JPH0861975A - Filter processing method and filter processing device for measuring data, and sectional continuous measuring system for physical quantity - Google Patents

Abstract

PURPOSE: To compensate a phase shift resulting from the phase characteristic of a filter without hurting the amplitude characteristic of the filter, by applying the first filtering process from the front end of a series of measured data, and the second filtering process from the terminal end of the processed measured data. CONSTITUTION: A series of measuring data obtained by a measuring device 3 is inputted to a storage device 6 once. The measuring data is outputted from the device 6 and the first filtering process is applied by a filter 1, and it is returned to the device 6. The data order of the measured data is reversed in the device 6, and it is delivered to the filter 1 again. The second filtering process with the same phase characteristic as the first filtering process is applied by the filter 1. The measured data processed twice by the filter 1 is returned to the device 6 again, and then delivered to an output device 7 after the data order is reversed in the device 6, or in the data order as it is. Consequently, an unnecessary data such as a noise can be removed without disturbing the characteristic of the waveform of the measuring data, for example.

Description

Detailed Description of the Invention

[0001]

FIELD OF THE INVENTION The present invention relates to a method and a device for processing a series of measurement data obtained by measurement.

[0002]

2. Description of the Related Art Generally, an apparatus for processing a series of measurement data obtained by measurement comprises a measuring means and a processing means for processing the measured data, and the measured data measured by the measuring means is sent to the processing means. Are processed. The measured data by this measuring means usually includes unnecessary information other than the desired information due to various causes such as noise. In addition, similar noise is often mixed in the process of transmitting the measurement data to the processing means.

Therefore, in order to remove such unnecessary information, it is widely practiced to provide a filter processing means in front of the processing means and filter the measurement data.

FIG. 6 schematically shows the case where noise is mixed in the measurement data.

For example, when the original measurement data is as shown in FIG. 6 (a), the impulsive noise 50 is mixed in, so that when the processing means is reached, the processing is performed as shown in FIG.
The waveform is as shown in (b). In this case, if the information to be obtained is, for example, the maximum value and the minimum value of the series of measurement data, the information to be obtained becomes incorrect due to the impulsive noise 50. Therefore, filter processing is performed to obtain correct measurement data.

However, in general, when noise or the like is removed by filtering, a phase shift (advance or delay) occurs due to the unavoidable phase characteristics of the filter, and the measured data after filtering is the same as that before mixing of noise or the like. It will be different.

Consider, for example, an analog filter. An analog filter handles electronic signals that are continuous both in terms of time and amplitude and is composed of electronic circuits. If the functional expression is H (jω) (where ω = 2πf: ω is an angular frequency, f is a frequency, j × j = −1, and π is a circular constant), then this filter has −20log | H (jω ) | Has an amplitude characteristic represented by |, and a phase characteristic represented by arg (H (jω)). This phase characteristic causes a phase shift in the measured data after the filtering process.

Further, the digital filter is composed of an electronic circuit which handles a digital signal which is discrete in time at a constant interval and discretely processes or a device which executes an operation. The digital filter is usually designed to obtain characteristics similar to those of the analog filter (amplitude characteristic, phase characteristic). For example, the input data string (analog signal) to the underlying analog filter is converted to a digital signal and processed by the digital filter,
When the processed data string (digital signal) is converted into an analog signal, the converted data string is designed to match the output data of the underlying analog filter. Therefore, also in the measurement data string (digital signal) processed by the digital filter, a phase shift occurs due to the phase characteristics of the analog filter which is the basis.

As described above, in the conventional filter processing method,
Due to this phase shift, the measured data after filtering is
Even if unnecessary information is removed, there is a problem that the measured data will be different from the original measured data.

[0010]

SUMMARY OF THE INVENTION The first object of the present invention is made in view of the above-mentioned points, and compensates the phase shift caused by the phase characteristic of the filter without impairing the amplitude characteristic of the filter. It is an object of the present invention to provide a method of filtering measurement data and a filter processing apparatus including a means for performing the filtering.

A second object of the present invention is to provide a piecewise continuous measurement system for physical quantities, which is capable of performing phase-compensated filter processing.

[0012]

The essence of the present invention for achieving the above object is as follows.

In the method of filtering measured data according to the present invention, a series of measured data is subjected to a first filtering process from the beginning, and the processed measured data subjected to the first filtering process is subjected to the first filtering process from the end. It is characterized in that a second filter process having substantially the same phase characteristic as the filter process is performed.

A preferred embodiment of the measurement data filtering method of the present invention is characterized in that the processed measurement data subjected to the second filtering is output from the end.

Another aspect of the method for filtering measurement data according to the present invention is that a series of measurement data is subjected to first filter processing from the end, and the processed measurement data subjected to the first filter processing is applied. A second filter process having substantially the same phase characteristics as the first filter process is performed from the beginning.

Another aspect of the method for filtering measured data according to the present invention is to perform a first filtering process from the beginning on each series of measured data of a group of series of measured data,
It is characterized in that each processed measurement data subjected to the filtering process is subjected to a second filtering process having substantially the same phase characteristics as the first filtering process from the end.

Another aspect of the method for filtering measurement data according to the present invention is that the first filtering process is applied to the measurement data sequence from the beginning, and the processed measurement data sequence subjected to the first filtering process is performed. A second filtering process having substantially the same phase characteristics as the first filtering process is performed from the end.

Another aspect of the method for filtering measured data according to the present invention is that the first filtered data is applied to the measured data string from the end and the processed data string subjected to the first filtered data is processed. A second filter process having substantially the same phase characteristics as the first filter process is performed from the beginning.

Another aspect of the measurement data filtering method of the present invention is that each measurement data of the measurement data string group is subjected to the first filter processing from the beginning, and each of the measurement data is subjected to the first filter processing. It is characterized in that the processed measurement data string is subjected to a second filtering process having substantially the same phase characteristics as the first filtering process from the end.

A preferred embodiment of the method for filtering measured data according to the present invention is characterized in that the first or second filtering is digital filtering.

Further, the measurement data filter processing apparatus of the present invention is provided with the measurement data acquisition means for obtaining a series of measurement data, the filter processing means for filtering the measurement data, and the processing by the filter processing means. It is characterized in that it is provided with a data order inverting means for the processed measurement data, and the filter processing means also has a function of filtering the output of the data order inverting means for the processed measurement data.

Another aspect of the measurement data filter processing apparatus of the present invention is a measurement data acquisition means for obtaining a series of measurement data, a first filter processing means for filtering the measurement data, and a first filter processing means. And a second filter processing means for filtering the output of the data order inversion means of the processed measurement data processed by the filter processing means, and the output of the data order inversion means of the processed measurement data.
The second filter processing means has substantially the same phase characteristic as that of the first filter processing means.

Further, the physical quantity piecewise continuous measuring system of the present invention comprises a physical quantity measuring means, a first filtering means for filtering the output of the measuring means, and the first filter at regular intervals. Data forward inverting means for inverting the output of the processing means in time, and second filter processing means for filtering the output of the data forward inverting means are provided.
Moreover, the second filter processing means has substantially the same phase characteristic as that of the first filter processing means.

A preferred embodiment of the physical quantity piecewise continuous measurement system of the present invention comprises a data re-reversal means for reversing the output of the second filter processing means with time at regular intervals. Is characterized by.

In a preferred embodiment of the physical quantity piecewise continuous measurement system of the present invention, the data forward inverting means also has a function of inverting the output of the second filter processing means at regular intervals. It is characterized by having both.

[0026]

In the filtering method and apparatus of the present invention, the data order of the series of measurement data that has been subjected to the first filtering process is reversed, and filtering is performed with substantially the same phase characteristics. Since the data order is reversed at the beginning of the second filter processing, the phase shift of each component by the first filter processing and the phase shift by the second filter processing are in opposite directions and have the same amount. Become. Therefore, the measurement data that has been subjected to this second filter processing has a phase shift of substantially zero and, in addition, has the same effect as the above-described filter processing performed twice in terms of amplitude characteristics.

Furthermore, if the data order of the measurement data that has been subjected to the second filter processing is reversed, the data order will also be the same as the first measurement data, and a filter that changes substantially only the amplitude characteristic without receiving a phase shift. Processing can be performed.

Hereinafter, with a low-pass filter having the amplitude-frequency characteristic shown in FIG. 7A and the phase characteristic shown in FIG. 7B, a series of measurement data shown in FIG. 8A is filtered. Taking the case as an example, the filtering method and apparatus of the present invention will be described.

Now, there is a series of measurement data as shown in FIG. 8A, and this data is mainly composed of components included in the frequency range 70 of the pass region of the filter shown in FIGS. 7A and 7B. The components of which are composed of offset (DC) components of the amplitude 74 at the beginning 72 of the measurement data, components that monotonically increase from the beginning 72 to the end 73 of the measurement data, the convex portion 80, and high-frequency noise not shown. It consists of The convex portion 80, the offset component, and the monotonically increasing component are components that have information to be obtained.

The measurement data is subjected to the first filter processing having the characteristics shown in FIG.

As a result, first, the offset component of the amplitude 74 at the head 72 of the data string and the component that monotonically increases in the direction from the head 72 to the tail 73 of the data string are mainly low frequency components. Phase shift hardly occurs by this first filter processing.

For example, the following two cases are assumed in digital filter processing, but in either case, phase shift hardly occurs. That is, in one embodiment of the digital filter processing, the condition before the beginning of the measurement data is set as a condition, and the setting of this condition is set as the continuous processing of the monotonically increasing component (even before the beginning of the measurement data, monotonically increasing. As described above, the phase shift after filtering the monotonically increasing component and the offset component is usually extremely small, and the phase delay hardly occurs. Further, when the setting of the above condition is set as the continuous processing of the offset component (assuming that the same value as the head before the head of the measurement data is assumed to be the same), the filter processing of the data string that changes from DC to monotonically increases And a slight phase delay after the filter processing occurs at the changing point (the head of the measurement data string). The filter processing result in this case is that the monotonically increasing component and the offset component are shifted from the beginning to the end of the measurement data by setting the above conditions,
The result is that the entire measurement data string has moved. In addition,
In this case, it is preferable to use a filter processing unit having a data processing amount having a margin larger than the data amount of the measurement data, a buffer of the measurement data after the processing, or the like in consideration of the shift of the measurement data after the first filtering. .

On the other hand, the convex portion 80 is mainly composed of a frequency component of a region which is included in the frequency range 70 of the pass region of the filter but is close to the boundary with the frequency range of the cutoff region. Therefore, when the first filtering process is performed, the amplitude is slightly attenuated but the large phase delay 78 is generated.
Receive.

The noise component has a high frequency, and the components mainly belonging to the frequency range of the cutoff region of the filter are
The first filtering of the characteristic of FIG. 7 results in a significant attenuation.

As a result, when subjected to the filtering process by the filter having the characteristic of FIG. 7, the measured data of FIG. 8A has only the convex portion 80 due to the phase delay 78 as shown in FIG. 8B. The data is relatively deviated from, and the noise component is considerably attenuated.

Subsequently, the measurement data after the first filtering (FIG. 8B) is subjected to the second filtering from the end 73 having the same characteristics as the first filtering. Then
Similarly to the above, the phase delay 79 shown in FIG. Since the filtering process is performed from the tail 73 of the data, the direction of the phase delay 79 is from the tail 73 to the head 72. The offset (direct current) component and the monotonically increasing component are hardly affected by the amplitude as well as the phase, as in the case of the first filter processing, and high-frequency noise is further attenuated.

Now, consider the amounts of the phase delay 78 and the phase delay 79. When the first filtering process is performed on the measurement data of FIG. 8A from the beginning 72 to obtain the result of FIG. 8B, and when the processed measurement data of FIG. 8B is processed from the end 73 to the second filtering process. To obtain the result shown in FIG. 8 (c),
Since the filtering process with the same phase characteristic is performed, the amount of the phase delay 78 of the convex portion 80 between FIG. 8A and FIG. 8B and the amount of the phase delay 78 between FIG. 8B and FIG. The amount of phase lag 79 in is the same in all frequency regions. Moreover, since the directions of the phase delay 78 and the phase delay 79 are opposite, these phase delays cancel each other out. Therefore, in principle, the phase shift caused by the first filtering process is completely removed by the second filtering process in all frequency regions.

Moreover, since the filtering process having the amplitude characteristic of FIG. 7A is performed twice, the components such as noises belonging to the frequency range 71 of the cutoff region in the measurement data are twice as steep as the characteristic of FIG. 7A. Attenuates due to various characteristics.

As a result, only the noise component can be attenuated with almost no loss of the convex portion 80 that carries the desired information, the offset component, and the monotonically increasing component.

The filtering process may be performed by a digital filter or an analog filter. Further, in the case of using a digital filter, a series of measurement data is a series of a large number of measurement data (a series of measurement data), but the above argument holds.

An embodiment of an apparatus for carrying out the measurement data processing method of the present invention will be described in detail below with reference to the drawings.

First, the configuration of the filter processing apparatus of this embodiment will be described.

FIG. 1 is a block diagram showing a schematic configuration of an embodiment of a filter processing device for measurement data. The filter processing device comprises a measuring means including a measuring device 3 and a process provided separately from the measuring device 3. The processing means 40 has a filter 1, a filter 2 having the same characteristics as the filter 1, a storage device 4, and an output device 5. Such a configuration is preferable when the measurement data is analog data.

A series of measurement data obtained by the measuring device 3 is first filtered by the filter 1 and then input to the storage device 4. The data order of the measurement data is inverted in the storage device 4 and sent to the filter 2. The measurement data, which has been subjected to the filtering process having substantially the same phase characteristic as that of the filter 1 by the filter 2, outputs the measured data
Is input to Further, in order to restore the data order to the order obtained by the measuring instrument 3, the data order may be output again after being inverted by the output device 5, or may be output without such data order restoration.

FIG. 2 shows a schematic configuration of another embodiment of the measurement data filtering device. In FIG. 2, the filter processing device is provided with a detecting means including the measuring device 3 and a processing device 40 provided separately from the measuring device. The processing device 40 includes the filter 1, the storage device 6, and the output device 7. have. Such a configuration is preferable when the measurement data is a digital measurement data string.

The series of measurement data obtained by the measuring device 3 is once input to the storage device 6. The measurement data is once subjected to the first filtering process from the storage device 6 by the filter 1 and then returned to the storage device 6. Then, the data order of the measurement data is inverted in the storage device 6 and is sent to the filter 1 again. The filter 1 performs a second filter process having the same phase characteristic as the first filter process. In this way, the measurement data that has been subjected to the filtering process by the filter 1 twice returns to the storage device 6 again, and the data order is inverted in the storage device 6 or sent to the output device 7 as it is. The inversion of the data order may be performed at the output device 7.

Further, the data order of the measurement data is inverted before passing through the filter 1 in the memory device 6, this data string is once passed through the filter 1 and returned to the memory device 6, and the memory device 6
Restore the data order before passing filter 1 again with
The data sequence whose order has been restored may be passed through the filter 1 again and returned to the storage device 6. Also in this case, the output device 7 outputs the data sequence without inverting the order. In this case, the first filtering process is performed from the end of the measurement data and the second filtering process is performed from the beginning. In this case as well, the operation of the present invention is exactly the same as the above case.

Further, instead of including the measuring means as in the above-described embodiment, the measured data is recorded on a recording medium such as a magnetic tape and is reproduced by a reading device or the like. The measurement data may be acquired.

The procedure in one embodiment of the filter processing apparatus of the present invention shown in FIG. 2 will be described below with reference to the flowcharts shown in FIGS. 3, 4 and 5. Note that here, the measurement data is assumed to be a digital measurement data string.

An outline of a series of measurement and filter processing is performed in the following procedure. The measuring instrument 3 starts the measurement, and the processing means 40 fetches the measurement data from the measuring instrument 3 and the processing means 4
It is recorded in the memory device 6 of 0 (step S1). After the data acquisition (step S1) is completed, the filtering process of the present invention is performed on the obtained measurement data string (step S).
2). After this, the processing result is output (step S3) and the series of operations is ended.

Hereinafter, details of execution of the filtering method according to the present invention (FIG. 3, step S2) will be described with reference to FIGS. 4 and 5.

Execution of filter processing (FIG. 3, step S)
When 2) is started, the processing means 40 sends the measurement data string from the storage device 6 to the filter 1 (FIG. 4, step S).
7). The measurement data string sent next is subjected to first filter processing by the filter 1 from the beginning (step S8). Then,
The measurement data string that has been subjected to the first filter processing is sent to the storage device 6, and the order of the measurement data string after processing in the storage device 6 is reversed (step S9). Next, the second filter process is performed from the end of the measured data string after the filter process with the same filter 1 as in step S8 (step S10).

Finally, the filtered data string is sent from the end to the storage device 6, and the order of the processed data string is reversed again. (Step S11).

The details of step S8 (step S10) for filtering the data string from the beginning (end) will be described below.

First, the value at the beginning (end) of the data string is set as the offset value (step S12).

Next, this offset is removed from each measured value of the measured data string (step S13). This is to subtract the offset value from each value in the data string,
For example, when the data string is X ₁ , X ₂ , ..., X _n (n is an integer) and the offset value is X ₁ , the data string after the offset removal is X ₁ -X ₁ , X ₂ -X ₁ , ..., X _n −X ₁ (n is an integer). This enables the filtering process based on the assumption that the value before the head of the measurement data string is the same as the offset value.

The data sequence after the offset is removed is filtered from the beginning (end) of the data sequence (step S1).
4). This filtering process will be described later. The offset of the filtered data string is restored (step S1
5) Then, the process of step S8 (step S10) of filtering the data string from the beginning (end) is completed.

This offset restoration means adding the same amount of offset value removed in step S13 to each value of the data string, thereby restoring the DC component of the measured data string.

Removal of such offset (step S
Instead of 13) or the restoration of the offset (step 15), the data before the beginning of the measurement data sequence can be set to the same value as the beginning of the measurement data sequence. In this case, the virtual data before the beginning is created. Also,
The convex portion 80 of FIG.
If it is known in advance that such a change will be superposed, the processing may be performed assuming that there is monotonically increasing data with the same slope as before and after the beginning of the measurement data sequence. . Further, in the above example, the same offset is removed from all the measurement data belonging to the measurement data string, but for example, the slope of the monotonic increase (decrease) from the measurement data at the beginning and the end (or by the least squares method) is used. It is also preferable to obtain the offset, set the offset for each measurement data, remove it, and restore it after the second filtering.

Here, the filter processing will be described. In this embodiment example, the filter 1 of the processing means 40 realizes the filter processing by using a digital filter. For example, the data string to be filtered is X ₁ , X ₂ , ..., X _n (n is an integer) The data string after filtering is Y ₁ , Y ₂ , ..., Y _n (n is an integer) ) And the functional expression of the filter is H (X _n-2 , X _n-1 , X _n , Y _n-2 , Y _n-1 ), the functional expression of the filtering process of this embodiment is , Y _n = H (X _n-2 , X _n-1 , X _n , Y _n-2 , Y _n-1 ). At this time, if sequentially calculating from the head X ₁ of the input measurement data string, Y ₁ = H (X _-1 , X ₀ , X ₁ , Y _-1 , Y ₀ ), Y ₂ = H (X ₀ , X ₁ , X ₂ , Y ₀ , Y ₁ ), ... _Yn = H (X _n-2 , X _n-1 , X _n , Y _n-2 , Y _n-1 ) The measurement data sequence of is obtained.

[0061] the top X ₁ than the previous data (X _-1, X ₀₎ of the data string to filter in this example of the value and the previous data from the head Y ₁ data string after filtering (Y _-1,
In order to set the value of Y ₀ ) to zero (X ₋₁ = X ₀ = Y ₋₁ = Y ₀ = 0), the value at the beginning (end) of the data string is used as an offset value (step 12) Column offset is removed (step S13) and the offset of the filtered and filtered data row is restored (step S15).
Processing is being carried out.

In order to carry out the above processing continuously or intermittently, the measurement data are divided at appropriate intervals to form a group of measurement data. This is achieved by performing the above-described filter processing on each of a series of divided measurement data (measurement data string) and connecting the obtained groups of filtered measurement data.
It should be noted that this interval need not always be constant, and may be changed depending on the value of data. Further, in order to perform the measurement seamlessly, it is preferable to store data for fetching the next series of measurement data during the inversion of the data order of the series of measurement data. Therefore, it is preferable to prepare two sets of means for inverting the data order of the measurement data and use them while sequentially switching them.

Further, the series of measurement data contained in the group of measurement data may be obtained by dividing the output of the same measuring means at appropriate time intervals, for example, and measuring the output of each independent measuring means at the same time. The processed products may be sequentially subjected to the above-mentioned processing. It is also possible to construct a piecewise continuous measurement system of physical quantities that continuously measures the measurement values of the same measurement means while performing such filter processing. In this case, the measurement results are obtained by sequentially measuring a large number of measurement targets that should be able to obtain roughly the same measurement results, and the above filter processing is performed for each measurement data sequence to obtain the measurement results for each measurement target. Good. Since the filtering process of such a measurement system only needs to reverse the data order, it can be performed at extremely high speed. Therefore, for example, it is possible to easily configure a real-time measurement system with a high-speed filtering function that can almost eliminate the phase shift.

In the present invention, the series of measurement data means
Refers to a series of measured values obtained by measuring a physical quantity by an appropriate measuring means. For example, it can be expressed as a function in which the position (time) is an independent variable and the displacement is a dependent variable, such as a measurement value obtained by continuously measuring the displacement of the object surface with respect to the position or time. The independent variable and the dependent variable may be digital quantities or analog quantities. When the independent variable is a digital quantity, a series of measurement data can also be called a measurement data string. Also,
Since the data order of a series of measurement data is reversed, the range of the independent variable of the measurement data needs to be finite. Therefore, when the independent variable is a digital quantity, the number of data contained in the measurement data string is always finite. Further, the independent variable may be, for example, an identification number of a measuring device when a physical quantity is measured using a large number of measuring devices. Further, in the present invention, a group of a series of measurement data refers to a group of measurement data having a plurality of series of measurement data as described above.

Further, in the present invention, the head of a series of measurement data refers to the beginning part of the measured or acquired measurement data, and the end refers to the last part of the measured or acquired measurement data. Even when the data order is reversed, the part corresponding to the part that was originally the beginning of the measurement data is called the beginning of the inverted data. Therefore, after the first filtering, the last part of the processed measurement data whose data order has been inverted is the "head" of this data. The same applies to the "tail".

Further, in the present invention, the data order of the measurement data may be reversed after the measurement, and the first filtering process may be performed from the end of the measurement data. In this case, the data order of the measurement data subjected to the second filter processing becomes the same as the data order of the first measurement data due to the inversion of the data order after the first filter processing. Also in this case, the same operation as in the case where the first filter processing is performed from the beginning of the measurement data is obtained. Therefore, it is preferable to measure the measurement data with the measuring device, store it in the storage means once, and read it to obtain the data because the processing time can be shortened.

In the present invention, the physical quantity means the length,
Any quantity that can be quantitatively measured such as temperature, mass, and time may be used. When these physical quantities are measured by measuring means,
In order to facilitate filter processing, it is often converted into a voltage, a current, a magnetic quantity, or the like, but in the present invention, a value that includes such a conversion can be treated as a measured value.

In the present invention, both analog and digital filter processing are preferably used as the filter processing. A digital filter that easily matches the characteristics of the first filter processing and the characteristics of the second filter processing and that easily reverses the data order is particularly preferably used. In addition, the characteristics of the filter include a low-pass filter, a high-pass filter, and a frequency characteristic of space or time.
Bandpass filters and band eliminate filters having all characteristics are preferably used.

In the present invention, the first filtering process and the second filtering process need to have substantially the same characteristics. Here, having substantially the same phase characteristic means that the phase characteristics are approximated to the extent that the effect of the phase shift due to the first filtering process can be removed to an extent sufficient to obtain the desired information. To do. When the measurement data is digital data, the phase characteristics of the first filtering process and the second filtering process can be substantially completely matched, which is preferable. Therefore, the reproducibility is also high.

In the present invention, the following means are preferably used as means for inverting the data order of a series of measurement data (or means for processing from the end of measurement data) or re-inversion means. That is, when the measurement data is analog (especially when the independent variable is analog), first, a series of measurement data is recorded on an analog recording medium such as a magnetic tape or a photographic film while the medium is rewound in the forward direction. This is realized by a method such as later reproducing the analog recording medium while rewinding it in the opposite direction. On the other hand, when the measured data is digital (especially when the independent variable is digital), a series of measured data is first stored in a digital storage medium such as a semiconductor memory in order of address, and then from the end of the memory in reverse order of address. It is realized by playing. From the viewpoint of the processing speed of the measurement data, it is preferable to use a digital measurement data sequence and invert the data order by the semiconductor memory or the like as described above. From this point as well, it is preferable that the measurement data be a digital quantity. When re-inverting the data, it is preferable to invert the data at the same intervals as the inversion means for the measurement data and the processed measurement data obtained by filtering the measurement data.

Further, in the present invention, as the measurement data acquisition means, a measuring instrument for measuring a physical quantity, a storage means or a recording medium recording the measurement data by the measuring instrument are preferably used.

[0072]

EXAMPLES Examples of the method for filtering measured data, the filtering apparatus, and the system for continuously measuring physical quantities in a piecewise manner according to the present invention will be described below.

Example 1 A shape measuring system for metal parts was manufactured using the measurement data filtering apparatus having the configuration shown in FIG.

As the measuring device 3, a contact profilometer for measuring the surface displacement of the metal parts was used, and an apparatus for successively measuring the shapes of the surfaces of many metal parts was constructed.
Each metal part is manufactured by mechanical cutting based on the same design (the surface is flat at the same height), and the surface shape is almost the same. Therefore, this device aims to measure the slight unevenness of each of these metal parts. In this device, one series of measurement data is acquired for each metal part, and the output of the profile meter becomes 0 in the middle of each part. Therefore, one measurement data group can be acquired as a whole. The output can be taken out as a voltage value.

The output of the measuring device 3 was converted into digital data by an A / D converter (not shown). As the filter 1, a low-pass Butterworth filter having the frequency characteristic shown in FIG. 7 is realized by a digital filter.

In this example, since the measurement data string sampled periodically is processed, the interval of the mathematically discrete data is set to be constant. In order to avoid the occurrence of aliasing error due to the relationship between the filter characteristic and the sampling period, the original analog filter characteristic is approximated by bilinear Z conversion.

As the storage device 4, a microcomputer and its memory device were used. The measurement data subjected to the first filter processing by the filter 1 can be temporarily stored in the memory device in the order of addresses, and then the data can be taken out from the end of the address of the data in the reverse order of the addresses and sent to the filter 2.

As the filter 2, a filter having exactly the same design as the filter 1 was used.

As the output device 5, the above-mentioned microcomputer screen or its screen hard copy was used. It should be noted that the measurement data string was re-inverted prior to output.

The surface shape of each part of the group of metal parts was measured by such an apparatus.

FIG. 9 shows, as a model, the measured value (output of the measuring device 3) of the surface shape of one of the groups of metal parts. The convex portion on the left side of FIG. 9 indicates the deformation of the metal part, and the component with a short cycle superimposed on this is apparently noise. This measurement data string was filtered by the above device. As a result, as shown in FIG. 10, noise mixed in the measurement data was removed without phase delay.

The above processing was performed on all metal parts belonging to the group. As a result, some metal parts having abnormal protrusions as shown in FIG. 10 were discovered.

In the above, the first filter processing was performed from the beginning of the measurement data, but conversely, when this was performed from the end and the second filter processing was performed from the beginning, almost the same results were obtained.

Embodiment 2 The embodiment is different from Embodiment 2 except that the output of the measuring device 3 is not subjected to A / D conversion, an analog filter having the same characteristics as the above digital filter is used, and the data order is reversed by a magnetic tape. A shape measuring system for metal parts similar to that of 1 was manufactured.

With this system, the shape of each metal part of the same group of metal parts as in Example 1 was measured, and the same results as in Example 1 could be obtained.

Example 3 The same metal part shape measuring system as in Example 1 was manufactured except that the filter 2 was not used and the second filtering was performed by the filter 1.

With this system, the shape of each metal part of the same group of metal parts as in Example 1 was measured, and the same results as in Example 1 could be obtained.

Comparative Example A metal part shape measuring system similar to that of Example 1 was manufactured except that the second filtering process and the data order inversion were not performed.

With this system, the shape of the same metal part as in Example 1 was measured, and a measurement data string similar to that in FIG. 9 was acquired. As a result of filtering the measurement data string with the filter 1, processed measurement data as shown in FIG. 11 was obtained. Due to the phase delay of the filter 1, the convex portion at the beginning of the data string moved toward the end of the data string, and the waveform was different from that of the first embodiment. Further, the noise component was recognized to be slightly larger than that in the case of FIG.

[0090]

According to the measurement data filtering method of the present invention, a series of measurement data is subjected to the first filter processing from the beginning (or the end), and the processed measurement data is processed from the end (or the beginning) to the first. Since the second filter processing having substantially the same phase characteristic as that of the second filter processing is performed, it is possible to realize the filter processing in which almost no phase shift occurs in any frequency region. Thereby, for example, unnecessary information such as noise can be removed without disturbing the characteristics of the waveform of the measurement data.

According to another aspect of the measurement data filtering method of the present invention, each series of measurement data in the series of measurement data is subjected to the first filter processing from the beginning,
Since each processed measurement data is subjected from the end to the second filtering process having substantially the same phase characteristic as the first filtering process, the filtering process that does not generate the phase shift in every frequency region is performed continuously or intermittently. Can be done

According to another aspect of the measurement data filtering method of the present invention, the measurement data string is subjected to the first filtering process from the beginning (or end) and the processed measurement data string is terminated (or Since the second filtering process having substantially the same phase characteristics as the first filtering process is performed from the first), the filtering process is fast and reproducible in digital data processing and hardly causes phase shift in all frequency regions. Can be realized.

According to another aspect of the measurement data filtering method of the present invention, each measurement data string in the group of measurement data strings is subjected to the first filter processing from the beginning, and each processed measurement data is added to the end. Since the second filter processing having substantially the same phase characteristic as that of the first filter processing is performed, the filter processing is continuously performed at high speed and reproducibility in digital data processing and almost no phase shift occurs in any frequency region. Can be done intermittently or intermittently.

Further, according to the measurement data filter processing apparatus of the present invention, the measurement data acquisition means for obtaining a series of measurement data, the filter processing means for filtering the measurement data, and the processing by the filter processing means are performed. And a filtering means for filtering the output of the data order reversing means for the processed measurement data. It is possible to realize a filter processing device that hardly generates a phase shift in a region.

According to another aspect of the measurement data filter processing apparatus of the present invention, a measurement data acquisition means for obtaining a series of measurement data, a first filter processing means for filtering the measurement data, and A data order inversion means for the processed measurement data that has been processed by the first filter processing means, and a second filter processing means for filtering the output of the data order inversion means for the processed measurement data.
And since the second filter processing means has substantially the same phase characteristics as the first filter processing means,
It is possible to realize a filter processing device that causes almost no phase shift in any frequency region.

According to the physical quantity piecewise continuous measuring system of the present invention, the physical quantity measuring means, the first filter processing means for filtering the output of the measuring means, and the first physical quantity measuring means at regular intervals. Data forward inverting means for inverting the output of the filter processing means in time, and second filter processing means for filtering the output of the data forward inverting means, and the second filter processing means is substantially Since it has the same phase characteristic as that of the first filter processing means, it is possible to perform the data processing of the group of measurement data while performing the filter processing with almost no phase shift for all frequency components. . As a result, for example, it is possible to configure a measurement system that sequentially filters the outputs of the same measurement device without phase shift in real time.

[Brief description of drawings]

FIG. 1 is a block diagram showing a schematic configuration of an embodiment of a measurement data filtering method and filtering apparatus of the present invention.

FIG. 2 is a block diagram showing a schematic configuration of an embodiment of a measurement data filtering method and filtering apparatus of the present invention.

FIG. 3 is a flowchart showing an outline of a procedure in one embodiment of the filtering method of the present invention.

FIG. 4 is a flowchart showing an outline of a procedure of filter processing in an embodiment of the filter processing method of the present invention.

FIG. 5 is a flowchart showing details of a procedure of filter processing in one embodiment of the filter processing method of the present invention.

FIG. 6 is a diagram showing an example for explaining the purpose of the filtering method of the present invention.

FIG. 7 is a diagram showing an example of filter characteristics of filter processing means used in the filter processing method and filter processing apparatus of the present invention.

FIG. 8 is a model diagram for explaining the operation of the filter processing method and filter processing apparatus of the present invention.

FIG. 9 is a model view showing the output of the measuring device in one embodiment of the filtering method, the filtering device and the physical quantity piecewise continuous measuring system of the present invention.

FIG. 10 is a diagram showing a model of a filtered output in one embodiment of the filtering method, the filtering device and the physical quantity piecewise continuous measuring system of the present invention.

FIG. 11 is a diagram showing a modeled output that has been subjected to filter processing when a conventional filter processing method is used.

[Explanation of symbols]

 1: Filter 2: Filter 3: Measuring device 4: Storage device 5: Output device 6: Storage device 7: Output device 40: Processing means 50: Impulsive noise 70: Frequency range of pass region 71: Frequency range of cut-off region 72 : Measurement data sequence head 73: Measurement data sequence end 74: Measurement data sequence start amplitude 75: Convex part amplitude 76: Convex part amplitude 77: Measurement data sequence end amplitude 78: Phase delay 79: Phase delay 80: Convex part

Claims (13)

[Claims] 1. A series of measurement data is subjected to a first filtering process from the beginning, and the processed measurement data subjected to the first filtering process has a phase substantially the same as that of the first filtering process from the end. A method for filtering measurement data, characterized by performing a second filtering process having characteristics. 2. The measurement data filtering method according to claim 1, wherein the processed measurement data subjected to the second filtering is output from the beginning. 3. A series of measurement data is subjected to a first filtering process from the end, and the processed data subjected to the first filtering process has a phase substantially the same as that of the first filtering process from the beginning. A method for filtering measurement data, characterized by performing a second filtering process having characteristics. 4. A first filter process is applied to each series of measurement data of a series of measurement data from the beginning, and each processed measurement data subjected to the first filter process is applied to the first from the end. A method for filtering measurement data, which comprises performing a second filtering process having substantially the same phase characteristics as the filtering process. 5. A first filtering process is applied to the measurement data sequence from the beginning, and a processed measurement data sequence subjected to the first filtering process has a phase substantially the same as that of the first filtering process from the end. A method for filtering measurement data, characterized by performing a second filtering process having characteristics. 6. A measurement data string is subjected to a first filtering process from the end, and a processed measurement data string subjected to the first filtering process has a phase substantially the same as that of the first filtering process from the beginning. A method for filtering measurement data, characterized by performing a second filtering process having characteristics. 7. A first filter process is applied to each measurement data of the measurement data string group from the beginning, and each processed measurement data string subjected to the first filter process is added to the first filter process from the end. A method for filtering measurement data, characterized by performing a second filtering process having substantially the same phase characteristics. 8. The method according to claim 5, wherein the first or second filter processing is digital filter processing.
7. The method for filtering measurement data according to any one of 7. 9. A measurement data acquisition means for obtaining a series of measurement data, a filter processing means for filtering the measurement data, and a data forward inverting means for the processed measurement data processed by the filter processing means. A filter processing apparatus, comprising: the filter processing means also having a function of filtering the output of the data forward inversion means of the processed measurement data. 10. A measurement data acquisition means for obtaining a series of measurement data, a first filter processing means for filtering the measurement data, and a processed measurement data processed by the first filter processing means. Data forward inverting means and second filter processing means for filtering the output of the data forward inverting means of the processed measurement data are provided, and the second filter processing means is substantially the first filter processing. A filtering apparatus having the same phase characteristic as that of the means. 11. A physical quantity measuring means, a first filter processing means for filtering the output of the measuring means, and a data sequence for inverting the output of the first filter processing means at regular intervals. It comprises an inverting means and a second filter processing means for filtering the output of the data forward inverting means, and the second filter processing means has substantially the same phase characteristic as the first filter processing means. A piecewise continuous measurement system for physical quantities, comprising: 12. A piecewise continuous physical quantity according to claim 11, further comprising data forward re-inversion means for inverting the output of said second filter processing means at regular intervals. Measuring system. 13. The data forward inverting means also has a function of inverting the output of the second filter processing means temporally at regular intervals. A system for continuous measurement of physical quantities.