed on such isolated noise having a strong randomness, most signals that are not affected by noise are also processed. Will be processed. That is, when an image signal is taken as an example, if a signal not affected by noise is processed in the same manner as described above, as a result, image quality deterioration such as image blur occurs.

In view of the above-mentioned circumstances, the present invention has an object of providing a noise removing apparatus and method capable of removing only isolated noise having strong impulse characteristics.

[0008]

The present invention has been proposed in order to achieve the above-mentioned object, and the noise removing apparatus of the present invention makes a prediction based on past input data and makes a present input. The first to calculate the prediction data corresponding to the data
Predicting means, second predicting means for predicting a threshold value based on past input data and past prediction data corresponding to the past input data, and current input data and the current input data. The difference from the corresponding predicted data is the second
Comparing means for comparing with the threshold value from the predicting means, and selecting means for selecting either the current input data or the prediction data corresponding to the current input data based on the comparison result of the comparing means. I have.

Here, the first predicting means uses a coefficient calculator for adaptively determining a prediction coefficient by applying a least squares operation to the past input data, the current input data and the prediction coefficient. And a signal value prediction unit that performs linear prediction.

Further, the noise removing apparatus of the present invention further comprises first storage means for storing input data, and the first storage means.
Second storage means for storing the prediction data from the prediction means is provided, and the second prediction means determines the difference between the data from the first storage means and the prediction data from the second storage means. Based on this, the threshold is adaptively changed.

It is preferable that the input data is divided into blocks for each of a plurality of lines and a plurality of pixels, and the processing such as the above prediction is performed using the data in block units.

Next, the noise removing method of the present invention obtains prediction data corresponding to the present input data based on the past input data, and calculates past input data and past prediction data corresponding to the past input data. Predict the threshold value based on, comparing the difference between the current input data and the predicted data corresponding to the current input data with the threshold value, based on the comparison result the current input data or Any of the prediction data corresponding to the current input data is selected.

Here, the prediction data is obtained by subjecting the past input data to a least-squares operation to adaptively determine a prediction coefficient and performing prediction using the current input data and the prediction coefficient. I am trying.

Further, the threshold value is set so that the input data and the prediction data are stored respectively, and the threshold value is adaptively changed based on the difference between the stored input data and the prediction data.

In other words, the noise removing device of the present invention is a device for removing noise of a digital signal,
A signal estimating means for estimating a signal to be processed, a threshold estimating means for estimating a reference of the likelihood of the output of the signal estimating means, and a signal estimating means based on the outputs of the signal estimating means and the threshold estimating means. And a switching means for switching the input signal relating to the processing.

Here, a memory for accumulating past input signals and outputs of past signal estimating means is provided, and the signal estimating means and the threshold estimating means use the accumulated past signals for estimation. I'm trying to do.

The state of the memory is controlled by the output of the switching means.

Further, the signal estimating means is a least-squares prediction coefficient determining means and a linear predicting means, and the threshold estimating means is a difference between a past input signal value and a corresponding past accumulated predicted value. A value proportional to the sum of absolute values of is set as the threshold value.

[0019]

According to the present invention, the threshold value is adaptively predicted based on the past input data and the corresponding past prediction data, and the threshold value and the difference between the current input data and the prediction data are calculated. To compare. At this time, if the input data includes noise having strong impulsiveness, for example, the value of the difference between the current input data and the prediction data is different from the threshold value. In this case, by outputting the prediction data instead of the input data, the obtained data becomes noise-free data (noise-free data). Further, by removing only the noise having a strong impulse property, it is possible to prevent the processing of a signal that does not include noise (low noise).

[0020]

Embodiments of the present invention will be described below with reference to the drawings.

As shown in FIG. 1, the noise removing apparatus of the embodiment to which the noise removing method of the present invention is applied, is a blocking processor 3 for blocking input data supplied through a terminal 11, and A data memory 9 (first storage means) for storing the block data from the block processor 3
And a coefficient calculator 13 for adaptively determining a prediction coefficient by performing a least squares operation on the past block data from the data memory 9, and the prediction coefficient for the current block data from the blocking processor 3 Signal value predictor 4 for multiplication
And (the first predicting means includes the coefficient calculator 13 and the signal value predictor 4), the data memory 10 (second storing means) for storing the predictive data from the signal value predictor 4,
A threshold predictor 5 (a first predictor) that adaptively predicts a threshold value based on the block data from the data memory 9 and the prediction data stored in the data memory 10 corresponding to the block data from the data memory 9 2 prediction method)
And a comparator 7 that compares the difference between the input data and the prediction data of the signal value predictor 4 with the threshold value from the threshold value predictor 5, and based on the comparison result of the comparator 7, And a selector 6 for selecting either the input data or the prediction data (for example, when the difference> threshold value, the prediction data is selected, and when the difference ≦ threshold value, the input data is selected).

The threshold predictor 5 adaptively changes the threshold based on the difference between the block data from the data memory 9 and the prediction data from the data memory 10. .

In this embodiment, assuming that a digital image signal is used as input data, noise is filtered by the digital VTR.
The case where noise is caused by a random bit error during reproduction of (a VTR that does not compress an image signal) will be described.

In FIG. 1, the input terminal 11 is supplied with an operation-converted one-dimensional digital image signal which is reproduced from a recording medium (magnetic tape) by a digital VTR and subjected to a predetermined process by a not-shown preceding structure. To be done. Therefore, the digital image signal supplied to the input terminal 11 may include noise due to a random bit error when the VTR is reproduced.

The digital image signal supplied to the input terminal 11 is sent to the blocking processor 3. The blocking processor 3 forms, for example, a two-dimensional block of 8 lines × 8 pixels, and performs order conversion so that the image signal in the two-dimensional block becomes a temporally continuous one-dimensional signal. The size of block formation can be arbitrarily selected, and for example, can be set to 3 × 3 pixels.

The output data si3 from the blocking processor 3 is a signal value predictor 4 and a data memory 9 which will be described later.
Sent to. As will be described later, the data memory 9 controls whether or not to store the data si3 from the blocking processing circuit 3 according to the comparison result signal si7 from the comparator 7. .

The data si9 read from the data memory 9 is sent to the coefficient calculator 13. The coefficient calculator 13 uses the data si9 read from the data memory 9 (ie, the data one block before the current output data si3 from the blocking processor 3), for example, using the least squares method. To calculate the prediction coefficient.

In other words, the above-mentioned prediction coefficient is obtained from data that has been processed earlier in time and is judged not to be noise, and the one stored in the memory 9 is used for the least squares method. It was decided by.

The prediction coefficient from the coefficient calculator 13 is sent to the signal value predictor 4.

The signal value predictor 4 is also supplied with the current output data si3 from the block processor 3 together with the prediction coefficient obtained by the coefficient calculator 13. Then, the calculation for multiplying the data si3 by the prediction coefficient is performed. That is, in the signal value predictor 4, the output data si3 from the blocking processor 3 for the current processing is stored in the data memory 9
The current prediction data (signal estimated value) si4 is obtained by multiplying the prediction coefficient obtained from the past data si3 stored in
To get

Here, using the flowchart of FIG.
The flow of processing in the coefficient calculator 13 and the signal value predictor 4 will be described.

In FIG. 2, in step S1, the past output data si3 of the blocking processor 3 stored in the data memory 9 is made into the form of simultaneous linear equations. In the next step S2, the predictive coefficient (coefficient solution) is obtained by solving the simultaneous linear equations. The processing of steps S1 and S2 is performed by the coefficient calculator 13.

In step S3, the current output data si3 from the blocking processor 3 and the coefficient solution (prediction coefficient)
Multiply with. The processing in step S3 is performed by the signal value predictor 4.

The prediction data si4 from the signal value predictor 4 obtained as described above is sent to the comparator 7 and also to the data memory 10.

The data si10 (predicted data) read from the data memory 10 is sent to the threshold predictor 5. Further, the output data si9 from the data memory 9 is also supplied to the threshold predictor 5. However, in this case, the data si10 read from the data memory 10 is data corresponding to the output data si9 from the data memory 9.

In the threshold predictor 5, the output data si9 (data processed earlier in time) from the data memory 9 is used for all the data stored in the data memories 9 and 10. , Subtraction is performed with the data si10 from the data memory 10 corresponding to the data si9, the absolute value of the subtraction value is summed, and the total value is divided, and then the value is predetermined. Multiplied by a predetermined proportional constant, the obtained value is used as a threshold value s
It is output as i5.

That is, the processing in the threshold predictor 5 is as shown in the flowchart of FIG.

In FIG. 3, in step S11, from the output data si3 (si9) of the blocking processor 3 stored in the data memory 9, the output data si4 (si10 of the signal value predictor 4 stored in the data memory 10). ) Is subtracted.

At the next step S12, the absolute value of the subtraction value is taken, and at step S13, all of them (for all data stored in the data memories 9 and 10).
Calculate the data sum of.

In step S14, the total value is divided by the total number of data, and in step S15, the divided value is multiplied by a predetermined proportional constant. The above-mentioned proportional constant is appropriately determined in advance and does not change during the process. A value larger than 1.0 such as 1.7, 2.0, ... Is used as the predetermined constant of proportionality.

The threshold value si5 obtained by the threshold value predictor 5 as described above and the prediction data si4 from the signal value predictor 4 described above are sent to the comparator 7.

In addition, the input terminal 1 is connected to the comparator 7.
An input digital image signal from 1 is also provided. In the comparator 7, the prediction data si from the signal value predictor 4
4 and the image signal (input signal value) from the input terminal 11 are taken, and the absolute value of the difference is compared using the threshold value from the threshold value predictor 5 (determined adaptively. Comparison based on the threshold value).

When the comparator 7 judges that the absolute value of the difference is larger than the threshold value in the comparison, the signal from the input terminal 11 contains noise (impulsive noise). On the contrary, when it is determined that the absolute value of the difference is not larger than the threshold value, comparison result data si7 indicating that the signal from the input terminal 11 does not include noise is output. The comparison result data si7 is, for example, 1 when it indicates that noise is included, and 0 when it indicates that noise is not included. Of course, it is 0 to indicate that noise is included,
It can be set to 1 to indicate that noise is not included.

This data si7 is sent to the selector 6 as a selection control signal. The selector 6 selects the prediction data si4 from the signal value predictor 4 and the input data from the input terminal 11 according to the selection control signal, and the selection control signal is If it is a signal indicating that noise is included, the prediction data si4 from the signal value predictor 4 is used. Conversely, if the selection control signal is a signal indicating that noise is not included, the input terminal 1 is used.
Select the input data from 1.

Therefore, the signal selected by the selector 6 becomes image data that does not contain noise (that is, noise is removed), and this noise-free data is taken out from the terminal 12.

The data si7 from the comparator 7 is also used.
Is also sent to the data memory 9 as a control signal for storing the data si3 from the blocking processor 3. That is, when the data si7 indicates that it includes noise, the output data si3 from the blocking processor 3 is not stored in the data memory 9, so that the data si3 (data si3) including the noise is stored.
9) is used to prevent the calculation processing in the coefficient calculator 13 from being performed. As a result, the signal value predictor 4
In, it is possible to prevent the prediction calculation using the prediction coefficient based on the signal including noise from being performed. The data si7 can also be considered as a control signal for determining whether to read the data from the data memory 9.

The coefficient calculator 13 and the threshold predictor 5 of the noise eliminator of this embodiment hold a predetermined prediction coefficient and threshold as initial values. Therefore, at the initial stage, for example, when the power is turned on, the signal value predictor 4 performs the prediction by the prediction coefficient that is predetermined and held in the coefficient calculator 13, and the threshold value prediction is performed. A comparison (determination as to whether or not there is noise) in the comparator 7 is made according to a threshold value that is predetermined and held in the device 5.
Then, as described above, the data memory 10 for the prediction data
Is stored and the next block data is processed.

That is, in the noise eliminator of the present embodiment, by performing the processing as described above, relatively large conspicuous noise is eliminated, and at the same time, the input signal and its prediction data (estimated signal) are The data is gradually stored in the data memories 9 and 10. In this way, after a certain amount of data is stored in the data memories 9 and 10, the prediction coefficient corresponding to the input data to be currently processed is It is determined using the least squares method from the past data. The prediction data predicted by the prediction coefficient determined by the least squares method and the input data are the threshold value judgment with the absolute value of the difference between the past prediction data (estimated value) and the corresponding input data, that is, the value proportional to the error. Then, it is determined whether or not it is noise. By repeating such an operation, the signal to be gradually removed approaches true noise (noise having strong impulsiveness).

In other words, the noise removing apparatus according to the embodiment of the present invention estimates the time or position of noise in advance, adaptively determines the presence or absence of noise according to the nature of the noise, and selects the presence or absence of noise. By performing the processing, a signal without noise is obtained. That is, in the noise removal apparatus of the present embodiment, the signal value that should be present is estimated from the past property of the signal value, and the difference between this estimated value and the actual signal value exceeds the variation that can be considered from the property of the signal. By correcting only the signal, only the noise (noise having strong impulsiveness) extremely deviating from the property of the signal is corrected, and the careless deterioration of the signal can be suppressed.

[0050]

As described above, in the present invention, the prediction data corresponding to the current input data is obtained based on the past input data, and the past input data and the past prediction corresponding to the past input data are obtained. Predict the threshold value based on the data, and compare the difference between the current input data and the predicted data corresponding to the current input data with this threshold value,
By selecting either the current input data or the prediction data corresponding to the current input data based on the comparison result, for example, it is possible to remove only isolated noise having strong impulse characteristics. .

[Brief description of drawings]

FIG. 1 is a block circuit diagram showing a schematic configuration of a noise removing device according to an embodiment of the present invention.

FIG. 2 is a flowchart showing a process of obtaining prediction data in this embodiment.

FIG. 3 is a flowchart showing a process of predicting a threshold value according to the present embodiment.

[Explanation of symbols]

 3 ... Blocking processor 4 ... Signal value predictor 5 ... Threshold predictor 6 ... Selector 7 ... Comparator 9, 10 ... Data memory 13 ...- Coefficient calculator

Claims (6)

[Claims] 1. A first prediction means for performing prediction based on past input data to calculate prediction data corresponding to current input data; and past input data and past past data corresponding to the past input data. A second predicting means for predicting a threshold value based on the predictive data, and a difference between the present input data and the predictive data corresponding to the present input data, the threshold value from the second predicting means. And a selection means for selecting either the current input data or the prediction data corresponding to the current input data based on the comparison result of the comparison means. apparatus. 2. The first predicting means uses a coefficient calculator for adaptively determining a prediction coefficient by performing a least squares operation on the past input data, and using the present input data and the prediction coefficient. The noise removing apparatus according to claim 1, comprising a signal value predicting unit that performs prediction. 3. A first storage means for storing the input data and a second storage means for storing the prediction data from the first prediction means are provided, and the second prediction means is the first storage means. 2. The noise eliminator according to claim 1, wherein the threshold value is adaptively changed based on a difference between the data from the storage means and the predicted data from the second storage means. 4. Predicting prediction data corresponding to current input data based on past input data, and predicting a threshold value based on past input data and past prediction data corresponding to the past input data. Then, the difference between the current input data and the prediction data corresponding to the current input data is compared with the above threshold value, and based on the comparison result, the current input data or the current input data is processed. A noise removal method characterized by selecting one of prediction data. 5. The prediction data is obtained by performing a least-squares operation on the past input data to adaptively determine a prediction coefficient, and performing prediction using the current input data and the prediction coefficient. The noise removing method according to claim 4, wherein the noise removing method is performed. 6. The input data and the prediction data are respectively stored, and the threshold value is adaptively changed based on a difference between the stored input data and the prediction data. Noise removal method.