JP7223661B2 - DATA PROCESSING DEVICE, DATA PROCESSING METHOD, PROGRAM AND ROTATION MEASURING DEVICE - Google Patents

Description

The present invention relates to a data processing device, a data processing method, a program thereof, and a rotation measuring device provided with the data processing device.

2. Description of the Related Art A measuring device is known that measures the rotation speed of a rotating body in real time based on changes in physical quantities (light, magnetism, etc.) accompanying rotation of the rotating body. For example, the rotation speed measuring device described in the following patent document selects a portion of an image of a rotating body captured by a camera by a user's operation, and information (luminance information) is subjected to Fourier transform (discrete Fourier transform) in the time-frequency domain, and the rotational speed is calculated from the frequency components obtained by the Fourier transform.

Patent No. 6521733

By the way, the discrete Fourier transform (hereinafter sometimes abbreviated as DFT) in the rotational speed measuring device of Patent Document 1 uses a data string consisting of a series of multiple pieces of luminance information. If the number (data length) of luminance information forming a data string is reduced, the DFT calculation error increases and the accuracy of the rotational speed decreases. Therefore, the data string used for DFT requires a certain amount of data length.

On the other hand, since luminance information is acquired one by one for each frame of video (moving image), when using a camera with a general frame rate, the amount of luminance information that can be acquired per unit time is about several tens. In this case, if the data length of one data string is about several hundred, the video period corresponding to one data string is about several seconds. If only one data string is acquired from one image period and DFT is performed, the DFT is performed only once every few seconds, so the rotation speed calculation result is also obtained only once every few seconds. .

One possible method for increasing the frequency of DFT execution is to overlap some luminance information in consecutive data strings. For example, if half of the luminance information in two consecutive data strings is overlapped, the period from the acquisition of one data string to the acquisition of the next data string is halved. In this case, the acquisition frequency of the data string is doubled, and the DFT execution frequency is also doubled.

The process of acquiring data strings such that some data (luminance information, etc.) in consecutive data strings overlap can be realized using, for example, a ring buffer as shown in FIG. 10A. The ring buffer shown in FIG. 10A has 16 memory areas from address A0 to address A15. Data sampled at regular time intervals are written to each memory area of the ring buffer in the order of addresses A0, A1, A2, . When one data is written to the last address A15, the next data is written to the first address A0.

FIG. 10B shows a state in which data D0-D11 are stored at addresses A0-A11. Assuming that 12 data D0 to D11 constitute one data string DS1, the data string DS1 can be read from addresses A0 to A11.

FIG. 10C shows a state in which eight pieces of data D12 to D19 have been written to the ring buffer from the state in FIG. 10B. As shown in FIG. 10C, 12 data D8-D19 written in addresses A12-A15 and A0-A3 form a data string DS2. can be read. In this case, four data D8 to D11 overlap in two continuous data strings DS1 and DS2.

By using the ring buffer in this way, it is possible to acquire data strings such that some data in consecutive data strings overlap. However, in the ring buffer, when the write destination address becomes the last address, the next write destination address returns to the beginning address, so the addresses are discontinuous in the series of data included in one data string. may become. For example, in the data string DS2 shown in FIG. 10C, the addresses A8 to A15 of the eight data D8 to D15 in the first half and the addresses A0 to A3 of the four data D16 to D19 in the latter half are discontinuous. In the arithmetic processing of DFT, sum-of-products are sequentially repeated in predetermined combinations of each data in the data string. Therefore, if the series of data addresses contained in the data string is discontinuous, processing such as conditional judgment is required when reading data from the ring buffer, and the overall computational load increases. profitable.

The present invention has been made in view of such circumstances, and its object is to obtain a data string periodically from time-series data and perform a predetermined calculation such as DFT, in which some data in a continuous data string To provide a data processing device, a data processing method, and a program capable of writing each data string in a continuous address range in a storage area even if there is overlap, and a rotation speed measuring device equipped with such a data processing device. is to provide

A first aspect of the present invention is a data processing device for processing time-series data in real time, comprising: a storage unit having a storage area for temporarily storing each data of the time-series data; and a processing unit. M pieces of data consecutive in the time-series data constitute one data string, and two consecutive data strings in the time-series data are N common to each other (N is more than M represents a small natural number), and the processing unit performs a process of sequentially acquiring each of the data of the time-series data, and a process of generating a write destination address of the acquired data in the storage area. a process of writing the acquired data to the generated address; a process of detecting completion of writing of each data string; performing a predetermined operation based on a data string, and generating the addresses by converting the M consecutive addresses corresponding to the M data constituting one data string into 1; One unit address range includes N first addresses corresponding to the first N data in the one data string and the second half (M (M−N) second addresses corresponding to the N) pieces of data, and the process of generating the addresses is performed after the one unit address range corresponding to the one data string. When the number of addresses is less than (M−N), the N consecutive addresses preceding the one unit address range are transferred to the next data string corresponding to the next data string following the one data string. The N first addresses in the unit address range.

According to this configuration, when the number of addresses after one unit address range corresponding to one data string is less than (MN), the storage area after the one unit address range , the (MN) second addresses in the next unit address range corresponding to the next data string following the one data string cannot be secured. In this case, the N consecutive addresses before the one unit address range are used as the N first addresses in the next unit address range. As a result, M consecutive addresses in the storage area are secured as the next unit address range. Further, since the N first addresses in the next unit address range are out of the one unit address range, even if the N new data are written to the N first addresses, The data already written in the one unit address range is not overwritten by the new data. Therefore, it is possible to write each of the data strings in a continuous address range (unit address range) of the storage area while overlapping the N pieces of data in the two consecutive data strings.

Preferably, when the number of addresses subsequent to one unit address range corresponding to one data string is (M−N) or more, the process of generating the addresses includes: The latter N addresses are set as the N first addresses in the next unit address range, and the (M−N) consecutive addresses following the one unit address range are set as the next unit address range. The (MN) second addresses in the unit address range.

According to this configuration, when the number of addresses after one unit address range corresponding to one data string is (MN) or more, the storage area after the one unit address range can secure the (MN) second addresses in the next unit address range corresponding to the next data string following the one data string. In this case, the latter N addresses in the one unit address range are set as the N first addresses in the next unit address range, and follow the one unit address range (MN). consecutive addresses are used as the (MN) second addresses in the next unit address range. As a result, the one unit address range and the next unit address range are secured overlapping each other in the storage area, so that the storage area can be used efficiently.

Preferably, in the process of generating the addresses, when the number of the addresses after one of the unit address ranges corresponding to one of the data strings is less than (MN), N The N addresses up to the th are set as the N first addresses in the next unit address range.

According to this configuration, when the number of addresses after one unit address range corresponding to one data string is less than (MN), the storage area after the one unit address range , the (MN) second addresses in the next unit address range corresponding to the next data string following the one data string cannot be secured. In this case, the N addresses from the top to the Nth address in the storage area are the N first addresses in the next unit address range. As a result, the range of the N first addresses in the next unit address range is maximally separated from the one unit address range. It is possible to lengthen the time until the column is overwritten by the other data, and it becomes easier to secure the processing time for the predetermined calculation.

Preferably, the process of generating the addresses is performed when the number of the addresses after one of the unit address ranges corresponding to one of the data strings being acquired is less than (MN), and , when generating the latter N addresses in the one unit address range, each time one data in the one data string is obtained, one address is generated from the latter N addresses and generating one address from the N first addresses in the next unit address range, and the number of the addresses after the one unit address range corresponding to the one data string being obtained. is (M−N) or more, each time one data in the one data string is acquired, one address corresponding to the one data is generated.

According to this configuration, when the number of addresses after one of the unit address ranges corresponding to one of the data strings being acquired is less than (MN), and the one unit When the latter N addresses in the address range are generated, the N first addresses in the next unit address range are also generated in parallel. That is, each time one data in the one data string is acquired, one address is generated from the latter N addresses, and the N first addresses in the next unit address range are generated. One said address is generated from the addresses. On the other hand, if the number of addresses after one of the unit address ranges corresponding to one of the data strings being acquired is (MN) or more, one of the data in the one data string is acquired. Each time, one address corresponding to the one data is generated. This limits the cases where the same data is written to different addresses in parallel.

Preferably, the storage area is {k×(MN)+N} or more and less than {(k+1)×(MN)+N} addresses (k is a natural number of 2 or more). In the initial state, the address generating process and the data writing process set the address indicated by the first variable to the top address of the storage area, and each time one piece of data is acquired. Then, the obtained one data is written to the address indicated by the first variable, and after the writing, the address indicated by the first variable is incremented, and each time one data is obtained, the first confirming whether the address indicated by the variable is the {k×(MN)+1}-th address of the storage area, and confirming whether the address indicated by the first variable is the {k×(MN)+1}-th address of the storage area; th address, the address indicated by the second variable is set as the leading address, and each time one of the data is obtained, the address indicated by the first variable is changed to {k× (M−N)+1}-th to {k×(M−N)+N}-th address range, and if the address indicated by the first variable is included in the address range, acquired The one data is written to the address indicated by the second variable, and after the writing, the address indicated by the second variable is incremented, and the address indicated by the first variable is {k×( MN)+N}th address, the address indicated by the first variable is set to the address indicated by the second variable.

According to this configuration, when the address indicated by the first variable is the {k×(M−N)+1}-th address, the N-th address from the address is the end of one of the unit address ranges. while the address that is Mth from the address deviates from the storage area. In this case, the (MN) second addresses in the next unit address range cannot be secured for the one unit address range in the storage area after the address. Therefore, when the address indicated by the first variable is included in the {k×(M−N)+1}-th to {k×(M−N)+N}-th address range of the storage area (the one is included in the unit address range), data is written to the next unit address range starting from the top address in parallel with data writing to the one unit address range.
When the address indicated by the first variable exceeds the {k×(MN)+N}-th address of the storage area, writing of the data string to the one unit address range is completed. In this case, by setting the address indicated by the second variable as the address indicated by the first variable, data writing to the next unit address range is continued.

Preferably, the storage area is composed of L addresses (L is an integer of 2×M or more), and the process of generating the addresses and the process of writing the data are performed in an initial state in a first The address interval between the address indicated by the variable and the address indicated by the second variable is set to M or more (LM) or less, and each time one of the data is obtained, the obtained one data is transferred to the write to the address indicated by the first variable and the address indicated by the second variable; after the writing, increment the address indicated by the first variable and the address indicated by the second variable; When the address indicated by one variable exceeds the maximum address of the storage area, the address indicated by the first variable is set to the top address of the storage area, and the address indicated by the second variable is set. exceeds the maximum address, the process of setting the address indicated by the second variable as the first address and detecting the completion of writing corresponds to one of the data strings by the process of generating the address. The process of specifying the one unit address range generated by the above based on the address indicated by the first variable and the address indicated by the second variable, and performing the predetermined operation includes: The data string used for the predetermined calculation is read out from the unit address range specified by the detection process.

According to this configuration, data writing to the address indicated by the first variable and data writing to the address indicated by the second variable are performed in parallel, and after the data writing, the first variable is The address indicated and the address indicated by the second variable are each incremented. Then, the address indicated by the first variable and the address indicated by the second variable are returned to the top address of the storage area after reaching the maximum address of the storage area.
The storage area is composed of L (≧M) addresses, and an address interval between the address indicated by the first variable and the address indicated by the second variable is M or more (L− M) or less, the address interval is maintained at M or more (LM) or less even after the initial state. As a result, at least one of the address indicated by the first variable and the address indicated by the second variable is incremented M times to generate M consecutive addresses (the unit address range).
The series of addresses obtained by incrementing the address of the first variable and the series of addresses obtained by incrementing the address of the second variable are in different address ranges in the storage area. Therefore, based on the address indicated by the first variable and the address indicated by the second variable (for example, from the magnitude relationship between the two), the series of addresses forming the unit address range is specified.

Preferably, said predetermined operation includes a discrete Fourier transform.
According to this configuration, since the data string is written in the unit address range consisting of the consecutive M addresses, the process for reading the data from the unit address range is performed at the top address of the unit address range. can always be expressed in the same way as an offset based on , which simplifies the process and reduces the load of arithmetic operations in the discrete Fourier transform.

A second aspect of the present invention is a data processing method in which one or more computers process time-series data in real time, wherein the one or more computers temporarily store each data of the time-series data. M pieces of continuous data in the time-series data form one data string, and two consecutive data strings in the time-series data share N common to each other. (N represents a natural number smaller than M) of the data, and the computer performs a process of sequentially acquiring each of the data of the time-series data, and a write destination of the acquired data in the storage area. A process of generating an address, a process of writing the acquired data to the generated address, a process of detecting completion of writing of each data string, and each time detection of completion of writing of one data string is performed. and the processing of performing a predetermined operation based on the one data string, and the processing of generating the address includes M consecutive data corresponding to the M data constituting one data string. The addresses are generated as one unit address range, and one unit address range includes N first addresses corresponding to the first N data in the one data string and the one data string. (M−N) second addresses corresponding to the latter (M−N) pieces of data, and the process of generating the addresses includes one of the unit addresses corresponding to one of the data strings If the number of addresses after the range is less than (MN), the N consecutive addresses before the one unit address range are transferred to the next data string following the one data string. The N first addresses in the corresponding next unit address range.

A third aspect of the present invention is a program that causes one or more computers to function as the processing unit in the data processing apparatus of the first aspect.

A fourth aspect of the present invention includes an imaging device for imaging a rotating body at a constant cycle, and the data processing device according to the first aspect, wherein the process of sequentially acquiring the data of the time-series data includes periodically acquiring numerical data corresponding to the values of pixels included in a specific region in the image of the body of revolution captured by the imaging device as the data of the time-series data; The rotation speed measuring device calculates the rotation speed of the rotating body based on the frequency component of the time-series data indicated by the result of the discrete Fourier transform performed as the predetermined calculation.

According to the present invention, when a data string is periodically acquired from time-series data and a predetermined calculation such as DFT is performed, even if some data overlap in the continuous data string, the continuous data in the storage area Each data column can be written to an address range.

FIG. 1 is a diagram showing an example of the configuration of a rotation measuring device according to this embodiment. FIG. 2 is a diagram illustrating an example of an operation of acquiring a data string from time-series data and executing a Fourier transform. FIG. 3 is a diagram illustrating an example of an operation of writing time-series data in a storage area in the rotation measuring device according to the first embodiment. FIG. 4 is a diagram illustrating an example of an operation of writing time-series data in a storage area in the rotation measuring device according to the first embodiment. FIG. 5 is a flowchart for explaining an example of processing of the rotation measuring device according to the first embodiment. FIG. 6 is a flowchart for explaining a first modified example of processing of the rotation measuring device according to the first embodiment. FIG. 7 is a flowchart for explaining a second modified example of processing of the rotation measuring device according to the first embodiment. FIG. 8 is a diagram illustrating an example of an operation of writing time-series data in a storage area in the rotation measuring device according to the second embodiment. FIG. 9 is a flowchart for explaining an example of processing of the rotation measuring device according to the second embodiment. 10A to 10C are diagrams illustrating an example of the operation of writing time-series data to the ring buffer.

<First Embodiment>
FIG. 1 is a diagram showing an example of the configuration of a rotation measuring device according to this embodiment. The rotation measurement device shown in FIG. and a data processing device 1 for calculating It should be noted that the rotational speed measurement target in this embodiment is not limited to the propeller 5 as shown in FIG.

The imaging device 2 directly or indirectly captures an image of at least a portion of the rotating body (propeller 5 or the like) and outputs the captured image data to the data processing device 1 . The imaging device 2 includes, for example, an imaging device such as a CMOS sensor or a CCD sensor.

Based on the imaging data input from the imaging device 2, the data processing device 1 acquires a series of data that changes with the rotation of the rotating body as time-series data, and transforms the time-series data into fast Fourier transform (hereinafter referred to as " The rotational speed of the rotating body is calculated by performing a DFT operation such as FFT.

The data processing apparatus 1 shown in FIG. 1 includes an input section 11, a display section 12, a storage section 13, and a processing section .

The input unit 11 is a device for inputting user's instructions to the processing unit 14 . For example, the input unit 11 includes one or more input devices such as a mouse, keyboard, touch pad, touch panel, push button, and rotary switch that generate signals according to user operations.

The display unit 12 is a device that displays an image based on the image data generated by the processing unit 14, and includes, for example, a liquid crystal display and an organic EL display.

The storage unit 13 stores a program 15 for causing a computer, which will be described later, of the processing unit 14 to execute processing. The storage unit 13 also stores constant data that is repeatedly used to execute the process, variable data that is temporarily saved during the process of executing the process, and the like. Specifically, the storage unit 13 has a storage area for temporarily storing each piece of time-series data based on the imaging data input from the imaging device 2 .

The program 15 stored in the storage unit 13 may be, for example, read from an external storage device (USB memory) connected to any input interface provided in the processing unit 14, or may be read from an external storage device (USB memory) connected to the processing unit 14. It may be read from a non-temporary recording medium (optical disc, etc.) by an arbitrary recording medium reading device, or may be downloaded from an external server device or the like via an arbitrary communication interface provided in the processing unit 14. .

The storage unit 13 includes one or more arbitrary storage devices such as DRAM, SRAM, ROM, flash memory, and hard disk.

The processing unit 14 is a device that controls the overall operation of the data processing device 1, and executes processing related to measurement of the rotation speed of the rotating body. The processing unit 14 includes one or more processing circuits (CPU, MPU, etc.) that execute processing according to instruction codes of one or more programs 15 stored in the storage unit 13 . When one or more processing circuits execute processing based on the program 15, the processing unit 14 operates as one or more computers. The processing unit 30 executes processing related to measurement of the rotation speed of the rotating body as one or more computers described above.

Note that the processing unit 14 may include one or more dedicated hardware (ASIC, FPGA, etc.) configured to implement specific functions. In this case, the processing unit 14 may execute at least part of the processing related to the measurement of the rotation speed of the rotating body in dedicated hardware.

The processing unit 14 executes detection range setting processing, time-series data acquisition processing, address generation processing, data write processing, write completion detection processing, arithmetic processing, and rotation speed calculation processing as processing related to rotation speed measurement.

(Detection range setting process)
The processing unit 14 sets a detection range used for measuring the rotational speed in the image of the rotating body captured by the imaging device 2 before starting to measure the rotational speed. For example, the processing unit 14 displays the image of the body of rotation captured by the imaging device 2 on the screen of the display unit 12 as a moving image, and detects the body of rotation from the image of the body of rotation on the screen in accordance with the user's instruction input through the input unit 11. Set range.

(Time-series data acquisition processing)
When starting to measure the rotation speed, the processing unit 14 sequentially acquires each piece of time-series data based on the imaging data input from the imaging device 2 . The processing unit 14 identifies one or more pixels included in the detection range from one frame of images in the imaging data, and based on the values of the identified one or more pixels, one of the time-series data. Acquire (sample) data. For example, the processing unit 14 calculates a numerical value indicating the average light intensity (luminance, etc.) in the detection range based on the values of one or more pixels included in the detection range, and converts this numerical value to 1 of the time-series data. obtained as one piece of data.

The processing unit 14 repeatedly acquires data strings from the time-series data through address generation processing, data write processing, and write completion detection processing, which will be described later. M consecutive data in the time-series data form one data string. "M" indicates the data length of the data string. Also, two continuous data strings in the time-series data include N pieces of data common to each other (N represents a natural number smaller than M). "N" indicates the number of data duplications in two consecutive data strings.

(address generation processing, data write processing)
When acquiring data in the time-series data acquisition process, the processing unit 14 generates an address to which the data is written in the storage area secured in the storage unit 13 . The processing unit 14 generates consecutive addresses in the storage area as a write destination for one data string. That is, the processing unit 14 generates M consecutive addresses corresponding to M data forming one data string. The processing unit 14 writes the data acquired in the time-series data acquisition process to the address generated in the above address generation process.

In the following description, M consecutive addresses to which one data string is written will be referred to as a "unit address range". Also, the first half N addresses in one unit address range are respectively called "first addresses", and the latter half (MN) addresses in one unit address range are respectively called "second addresses". call. The N first addresses are addresses corresponding to the first half N data in one data string, and the (MN) second addresses are the second half (MN) addresses in one data string. This is an address corresponding to individual data.

Further, in the following description, of two consecutive data strings arbitrarily selected in the time-series data, the first data string is called the "first data string", and the latter data string is called the "second data string". call. Also, here, the unit address range to which the first data string is written is referred to as the "first unit address range", and the unit address range to which the second data string is written is referred to as the "second unit address range". call. Furthermore, the N pieces of data that are common to each other in the first data string and the second data string are respectively called "common data", and the address to which the common data is written in the first data string is called "common data write address". .

As described above, the first data string and the second data string include N pieces of common data that are common to each other. The latter N addresses (second addresses) in the first unit address range are common data write addresses, and N common data are written therein. In addition, N pieces of common data are also written to N pieces of first addresses in the second unit address range.

The processing unit 14 changes the position of the second unit address range in the storage area according to the number of addresses remaining after the first unit address range in the storage area (hereinafter sometimes referred to as "number of remaining addresses"). Let

When the number of remaining addresses is greater than (MN), (MN) second addresses in the second unit address range can be secured after the first unit address range in the storage area. In this case, the processing unit 14 sets the range overlapping the first unit address range in the storage area as the second unit address range. That is, when the number of remaining addresses is (M−N) or more, the processing unit 14 converts the N common data write addresses in the first unit address range (the latter N addresses in the first unit address range) to Assume that there are N first addresses in the second unit address range. In addition, the processing unit 14 treats (MN) consecutive addresses following the first unit address range as (MN) second addresses in the second unit address range.

If the number of remaining addresses is greater than (MN), the processing unit 14 generates one address each time it acquires one piece of data. That is, in this case, the processing unit 14 generates one address (address in the first unit address range) corresponding to the one piece of data each time it acquires one piece of data while acquiring the first data string. .

On the other hand, if the number of remaining addresses is less than (MN), (MN) second addresses in the second unit address range cannot be secured after the first unit address range in the storage area. In this case, the processing unit 14 sets the N consecutive addresses preceding the first unit address range in the storage area as the N first addresses in the second unit address range. For example, the processing unit 14 sets the N addresses from the top to the Nth in the storage area as the N first addresses in the second unit address range.

When N pieces of data are written to N first addresses in the second unit address range, writing of the first data string is completed in the first unit address range. At this time, since the N first addresses in the second unit address range are addresses before the first unit address range, the data written in the second unit address range is different from each data in the first unit address range ( first data column) is not overwritten. Therefore, the first data string for which writing has been completed can be normally read in the FFT operation processing described later.

When the number of remaining addresses is less than (MN) and when generating the latter N addresses (common data write addresses) in the first unit address range, the processing unit 14 generates two addresses are generated in parallel. That is, in this case, the processing unit 14 generates one address from the last N addresses in the first unit address range each time it acquires one piece of data in the middle of acquiring the first data string, An address is generated from the N first addresses in the two-unit address range.

Here, the storage area reserved for temporary storage of time-series data is {k×(M−N)+N} or more and less than {(k+1)×(M−N)+N} addresses (k represents a natural number of 2 or more). Also, the {k×(MN)+1}th address from the top of the storage area is called a “first boundary address At1”, and the {k×(MN)+N}th address from the top of the storage area is called “first boundary address At1”. shall be referred to as a "second boundary address At2".

The first boundary address At1 ({k×(MN)+1}-th address from the beginning) is the unit address range closest to the end of the storage area (hereinafter sometimes referred to as the “end side unit address range”). )include. The N-th address counted from the first boundary address At1 is the last address in the last side unit address range. When the end-side unit address range is regarded as the first unit address range, the first boundary address At1 is the first address among the N common data write addresses.

The second boundary address At2 ({k×(M−N)+N}th address from the beginning) is the Nth address counted from the first boundary address At1, and is the last address in the end side unit address range. .

For example, the processing unit 14 uses two variables (a first variable W1 and a second variable W2) each indicating a write destination address to perform an address generation process and a data write process.
First, in the initial state, the processing unit 14 sets the address indicated by the first variable W1 to the top address of the storage area (hereinafter sometimes referred to as "top address Ah").
Each time one piece of data is acquired in the time-series data acquisition process, the processing unit 14 writes the acquired piece of data to the address indicated by the first variable W1, and after the writing, increments the address indicated by the first variable W1. do.
Further, the processing unit 14 checks whether the address indicated by the first variable W1 is the first boundary address At1 each time one piece of data is obtained in the time-series data obtaining process, and the address indicated by the first variable W1 is the first boundary address At1. If it is 1 boundary address At1, the address indicated by the second variable W2 is set as the top address Ah.
Furthermore, the processing unit 14 checks whether the address indicated by the first variable W1 is included in the address range from the first boundary address At1 to the second boundary address At2 each time one piece of data is acquired in the time-series data acquisition process. do. If the address indicated by the first variable W1 is included in this address range (At1 to At2), the processing unit 14 merges the acquired data into the address indicated by the first variable W1 and the address indicated by the second variable W2. After each write, the address indicated by the first variable W1 and the address indicated by the second variable W2 are incremented.
As a result of incrementing the address indicated by the first variable W1, if the address indicated by the first variable W1 exceeds the second boundary address At2, the processing unit 14 changes the address indicated by the first variable W1 to the address indicated by the second variable W2. set to address.

(Write completion detection processing)
The processing unit 14 detects completion of writing each data string to the unit address range in the process of the address generation process and the data write process described above. For example, when starting to write data from the top address Ah of the storage area, the processing unit 14 starts counting the number of times data is written to the storage area, and based on the counted value, the writing of each data string is completed. to detect

Further, when detecting that the writing of one data string is completed, the processing unit 14 specifies the unit address range in which the one data string is written. For example, the processing unit 14 stores the start address of the unit address range in a predetermined variable, and each time it detects that writing of one data string is completed, it changes the predetermined address so as to indicate the start address of the next unit address range. update the variables in

The first variable W1 described above indicates the latest write destination address, and based on this write destination address, it is possible to grasp to which address in the unit address range the data has been written. be. Therefore, the processing unit 14 may detect completion of writing of the data string by detecting writing of data at the last address of the unit address range based on the address indicated by the first variable W1. . In this case, based on the address indicated by the first variable W1 when the writing of one data string is completed, the processing unit 14 determines the unit address range (for example, the first address in the unit address range) where the one data string is written. ) may be specified.

(FFT calculation processing, rotation speed calculation processing)
The processing unit 14 performs an FFT operation based on one data string each time it detects completion of writing of one data string in the write completion detection process. The processing unit 14 reads a data string to be used for FFT calculation from the unit address range specified in the write completion detection process. When the result of the FFT calculation is obtained, the processing unit 14 calculates the rotation speed of the rotating body based on the frequency component of the time-series data indicated by this calculation result. For example, in the propeller 5 as shown in FIG. 1, when the position where five symmetrical blades 51 pass is set as the detection range, the processing unit 14 detects the frequency of the maximum peak in the frequency component obtained by the FFT calculation. The one-fifth frequency is calculated as the rotational speed of the propeller 5 (the number of revolutions per unit time).

Next, the operation of the rotation speed detection device according to the present embodiment described above will be described.

FIG. 2 is a diagram illustrating an example of an operation of acquiring a data string from time-series data and executing FFT. "D(1)", "D(2)", . . . in FIG. 2 indicate individual data in the time-series data. In the following description, individual data (D(1), D(2), . . . ) may be referred to as "data D" without distinction. Also, “DS1”, “DS2”, . . . in FIG. 2 indicate individual data strings in the time-series data. In the following description, each data string (DS1, DS2, . . . ) may be referred to as "data string DS" without distinction. In the example of FIG. 2, the data length M of the data string DS is "8", and the number of data overlaps N in two consecutive data strings DS is "2". As shown in FIG. 2, the processing unit 14 of the data processing device 1 performs FFT calculation each time it acquires one data string DS from the time-series data.

FIG. 3 is a diagram illustrating a first example of the operation of writing the time-series data (M=8, N=2) shown in FIG. 2 to the storage area in the rotation speed measuring device according to the first embodiment. . In the example of FIG. 3, the size of the storage area (the number of assigned addresses) is "16". "R1", "R2", . . . in FIG. In the following description, each unit address range (R1, R2, . . . ) may be referred to as "unit address range R" without distinction.

In the initial state, the head address Ah of the storage area is set in the first variable W1. Each time new data D of the time-series data is acquired, the acquired data D is written to the address indicated by the first variable W1, and after the data D is written, the address of the first variable W1 is incremented. When the data D(8) is written to the address (Ah+7), the writing of the data string DS1 (D(1) to D(8)) to the unit address range R1 (Ah to Ah+7) is completed. At this time, the variable Ab indicates the head address Ah of the unit address range R1, and the unit address range R1 is specified by the variable Ab. When the writing of the data string DS1 is completed, the data string DS1 is read from the unit address range R1 specified by the variable Ab (address Ah), and the FFT operation is performed.

The unit address range R2 (Ah+6 to Ah+13) of the next data string DS2 following the data string DS overlaps the unit address range R1 as shown in FIG. The unit address range R2 is the end side unit address range closest to the end address (Ah+15) of the storage area. When the data D(12) is written to the address (Ah+12) in the unit address range R2, the first variable W1 indicates the address (Ah+12). This address (Ah+12) is the first boundary address At1. That is, when the data length M is "8", the data duplication number N is "2", and the storage area size is "16=2×(MN)+4", the first boundary address At1 is It is the {2×(MN)+1}th address and matches the address (Ah+12). When the address indicated by the first variable W1 is the first boundary address At1, the top address Ah is set to the second variable W2.

The data D(13) obtained following the data D(12) are written to the address (Ah+12) indicated by the first variable W1 and the leading address Ah indicated by the second variable W2. After this writing, when the first variable W1 and the second variable W2 are incremented, the second variable W2 indicates the address (Ah+1) and the first variable W1 indicates the address (Ah+13). Here, the address (Ah+13) indicated by the first variable W1 is the second boundary address At2. That is, when the data length M is "8", the data duplication number N is "2", and the storage area size is "16=2×(MN)+4", the second boundary address At2 is It is the {2×(MN)+N}-th address and matches the address (Ah+13).

Since the address (Ah+13) indicated by the first variable W1 is included in the address range from the first boundary address At1 to the second boundary address At2, the data D(14) acquired following the data D(13) is It is written to the address (Ah+13) indicated by the first variable W1 and the address (Ah+1) indicated by the second variable W2. This completes the writing of the data string DS2 to the unit address range R2. At this time, since the variable Ab indicates the top address (Ah+6) of the unit address range R2, the data string DS2 is read out from the unit address range R2 specified by the variable Ab, and the FFT operation is performed.

When the first variable W1 and the second variable W2 are incremented after writing the data D(14), the second variable W2 indicates the address (Ah+2) and the first variable W1 indicates the address (Ah+14). Here, since the address (Ah+14) indicated by the first variable W1 exceeds the second boundary address At2, the address (Ah+2) of the second variable W2 is set to the first variable W1. As a result, data D(15), D(16), . . . following data D(14) are written to addresses (Ah+2), (Ah+3), . The unit address range R3 is shifted to the top address Ah side with respect to the unit address range R2, and the data D is not written to the addresses (Ah+14, Ah+15) after the unit address range R2.

FIG. 4 is a diagram illustrating a second example of the operation of writing the time series data (M=8, N=2) shown in FIG. 2 in the storage area in the rotation speed measuring device according to the present embodiment. In the example of FIG. 4, the size of the storage area is "27", which is larger than the example of FIG. When the data length M is "8", the data duplication number N is "2", and the storage area size is "27=4×(MN)+3", as shown in FIG. 4, the first boundary address At1 is The address is (Ah+24) and the second boundary address At2 is the address (Ah+25).

In the example of FIG. 4, four unit address ranges (R1 to R4) are assigned consecutively from the head side to the tail side of the storage area. The fourth unit address range R4 from the head side is the tail side unit address range. The first boundary address At1 and the second boundary address At2 are included in the unit address range R4, which is the end side unit address range. When the first variable W1 indicates the address (Ah+24), the address (Ah+24) is the first boundary address, so the top address Ah is set to the second variable W2. Then, when the address indicated by the first variable W1 is included in the address range (Ah+24 to Ah+25) from the first boundary address At1 to the second boundary address At2, the addresses indicated by the first variable W1 and the second variable W2 are Data D is written. When the address indicated by the first variable W1 is greater than the second boundary address At2 (address (Ah+25)), the first variable W1 is set to the address (Ah+2) of the second variable W2. As a result, as shown in FIG. 4, the unit address range R5 is shifted toward the top address Ah, and the data D is not written to the address (Ah+26) after the unit address range R4.

FIG. 5 is a flowchart for explaining an example of processing of the rotation measuring device according to the present embodiment, and shows processing when measuring the rotation speed of a rotating body in real time.

The processing unit 14 performs initial setting in an initial state before starting measurement of the rotational speed (ST100). In the initial setting, the processing unit 14 sets the size of the storage area to {k×(M−N)+N} or more and less than {(k+1)×(M−N)+N} (k is a natural number of 2 or more). ), and sets the first variable W1 to the head address Ah of the storage area. The processing unit 14 also sets the variable Bt indicating the number of times data D is written to the storage area to zero, and sets the variable Ab indicating the top address of the unit address range R to "Ah-(MN)". .

Furthermore, the processing unit 14 sets the first boundary address At1 to the address {Ah+k×(M−N)}, and sets the second boundary address At2 to the address {Ah+k×(M−N)+(N−1)}. do. The first boundary address At1 is the {k×(M−N)+1}-th address counting from the top address Ah, and the second boundary address At2 is {k×(M−N) counting from the top address Ah. +N}th address.

The processing unit 14 acquires one piece of data D based on the imaging data of the imaging device 2 (ST105), and writes the data D to the address indicated by the first variable W1 (ST110). The processing unit 14 also checks whether the first variable W1 indicates the first boundary address At1 (ST115), and if the first variable W1 indicates the first boundary address At1 (Yes in ST115), the second variable W2 A leading address Ah is set (ST120). Further, the processing unit 14 compares the first variable W1 and the second variable W2 (ST125), and if the addresses of both do not match (No in ST125), writes the data D to the address indicated by the second variable W2 (ST130). ).

After writing the data D to the storage area, the processing unit 14 increments the first variable W1 and the second variable W2, respectively, and also increments the variable Bt indicating the number of times the data D has been written (ST135).

The processing unit 14 checks whether the incremented first variable W1 indicates the second boundary address At2 (ST140), and if the address indicated by the first variable W1 matches the second boundary address At2 (Yes in ST140), The address indicated by the first variable W1 is set to the address indicated by the second variable W2 (ST145). In this case, the processing unit 14 sets the variable Ab indicating the top address of the unit address range R to "Ah-(MN)", which is the same as the initial setting (ST100).

Next, the processing section 14 confirms whether the variable Bt is equal to or greater than the data length M (ST150). If the variable Bt has not reached the data length M (No in ST150), the processing unit 14 returns to step ST105 and repeats the processes after step ST105. If the variable Bt is greater than or equal to the data length M (Yes in ST150), the processing section 14 proceeds to the next step ST155.

In step ST155, the processing unit 14 confirms whether the difference (Bt-N) between the variable Bt and the number of data overlaps N is divisible by the difference (MN) between the data length M and the number of data overlaps N. If the difference (Bt−N) is not divisible by the difference (MN) (No in ST155), the processing unit 14 returns to step ST105 and repeats the processing after step ST105. If the difference (Bt−N) is divisible by the difference (MN) (Yes in ST155), the processing unit 14 detects that the writing of the data string DS has been completed by the most recent writing of the data D (ST110), The process proceeds to the next step ST160.

The processing unit 14 adds “MN” to the variable Ab indicating the top address of the unit address range R in step ST160. Since the unit address range R is shifted by "MN" to the end of the storage area, the variable Ab obtained by adding "MN" is the unit address range R of the data string DS for which the completion of writing was detected in step ST155 indicates the beginning of the

Next, the processing unit 14 performs an FFT operation based on the data string DS of the unit address range R specified by the variable Ab. The processing unit 14 also calculates the rotation speed of the rotating body based on the frequency component of the time-series data obtained by the FFT calculation (ST165). The processing unit 14 outputs information on the calculated rotational speed to the screen of the display unit 12 or the like (ST170). After that, the processing unit 14 returns to step ST105 and repeats the processes after step ST105.

As described above, according to the rotation speed measuring device according to the present embodiment, the number of addresses (remaining address) is less than (M−N), the second unit address range to which the second data string (for example, the data string DS5 in FIG. 4) is written is stored in the storage area after the first unit address range. (M−N) of second addresses cannot be reserved. In this case, N consecutive addresses before the first unit address range in the storage area are set as N first addresses in the second unit address range (for example, addresses Ah and (Ah+1)). As a result, M consecutive addresses of the storage area are secured as the second unit address range. Also, since the N first addresses in the second unit address range are out of the first unit address range, N new data D are written to the N first addresses in the second unit address range. Also, the new data D does not overwrite the data D already written in the first unit address range. Therefore, the first data string and the second data string are written in consecutive address ranges (unit address ranges) of the storage area while overlapping the N pieces of data D in the consecutive first data string and the second data string. can be done.

Further, according to the rotation speed measuring device according to the present embodiment, the number of addresses (number of remaining addresses) after the first unit address range to which the first data string (for example, the data string DS1 in FIG. 4) is written is In the case of (M−N) or more, in the storage area after the first unit address range, the second data string following the first data string (for example, the data string DS2 in FIG. 4) is written. (MN) second addresses in a two-unit address range can be reserved. In this case, the last N addresses in the second unit address range are used as the N first addresses in the second unit address range (for example, the addresses (Ah+6) and (Ah+7) in the unit address range R2 in FIG. 4). , (M−N) consecutive addresses following the first unit address range are set as (MN) second addresses in the second unit address range (for example, in the unit address range R2 in FIG. 4). addresses (Ah+8) to (Ah+13)). As a result, the first unit address range and the second unit address range are secured while overlapping each other in the storage area, so that the storage area can be used efficiently.

Further, according to the rotation speed measuring device according to the present embodiment, the number of addresses (number of remaining addresses) after the first unit address range (for example, unit address range R4 in FIG. 4) is less than (MN). In this case, the N addresses from the top address Ah to the Nth address in the storage area are set as the N first addresses in the second unit address range (for example, the addresses Ah and (Ah+1) in the unit address range R5 in FIG. ). As a result, since the range of N first addresses in the second unit address range is separated from the first unit address range as much as possible, the first data string that has been written to the first unit address range is separated. It is possible to lengthen the time until the data D is overwritten. For example, in the case of FIG. 4, when the data D(26) is written to the address (Ah+1) in the unit address range R5, the writing of the data string DS4 in the unit address range R4 is completed. After the writing of the data string DS4 is completed, the data string DS4 in the unit address range R4 is newly written until the 16 data D(27) to D(42) are written to the addresses (Ah+2) to (Ah+17). Data D is not overwritten. As a result, the period during which the written data string is stored in the storage area without being overwritten becomes longer, and it becomes easier to secure the processing time for computation such as FFT.

Further, according to the rotation speed measuring device according to the present embodiment, the number of addresses after the first unit address range (remaining address number) is less than (M−N), and when generating the latter N addresses in the first unit address range (for example, addresses (Ah+24) and (Ah+25)), and N first addresses in the second unit address range are also generated in parallel (for example, addresses Ah and (Ah+1) in the unit address range R5 in FIG. 4). In this case, each time one piece of data D in the first data string is obtained, one address is generated from the latter N addresses in the first unit address range, and N pieces of data in the second unit address range are generated. An address is generated from the first address. On the other hand, when the number of remaining addresses after the first unit address range is (M−N) or more, each time one data D in the first data string is obtained, one data D corresponding to the one data D is obtained. An address is generated. This limits the cases where the same data D is written to different addresses in parallel, so that the load of the write operation on the storage unit 13 can be reduced.

Next, a modified example of the processing of the rotational speed measuring device according to the present embodiment will be described with reference to the flowcharts of FIGS. 6 and 7. FIG.

FIG. 6 is a diagram showing a first modified example of processing of the rotation speed measuring device according to the present embodiment. Steps ST100 to ST160 in the flowchart shown in FIG. 6 are the same as those in the flowchart shown in FIG. In the first modification shown in FIG. 6, the processing (computation processing, output processing) of steps ST165 and ST170 in the flowchart shown in FIG. ) in parallel. That is, after updating the variable Ab in step ST160, the processing unit 14 asynchronously calls and executes the processing of the sub-thread (ST175). In the secondary thread, the processing unit 14 performs an FFT operation based on the data sequence DS of the unit address range R specified by the variable Ab, and calculates the rotation speed based on the frequency component of the time-series data obtained by the FFT operation. (ST180). The processing unit 14 outputs information on the calculated rotational speed to the screen of the display unit 12 (ST185).

According to the first modified example described above, a thread that acquires time-series data based on imaging data and writes a data string DS in the unit address range R of the storage area, reads the data string DS from the unit address range R, performs FFT, etc. are executed in parallel. Therefore, it becomes easier to secure processing time for arithmetic processing (such as FFT). Moreover, even if the arithmetic processing (such as FFT) requires a relatively long time, it is possible to obtain the result of the arithmetic processing in real time without interfering with the cycle of acquiring the data string DS from the time-series data.

In the example of FIG. 6, there may be two or more secondary threads that execute the processes of steps ST180 and ST185. As a result, a plurality of arithmetic processes (such as FFT) can be executed in parallel, making it easier to secure processing time for the arithmetic processes (such as FFT).

FIG. 7 is a diagram showing a second modified example of processing of the rotation speed measuring device according to the present embodiment. The flowchart shown in FIG. 6 is obtained by replacing step ST125 in the flowchart shown in FIG. 6 with step ST125A, and other steps are the same as the flowchart shown in FIG.

In step ST125 of the flow charts shown in FIGS. 5 and 6, if the address indicated by the first variable W1 and the address indicated by the second variable W2 do not match (No in ST125), data is stored in the address indicated by the second variable W2. D is written (ST130). The state where the address indicated by the first variable W1 and the address indicated by the second variable W2 do not match starts when the second variable W2 is set to the head address Ah in step ST120, and the address of the first variable W1 is set in step ST145. is set to the address of the second variable W2. Therefore, this state is the same as the state where the address of the first variable W1 is greater than or equal to the first boundary address At1 and less than or equal to the second boundary address At2. Therefore, in the second modification shown in the flowchart of FIG. 7, in step ST125A that replaces step ST125, it is determined whether the address of the first variable W1 is greater than or equal to the first boundary address At1 and less than or equal to the second boundary address At2. If the address of the first variable W1 is greater than or equal to the first boundary address At1 and less than or equal to the second boundary address At2 (Yes in ST125A), data D is written to the address indicated by the second variable W2.

<Second embodiment>
Next, a rotational speed measuring device according to a second embodiment of the present invention will be described. The rotation speed measurement device according to the second embodiment is obtained by changing the processing of the processing unit 14 in the rotation speed measurement device according to the first embodiment, and the overall configuration is the rotation speed measurement device according to the first embodiment. It is almost the same as the device. Data processing in the processing unit 14 will be mainly described below.

The processing unit 14 performs, as in the first embodiment, detection range setting processing, time-series data acquisition processing, address generation processing, data writing processing, write completion detection processing, FFT calculation processing, Then, rotational speed calculation processing is executed. Of these, the detection range setting process, the time-series data acquisition process, the FFT calculation process, and the rotation speed calculation process are substantially the same as those already described in the first embodiment, and thus descriptions thereof will be omitted.

(address generation processing, data write processing)
The processing unit 14 uses two variables (a first variable W1 and a second variable W2) each indicating a write destination address to perform an address generation process and a data write process. Assuming that the size of the storage area (the number of addresses) is "L", the processing unit 14 sets the size L of the storage area to at least twice the data length M of the data string DS (L≧2×M). If the interval between the address indicated by the first variable W1 and the address indicated by the second variable W2 is "G", the processing unit 14 sets the interval G to be equal to or greater than the data length M and equal to or less than (LM). (M≦G≦LM).

First, in the initial state, the processing unit 14 sets the address interval G to "M≦G≦LM". Each time one piece of data is acquired in the time-series data acquisition process, the processing unit 14 writes the acquired piece of data to the address indicated by the first variable W1 and the address indicated by the second variable W2. , the address indicated by the first variable W1 and the address indicated by the second variable W2 are each incremented. Then, when the address indicated by the first variable W1 exceeds the maximum address of the storage area, the address indicated by the first variable W1 is set as the top address of the storage area. Also, when the address indicated by the second variable W2 exceeds the maximum address of the storage area, the address indicated by the second variable W2 is set as the top address of the storage area.

(Write completion detection processing)
The processing unit 14 detects completion of writing each data string to the unit address range in the process of the address generation process and the data write process described above. For example, similar to the first embodiment, the processing unit 14 detects completion of writing of each data string based on the count value obtained by counting the number of times data is written to the storage area.

Further, when the processing unit 14 detects that writing of one data string is completed, the one data string has been written based on the address indicated by the first variable W1 and the address indicated by the second variable W2. Identify the unit address range.

By the address generating process and the data writing process described above, the same data is stored in the storage area at two locations, the address indicated by the first variable W1 and the address indicated by the second variable W2. Therefore, based on the first variable W1 and the second variable W2 when the completion of writing of the data string is detected, it is possible to specify two address ranges of the storage area in which the same data string is written. . Also, by comparing the leading addresses or trailing addresses in these two address ranges, a unit address range in which all addresses are continuous (not including the trailing end and leading end of the storage area) can be specified. is possible. That is, of the two address ranges, all the addresses are continuous in the address range whose top address is close to the top of the storage area. Further, of the two address ranges, all addresses are continuous in the address range whose end address is close to the end of the storage area. Therefore, based on the first variable W1 and the second variable W2 when the writing of one data string is completed (for example, by comparing the magnitude relationship of the first variable W1 and the second variable W2), all addresses are It is possible to specify a continuous unit address range.

FIG. 8 is a diagram illustrating an example of the operation of writing the time-series data (M=8, N=2) shown in FIG. 2 to the storage area in the rotation measuring device according to the second embodiment. In the example of FIG. 8, the size of the storage area (the number of assigned addresses) is "16".

In the initial state, the first variable W1 is set to the head address Ah of the storage area, and the second variable W2 is set to the address (Ah+8). The interval G between the address indicated by the first variable W1 and the address indicated by the second variable W2 is "8". Each time new data D of the time-series data is acquired, the data D is written to the address indicated by the first variable W1 and the address indicated by the second variable W2. After the data D is written, the first variable W1 and The address of the second variable W2 is incremented.

When data D(8) is written, data sequence DS1 is written in the range of addresses Ah to (Ah+7) and the range of addresses (Ah+8) to (Ah+15). The first variable W1 indicates the leading address (Ah+8) in one address range (Ah+8 to Ah+15), and the second variable W2 indicates the leading address Ah in the other address range (Ah to Ah+7).
When the data D(14) is written, the data string DS2 is written in the range of addresses (Ah+6) to (Ah+13) and the range of addresses (Ah+14), (Ah+15), and Ah to (Ah+5). . The first variable W1 indicates the first address (Ah+14) in one address range (Ah+14, Ah+15, Ah to Ah+5), and the second variable W2 indicates the first address (Ah+6) in the other address range (Ah+6 to Ah+13). ).
Thus, the first variable W1 and the second variable W2 indicate the top address of the two address ranges in which the same data D is written. Therefore, when the writing of one data string DS is completed, the one data string DS is written to the address indicated by the first variable W1 or the address indicated by the second variable W2, whichever is smaller. can be specified as the top address of the unit address range.

FIG. 9 is a flowchart for explaining an example of processing of the rotation measuring device according to the second embodiment, and shows processing when measuring the rotational speed of a rotating body in real time.

The processing unit 14 performs initial setting in an initial state before starting measurement of the rotational speed (ST200). In the initial setting, the processing unit 14 sets the size L of the storage area to two times or more of the data length M (L≧2×M), sets the first variable W1 to the top address Ah of the storage area, and sets the second Set variable W2 to address (Ah+M). The interval G between the address indicated by the first variable W1 and the address indicated by the second variable W2 is "M". The processing unit 14 also sets a variable Bt indicating the number of times data D is written to the storage area to zero.

When the processing unit 14 acquires one data D based on the imaging data of the imaging device 2 (ST205), the processing unit 14 writes the data D to the address indicated by the first variable W1 and the address indicated by the second variable W2 (ST210). After writing the data D to the storage area, the processing section 14 increments the first variable W1 and the second variable W2, respectively, and also increments the variable Bt indicating the number of times the data D is written (ST215).

If the incremented first variable W1 indicates an address (Ah+L) that deviates from the storage area (Yes in ST220), the processing unit 14 sets the first variable W1 to the head address Ah of the storage area (ST225). If the incremented second variable W2 indicates an address (Ah+L) that deviates from the storage area (Yes in ST230), the processing unit 14 sets the second variable W2 to the head address Ah of the storage area (ST235).

Next, the processing section 14 confirms whether the variable Bt is equal to or greater than the data length M (ST240). If the variable Bt has not reached the data length M (No in ST240), the processing unit 14 returns to step ST205, and repeats the processes after step ST205. If the variable Bt is greater than or equal to the data length M (Yes in ST240), the processing section 14 proceeds to the next step ST245.

In step ST245, the processing unit 14 confirms whether the difference (Bt-N) between the variable Bt and the number of data overlaps N is divisible by the difference (MN) between the data length M and the number of data overlaps N. If the difference (Bt−N) is not divisible by the difference (MN) (No in ST245), the processing unit 14 returns to step ST205 and repeats the processes after step ST205. If the difference (Bt−N) is divisible by the difference (MN) (Yes in ST245), the processing unit 14 detects that the writing of the data string DS has been completed by the most recent writing of the data D (ST210), The process proceeds to the next step ST250.

In step ST250, the processing unit 14 sets the smaller one of the address indicated by the first variable W1 and the address indicated by the second variable W2 to the variable Ab. A variable Ab indicates the top address in the unit address range R of the data string DS for which writing has been completed.

The processing unit 14 performs an FFT operation based on the data sequence DS of the unit address range R specified by the variable Ab. The processing unit 14 also calculates the rotation speed of the rotating body based on the frequency component of the time-series data obtained by the FFT calculation (ST255). The processing unit 14 outputs information on the calculated rotational speed to the screen of the display unit 12 or the like (ST260). After that, the processing section 14 returns to step ST205 and repeats the processes after step ST205.

As described above, according to the rotation speed measuring device according to the second embodiment, the number of addresses after the first unit address range to which the first data string (for example, the data string DS2 in FIG. 8) is written When the (number of remaining addresses) is less than (MN), the storage area after the first unit address range contains the second unit to which the second data string (for example, the data string DS3 in FIG. 8) is written. (MN) second addresses in the address range cannot be reserved. In this case, the N consecutive addresses before the first unit address range in the storage area are set as the N first addresses in the second unit address range (for example, the address (Ah+4) in the unit address range R3 in FIG. ) and (Ah+5)). As a result, M consecutive addresses of the storage area are secured as the second unit address range. Also, since the N first addresses in the second unit address range are out of the first unit address range, N new data D are written to the N first addresses in the second unit address range. Also, the new data D does not overwrite the data D already written in the first unit address range. Therefore, the first data string and the second data string are written in consecutive address ranges (unit address ranges) of the storage area while overlapping N pieces of data D in the consecutive first data string and second data string. can be done.

Although several embodiments of the present invention have been described above, the present invention is not limited to the above-described embodiments and includes various variations.

For example, in the above description of the embodiment, an example of a rotational speed measuring device was given, but the present invention is not limited to this example. That is, the present invention is widely applicable to various data processing devices that process time-series data in real time, and various devices such as measurement devices equipped with data processing devices.

In addition, in the description of the above-described embodiment, an example of performing FFT calculation based on a data string obtained from time-series data was given, but various calculation processing (matrix calculation, statistical processing, etc.) not limited to FFT is performed. good too.

DESCRIPTION OF SYMBOLS 1... Data processing apparatus 11... Input part 12... Display part 13... Storage part 14... Processing part 15... Program 2... Imaging device 5... Propeller 51... Blade DS1-DS5... Data string, D(1) to D(32)... data, R1 to R5... unit address range, W1... first variable, W2... second variable, Ab, Bt... variable, Ah... top address, At1... first boundary address, At2... second boundary address, M... data length, N... number of data duplications

Claims (10)

A data processing device that processes time-series data in real time,
a storage unit having a storage area for temporarily storing each data of the time-series data;
a processing unit;
M consecutive data in the time-series data constitute one data string,
Two of the data strings that are continuous in the time-series data include N pieces of the data that are common to each other (N is a natural number smaller than M),
The processing unit is
A process of sequentially acquiring each of the data of the time-series data;
a process of generating an address to which the acquired data is written in the storage area;
a process of writing the obtained data to the generated address;
a process of detecting completion of writing of each data string;
each time the completion of writing of one data string is detected, a process of performing a predetermined calculation based on the one data string;
The process of generating the addresses includes generating M consecutive addresses corresponding to the M data constituting one data string as one unit address range,
One unit address range is
N first addresses corresponding to the first half N data in the one data string;
and (MN) second addresses corresponding to the latter (MN) data in the one data string,
When the number of addresses after one unit address range corresponding to one data string is less than (M−N), the process of generating the addresses includes N the consecutive addresses are the N first addresses in the next unit address range corresponding to the next data string following the one data string;
Data processing equipment. When the number of addresses subsequent to one unit address range corresponding to one data string is (M−N) or more, the processing for generating the addresses includes the following N of said addresses are said to be said N first addresses in said next unit address range, and (MN) said consecutive addresses following said one unit address range are said to be said next unit address range. Let the (M−N) second addresses in
2. A data processing apparatus according to claim 1. When the number of addresses after one of the unit address ranges corresponding to one of the data strings is less than (M−N), the process of generating the addresses is performed on the Nth address from the beginning in the storage area. Let the N addresses be the N first addresses in the next unit address range;
3. A data processing apparatus according to claim 1 or 2. The process of generating the address includes:
A case where the number of addresses after one of the unit address ranges corresponding to one of the data strings being acquired is less than (M−N), and the latter half of the one unit address range in the case of generating the addresses, each time one of the data in the one data string is obtained, one of the addresses is generated from the latter N addresses, and the address in the next unit address range is generated generating one said address from N first addresses;
When the number of addresses after one of the unit address ranges corresponding to one of the data strings being acquired is (M−N) or more, each time one of the data in the one data string is acquired , generating one said address corresponding to said one data;
4. A data processing apparatus according to claim 3. The storage area is composed of {k×(MN)+N} or more and less than {(k+1)×(MN)+N} addresses (where k is a natural number of 2 or more),
The process of generating the address and the process of writing the data include:
In an initial state, the address indicated by the first variable is set to the top address of the storage area;
each time one of the data is acquired, writing the acquired one data to the address indicated by the first variable, and after the writing, incrementing the address indicated by the first variable;
Each time one of the data is obtained, it is checked whether the address indicated by the first variable is the {k×(MN)+1}-th address of the storage area, and the address indicated by the first variable is checked. if the address is the {k×(M−N)+1}-th address, setting the address indicated by the second variable as the first address;
Each time one piece of data is obtained, the address indicated by the first variable is the {k×(MN)+1}th to {k×(MN)+N}th addresses of the storage area. is included in the range, and if the address indicated by the first variable is included in the address range, the obtained one data is written to the address indicated by the second variable; 2 increment the address indicated by the variable,
when the address indicated by the first variable exceeds the {k×(MN)+N}-th address of the storage area, the address indicated by the first variable is changed to the address indicated by the second variable; set,
5. A data processing apparatus according to claim 3 or 4. The storage area is composed of L addresses (L is an integer of 2×M or more) of the addresses,
The process of generating the address and the process of writing the data include:
In an initial state, setting an address interval between the address indicated by the first variable and the address indicated by the second variable to M or more (LM) or less;
Each time one of the data is acquired, the acquired one data is written to the address indicated by the first variable and the address indicated by the second variable, and after the writing, the first variable is indicated. each incrementing the address and the address indicated by the second variable;
when the address indicated by the first variable exceeds the maximum address of the storage area, setting the address indicated by the first variable to the top address of the storage area;
when the address indicated by the second variable exceeds the maximum address, setting the address indicated by the second variable as the top address;
The process of detecting the completion of writing is performed by dividing one unit address range generated corresponding to one of the data strings by the process of generating the address, where the address indicated by the first variable and the second variable are identified based on said address indicated and
The process of performing the predetermined calculation reads the data string used for the predetermined calculation from the unit address range specified by the process of detecting the completion of writing.
3. A data processing apparatus according to claim 1 or 2. the predetermined operation comprises a discrete Fourier transform;
A data processing device according to any one of claims 1 to 6. A data processing method in which one or more computers process time-series data in real time,
The one or more computers include a storage unit having a storage area for temporarily storing each data of the time-series data,
M consecutive data in the time-series data constitute one data string,
Two of the data strings that are continuous in the time-series data include N pieces of the data that are common to each other (N is a natural number smaller than M),
The computer
A process of sequentially acquiring each of the data of the time-series data;
a process of generating an address to which the acquired data is written in the storage area;
a process of writing the obtained data to the generated address;
a process of detecting completion of writing of each data string;
each time the completion of writing of one of the data strings is detected, a process of performing a predetermined operation based on the one data string;
The process of generating the addresses includes generating the M consecutive addresses corresponding to the M data constituting one data string as one unit address range,
One unit address range is
N first addresses corresponding to the first half N data in the one data string;
and (MN) second addresses corresponding to the latter (MN) data in the one data string,
In the process of generating the addresses, when the number of addresses after one unit address range corresponding to one data string is less than (MN), N before the one unit address range the consecutive addresses are the N first addresses in the next unit address range corresponding to the next data string following the one data string;
Data processing method. A program that causes one or more computers to function as the processing unit in the data processing apparatus according to any one of claims 1 to 7. an imaging device that captures an image of a rotating body at regular intervals;
A data processing device according to claim 7,
The process of sequentially acquiring the data of the time-series data includes numerical data corresponding to the values of pixels included in a specific region in the image of the body of rotation captured by the imaging device. obtained periodically as said data;
The processing unit calculates the rotation speed of the rotating body based on the frequency component of the time-series data indicated by the result of the discrete Fourier transform performed as the predetermined calculation.
Rotational speed measuring device.

Images