JP7044604B2 - Signal processing method for photoelectric encoder - Google Patents

Description

The present invention relates to a signal processing method of a photoelectric encoder.

FIG. 1 shows the main configuration of the photoelectric encoder 100.
A photoelectric encoder is widely used as a position or displacement measuring device.
The photoelectric encoder 100 is located from a light source 110, a main scale 200 having a scale pattern 210, a light receiving detector 300 that receives reflected light or transmitted light from the main scale 200, and a light / dark pattern received by the light receiving detector 300. Alternatively, it is provided with a signal processing unit 400 for obtaining displacement.

Here, the light from the light source 110 is made into parallel light by the collimated lens 112, and then the main scale 200 is irradiated with the light. Further, the light source 110, the lens 112 and the light receiving detector are unitized as a detection head, and the detection head is provided so as to be relatively displaceable with respect to the main scale 200.

In the scale pattern 210, if it is a transmissive photoelectric encoder, transmissive portions and non-transmissive portions are alternately arranged at a predetermined pitch or a random pitch. Further, in the case of the reflective photoelectric encoder, the scale pattern is such that the reflective portions and the non-reflective portions are alternately arranged at a predetermined pitch or a random pitch.

For example, as shown in FIG. 1, a light receiving detector 300 irradiates a scale pattern 210 in which transmissive portions and non-transmissive portions are alternately arranged at random pitches, and a light / dark pattern formed by the transmitted light is used as an image. Suppose you got it in. Ideally, as illustrated in FIG. 2, the acquired light intensity distribution of the light / dark pattern has the same maximum light intensity corresponding to the transmissive portion and the same minimum light intensity corresponding to the non-transmissive portion. Should be. With such an ideal light intensity distribution, a bright part and a dark part can be discriminated by one fixed threshold value.

However, in reality, the light intensity distribution of the light-dark pattern is distorted as illustrated in FIG. 3 due to various factors such as the light amount distribution of the light source 110, the distortion of the lens 112, and the boundary portion between the lenses of the microlens array. When a fixed threshold value is applied to such a distorted light-dark pattern, it is not possible to accurately distinguish between a bright part and a dark part.

Therefore, conventionally, light and dark are discriminated by calculating a threshold value by a filtering process that removes long-period distortion by Fourier transform.

JP-A-2017-58138

However, in the filtering by the Fourier transform, there is a problem that light and dark cannot be correctly discriminated when the light intensity distribution includes a local bias. For example, in the circled part in FIG. 4, the peaks of the bright parts on both sides are high, so that the bright parts in between cannot be detected.
In addition, the Fourier transform has a large amount of calculation and a large calculation load. Therefore, it takes a long time to perform each measurement, which is an obstacle to speeding up.
Further, in order to speed up the Fourier transform, the number of light receiving elements of the light receiving detector 300 must be 2 ⁿ . This has been a big limitation for the design of the light receiving detector 300.

Therefore, there is a demand for a photoelectric encoder capable of improving the reliability of light-dark analysis, speeding up measurement, and improving the degree of freedom in design.

An object of the present invention is to provide a signal processing method for a photoelectric encoder capable of accurately discriminating between light and dark by setting an appropriate threshold value even from a light and dark pattern in which the light intensity distribution is distorted.

The signal processing method of the photoelectric encoder of the present invention is
A scale in which a two-level code pattern according to a pseudo-random code sequence is provided along the length measurement direction, and
It is a signal processing method of a photoelectric encoder provided with a detection unit which is provided so as to be relatively movable along the scale and detects an absolute position on the scale based on a pseudo-random code sequence on the scale. ,
When one code of the two-level code pattern is the code "1" or the code "0" and one of the two bits representing one code is L or H,
The detector is
Acquire the detection image of the scale and
The detected image is subjected to morphology processing to calculate the lower envelope and the upper envelope.
The average line between the lower envelope and the upper envelope is obtained, and this is used as a threshold line.
The binarization process is performed by comparing the detected image with the threshold line.
In the morphology process, the sub-section having a predetermined width L is moved from one end to the other end of the detection image, and the minimum value of the light intensity is extracted from the data in the sub-section to extract the lower envelope. It is characterized in that a line is obtained and the maximum value of light intensity is extracted from the data in the sub-section to obtain the upper envelope.

In one embodiment of the invention
The width of one bit is W / 2,
The code "1" is represented by the A pattern, which is a set of L and H.
The symbol "0" is represented by a B pattern, which is a set of L and L, and a C pattern, which is a set of H and H.
When the reference numerals "0" are continuous, the B pattern and the C pattern are arranged alternately.
The width L of the sub-section is preferably 1.5 W <L <2 W.

In one embodiment of the invention
The width of one bit is W / 2,
The code "1" is represented by the A pattern, which is a set of L and H.
The symbol "0" is represented by a B pattern, which is a set of L and L, and a C pattern, which is a set of H and H.
When the reference numerals "0" are continuous, the B pattern and the C pattern are allowed to be continuous up to K, respectively.
The width L of the sub-section is preferably (K + 0.5) W <L <(K + 1) W.

It is a figure which shows the main structure of a photoelectric encoder. It is a figure which shows an example of a light receiving pattern. It is a figure which shows an example of a light receiving pattern with distortion. It is a figure which shows an example of a light receiving pattern with distortion. It is a figure which shows an example of a scale pattern. It is a figure which exemplifies the combination of the bit which expresses the symbol 1 and the symbol 0. It is a flowchart of the procedure of calculating an absolute position from the acquired light-dark pattern. It is a figure which illustrates the coded result. It is a flowchart which shows the whole flow of the quantization process ST200. It is a flowchart which shows the procedure of the lower envelope calculation process. It is a figure which shows typically the process of calculating the lower envelope EL. It is a figure which illustrated the lower envelope, the upper envelope and the threshold line.

An embodiment of the present invention will be illustrated and described with reference to the reference numerals attached to each element in the figure.
(First Embodiment)
The present invention relates to a signal processing method for a photoelectric encoder capable of accurately discriminating between light and dark even from a light and dark pattern in which the light intensity distribution is distorted.
The point of the present invention lies in "setting an appropriate threshold value", but as a premise that the signal processing (threshold value setting) of the present invention can be applied, there is a rule (design rule) in how to provide the scale pattern 210.
Therefore, first, how to provide the scale pattern 210 will be described.

(Explanation of scale pattern)
FIG. 5 is an example of the scale pattern 210.
This scale pattern 210 is an ABS (Absolute) scale pattern using an M-sequence code which is a kind of pseudo-random code sequence.
When the consecutive N codes in the M-sequence code pattern are taken out, the same code pattern of the N codes appears only in one place in one cycle of the M-sequence code pattern.
In the example of FIG. 5, in the ABS scale pattern in which "1" and "0" are randomly arranged, the codes of "1" and "0" are represented by 2 bits, respectively.
The symbol "1" is a combination of a non-transparent portion and a transmissive portion.
On the other hand, the reference numeral "0" is a transparent portion for both 2 bits or a non-transparent portion for both bits.

Here, for later explanation, the transparent portion is referred to as a “bright portion” (or “H”), and the non-transparent portion is referred to as a “dark portion” (or “L”).

The feature of this scale pattern 210 is that two kinds of patterns (transparent part for both 2 bits and non-transparent part for both bits) are used to express the symbol "0" (FIG. 6).
Now, the non-transparent portion of both bits is referred to as a B pattern, and the transmissive portion of both bits is referred to as a C pattern.
In expressing the symbol "0", the B pattern and the C pattern are used alternately. That is, the B pattern and the C pattern are arranged alternately at the positions where the reference numerals "0" are continuous. When this arrangement rule is applied, the bright part (H) and the dark part (L) should be continuous at a maximum of three.

(Explanation of operation)
The procedure for calculating the absolute position from the light / dark pattern acquired by the light receiving detector 300 is shown in the flowchart of FIG.
This arithmetic processing operation is executed by the signal processing unit 400.
First, the signal processing unit 400 sequentially sweeps the signals from the light receiving element array of the light receiving detector 300 and acquires the transmitted light pattern of the ABS scale 200 as an image image (ST110). Then, the acquired detection images are sequentially quantized (ST200). That is, the threshold value is determined for the light intensity of the acquired light / dark pattern, and the dark part and the bright part are distinguished and binarized.
The threshold setting at this time will be described later.
The dark part is "L" and the bright part is "H".

Subsequently, encoding is performed (ST310).
Two bits represent one code.
The set of (L, H) is converted to the code "1".
The pair of (L, L) and (H, H) is converted to the sign "0".
FIG. 8 is an example of the coded result.

Correlation calculation with the reference pattern is performed using the data encoded in this way (ST320). The position showing the highest correlation in the correlation operation is obtained as the current absolute position (ST330).
As the reference pattern, for example, a scale pattern provided according to the design rule may be used.

(Quantization process ST200)
Now, the overall flow of the quantization step ST200 is shown in the flowchart of FIG.
The quantization step ST200 includes a threshold line calculation step ST201 and a light / dark determination step ST240.
Here, the threshold line calculation step ST201 is a center line calculation step ST230 for calculating an intermediate line between the lower envelope calculation step ST210, the upper envelope calculation step ST220, and the upper envelope EU and the lower envelope EL. And have.

FIG. 10 is a flowchart showing the procedure of the lower envelope calculation step ST210. Further, FIG. 11 is a diagram schematically showing the process of calculating the lower envelope EL.
In the lower envelope calculation step ST210, first, a sub-section having a width L is set (ST211). The width L of the sub-section is expressed as follows, where W is the width of one sign of the scale pattern 210.

1.5W <L <2W

By the way, as a design rule of the scale pattern 210, the B pattern and the C pattern are arranged alternately in the place where the sign "0" is continuous, so if the width L of the sub-section is set by the above formula, it is appropriate for setting the threshold value. Envelopes (upper envelope EU, lower envelope EL) can be obtained.
If the design rule is adopted that the B pattern and the C pattern are not "alternate" but are allowed to be continuous up to K, the range of the width L is as follows.

(K + 0.5) W <L <(K + 1) W

While moving this sub-section from the left to the right of the light receiving region (ST212, ST216), the minimum value of the light intensity is extracted from the data in the sub-section (ST213), and the extracted data are sequentially extracted from the lower envelope EL. It is registered as the data of (ST214).
The shift amount in the sub section is equal to the pitch P of the light receiving element array (ST216).
This procedure gives the lower envelope EL (FIGS. 11 and 12).

Although expressed as a "line", the lower envelope EL obtained by the above processing is a sequence of discrete point data. Smoothing or interpolation processing may be performed on this discrete point data, or an approximate curve such as a spline curve may be obtained. However, even if interpolation or curve fitting is not performed, the upper envelope EU, which will be described later, has the same number of data as the lower envelope EL, so the center line (threshold line) from the upper envelope EU and the lower envelope EL. TL) is required.
Furthermore, each data constituting the center line (threshold line TL) corresponds to each data of the light / dark pattern acquired by the light receiving detector (light receiving element array) in order from the left (one point at the right end is discarded. It may be, but it is trivial.)
Therefore, the light / dark determination can be made by comparing each data of the light / dark pattern with each data of the threshold line TL.
Therefore, operations such as interpolation and curve fitting are unnecessary.

The upper envelope EU is also calculated by the same procedure (ST220). The only difference from the case of the lower envelope EL is that the maximum value of the light intensity is extracted from the data in the sub-section.

Of course, the maximum value and the minimum value may be extracted at the same time when the data in the sub-section is viewed. If the maximum value is registered as the point of the upper envelope EU and the minimum value is registered as the point of the lower envelope EL, when the sub-section moves from the left end to the right end, the upper envelope EU and the lower envelope EL become Each can be obtained.

In the center line calculation step (ST230), a line just intermediate between the upper envelope EU and the lower envelope EL is obtained. That is, the points of the upper envelope EU and the points of the lower envelope EL may be added by the corresponding points and divided by 2 (that is, the average is obtained). The center line thus obtained is defined as the threshold line TL.

The light-dark pattern acquired by the light-receiving detector 300 is compared with the threshold line TL, and whether it is light or dark is binarized. Now, the light-dark pattern acquired by the light-receiving detector 300 is quantized. After this, since coding and correlation calculation are known techniques, explanations are omitted.

According to this embodiment, the following effects are obtained.
The upper and lower envelopes (EU, EL) are obtained for the light-dark pattern acquired by the light-receiving detector 300, and the middle between the upper envelope EU and the lower envelope EL is set as the threshold line TL.
At this time, since the width L of the sub-section was appropriately set, the upper and lower envelopes (EU, EL) that appropriately followed the distortion of the acquired light-dark pattern were obtained, and therefore, the distortion of the light-dark pattern was appropriately followed. The threshold line TL is obtained.
(It should be noted that the reason why the width L of the sub-section can be appropriately set in this way is that the upper limit for continuous "bright" and the upper limit for continuous "dark" are set as the design rule of the scale pattern 210.)
This provides a threshold line TL that accurately discriminates between all "bright" and all "dark", no matter how distorted the light-dark pattern of the detected image. Accurate quantization leads to accurate coding and correlation calculation, which leads to improvement of measurement accuracy and measurement reliability of photoelectric encoder.

Further, in this embodiment, the calculation is simpler than the Fourier transform, and there is no limitation on the number of light receiving elements.

The present invention is not limited to the above embodiment, and can be appropriately modified without departing from the spirit.
In the above embodiment, the ABS scale is a transmissive type, but a reflective scale may be used.
Although linear scales and encoders have been exemplified in the above embodiments, the present invention can also be applied to rotary encoders.

100 ... photoelectric encoder,
110 ... light source, 112 ... lens,
200 ... main scale, 210 ... scale pattern,
300 ... light receiving detector,
400 ... Signal processing unit.

Claims (3)

A scale in which a two-level code pattern according to a pseudo-random code sequence is provided along the length measurement direction, and
It is a signal processing method of a photoelectric encoder provided with a detection unit which is provided so as to be relatively movable along the scale and detects an absolute position on the scale based on a pseudo-random code sequence on the scale. ,
When one code of the two-level code pattern is the code "1" or the code "0" and one of the two bits representing one code is L or H,
The detector is
Acquire the detection image of the scale and
The lower envelope and the upper envelope of the detected image are calculated, and the envelope is calculated.
The line between the lower envelope and the upper envelope is set as the threshold line.
The binarization process is performed by comparing the detected image with the threshold line.
The step of calculating the lower envelope and the upper envelope of the detection image is to move the sub-section having a predetermined width L from one end of the detection image to the other end of the sub-section. The minimum value of the light intensity is extracted from the data in the subsection to obtain the lower envelope, and the maximum value of the light intensity is extracted from the data in the sub-section to obtain the upper envelope .
A signal processing method for a photoelectric encoder. In the signal processing method of the photoelectric encoder according to claim 1,
The width of one bit is W / 2,
The code "1" is represented by the A pattern, which is a set of L and H.
The symbol "0" is represented by a B pattern, which is a set of L and L, and a C pattern, which is a set of H and H.
When the reference numerals "0" are continuous, the B pattern and the C pattern are arranged alternately.
A signal processing method for a photoelectric encoder, wherein the width L of the sub section is 1.5 W <L <2 W. In the signal processing method of the photoelectric encoder according to claim 1,
The width of one bit is W / 2,
The code "1" is represented by the A pattern, which is a set of L and H.
The symbol "0" is represented by a B pattern, which is a set of L and L, and a C pattern, which is a set of H and H.
When the reference numerals "0" are continuous, the B pattern and the C pattern are allowed to be continuous up to K, respectively.
A signal processing method for a photoelectric encoder, wherein the width L of the sub-section is (K + 0.5) W <L <(K + 1) W.

Images