JP7229838B2 - Flow estimation device - Google Patents

Description

Embodiments relate to droplet flow rate estimation.

Conventionally, for example, in intravenous drip injection using an infusion pump, in order to administer a drug solution at an accurate infusion rate (i.e., flow rate), a drip sensor is used to detect the drop of a droplet in the drip cylinder, and the droplet Feedback control is performed on the infusion pump based on the number of drops. Feedback control of the injection rate typically assumes that the volume of the falling droplets is constant, but if the volume is accurately calculated sequentially, it is possible to estimate the flow rate of the droplets with higher accuracy. can.

The area around the liquid drop nozzle is continuously photographed, the image immediately after the liquid drop leaves the drop nozzle is specified from the captured image, and the volume of the liquid drop reflected in the image is calculated by the piecewise quadrature method. techniques are known. However, the drop speed of the droplet increases approximately in proportion to the elapsed time from the start of drop, and the higher the drop speed, the more blurred the droplet is photographed, so the error in the calculated volume tends to increase. On the other hand, in order to reliably capture the image immediately after the droplet leaves the dropping nozzle, it is necessary to have a wide imaging range (that is, a high resolution) in the direction in which the droplet falls (usually the vertical direction) and a high frame rate. It is necessary to shoot using a high-rate two-dimensional image sensor. Generally speaking, there is a trade-off relationship between the accuracy of droplet volume calculation and the cost of equipment for photographing the droplet.

In addition, this technique cannot estimate the flow rate of a drop until at least one drop has fallen. Therefore, the longer the drop cycle of the droplets, the longer it takes to reflect the latest estimation results after adjusting the flow rate of the droplets. This is inconvenient in the use case of adjusting the flow rate to a target value. Also, since the droplet volume increases nonlinearly rather than linearly until it falls, it is not easy to estimate an accurate flow rate based on observations over part of the fall period.

Japanese Patent No. 5583939

Embodiments aim to estimate the flow rate of a droplet based on observations over a period shorter than the drop period of the droplet.

According to the embodiment, the flow rate estimation device includes an imaging section, a volume calculation section, an approximation calculation section, and a flow rate calculation section. The imaging unit generates a series of images by continuously photographing the ejection port that ejects at least one droplet including the first droplet. The volume calculation unit calculates a plurality of images associated with a second period, which is a part of the first period from when the first droplet adheres to the tip of the ejection port until it falls, from the series of images. Calculate the volume of the first drop reflected on each. The approximation calculator approximates the time change of the volume of the first droplet in the second period with a first straight line. The flow rate calculation unit determines a correction coefficient corresponding to any one of the volumes of the first droplet in the second period according to a predetermined relationship between the volume and the correction coefficient, and calculates the correction coefficient on the first straight line. By applying the slope, an estimated flow rate based on observations of the first drop over the second time period is calculated.

The block diagram which illustrates the flow estimation apparatus which concerns on 1st Embodiment. 2 is a flowchart illustrating the operation of the flow rate estimation device of FIG. 1; FIG. 2 is a diagram exemplifying a series of images captured by the imaging unit in FIG. 1; 4A and 4B are graphs illustrating changes in droplet volume over time calculated based on the series of images in FIG. 3 and straight lines and cubic curves that approximate the change in volume over time; 5 is a graph exemplifying the relationship between time (frame) and the slope of the approximation straight line of the temporal change in volume around that time, which is derived based on the graph of FIG. 4 ; 6 is a graph exemplifying the relationship between the volume and the slope of the approximation straight line of the volume change over time in the vicinity of the volume, which is derived based on the graphs of FIGS. 4 and 5; 2 is a graph illustrating the relationship between a correction factor and volume used by the flow rate estimator of FIG. 1; 2 is a flowchart illustrating the operation of the flow rate estimation device of FIG. 1 in the process of deriving a correction coefficient;

Descriptions of the embodiments are given below with reference to the drawings. Elements that are the same as or similar to those already explained are denoted by the same or similar reference numerals, and overlapping explanations are basically omitted.

The flow rate estimating device according to each embodiment is used, for example, to estimate the flow rate of a drug liquid droplet in intravenous drip injection, but the flow rate of any other type of droplet (or the discharge amount per dropping cycle) may be used to estimate For example, this flow rate estimating device may be used to estimate the flow rate of ink droplets in an inkjet printer, chemical liquid droplets in a semiconductor manufacturing apparatus, and the like.

(First embodiment)
As illustrated in FIG. 1, the flow estimation apparatus according to the first embodiment includes an imaging unit 101, a volume calculation unit 102, a volume storage unit 103, an approximation calculation unit 104, and a correction coefficient derivation unit 105. , a correction coefficient storage unit 106 and a flow rate calculation unit 107 .

The photographing unit 101 continuously photographs the vicinity of the tip of a nozzle (that is, an ejection port) that ejects liquid droplets such as a chemical solution, so that the liquid droplets adhering to the tip of the nozzle gradually increase in volume. Generate a series of images representing . A series of images, for example, as shown in FIG. 3, is a plurality of images obtained by photographing at a predetermined frame rate how a droplet adheres to the tip of a nozzle and then gradually increases in volume. .

The number given to each image in FIG. 3 represents the frame number of the image. In the example of FIG. 3, images from frame number 1 to frame number 54 represent droplet volume growth over one drop cycle.

Since the frame rate at which the sequence of images shown in FIG. 3 was captured was 20 fps, the length of the fall cycle is approximately 54/20=2.7, or just under 3 seconds. Therefore, assuming the example of FIG. 3 and estimating the flow rate based on observations over the entire drop period, the flow rate estimation result will be updated at intervals of a little less than 3 seconds. On the other hand, if the flow rate is estimated based on observations with a unit period of 5 frames or 10 frames, the estimated flow rate is approximately 0.25 (=5/20) seconds or 0.5 (=10/20) seconds. will be updated in the cycle of Thus, for example, in the use case of manually or automatically adjusting the droplet flow rate to a target value based on the estimated flow rate, the latter may complete the task in less time than the former. be.

The imaging unit 101 outputs a series of images to the volume calculation unit 102 . Note that the image capturing unit 101 may sequentially output to the volume calculation unit 102 each time an image is captured, or may output a predetermined number of images (for example, 5 or 10) to the volume calculation unit 102 after capturing a predetermined number of images. You can also batch output to .

The photographing unit 101 is required to photograph the state in which the volume of the droplet adhering to the tip of the nozzle gradually increases. rate is not high. Furthermore, the image capturing range of the image capturing unit 101 may include the entire droplet adhering to the tip of the nozzle. For example, the imaging unit 101 can be implemented using a small and inexpensive two-dimensional image sensor (camera) with a low frame rate and low resolution. Furthermore, the lower the frame rate and resolution of the imaging unit 101, the smaller the data size of a series of images, so there is also the advantage that the time and memory capacity required for various calculation processes to be described later can be reduced.

A volume calculator 102 receives a series of images from the imaging unit 101 . A volume calculator 102 calculates the volume of the droplet in each of the series of images. The volume calculation unit 102 may perform volume calculation by various methods such as piecewise quadrature. The volume calculation unit 102 stores the volume calculation result in the volume storage unit 103 .

Specifically, the volume calculation unit 102 calculates a plurality of images associated with a unit period of flow rate estimation, which is a part of the drop cycle from when the droplet adheres to the tip of the nozzle until it falls, from the series of images. Calculate the volume of the droplet in each of the images. This unit period corresponds to, for example, 5 frames, 10 frames, etc., and the flow rate estimation apparatus in FIG. 1 can estimate the latest flow rate of droplets for each unit period. That is, the flow rate estimation device of FIG. 1 can estimate the flow rate of the droplet based on observation over a unit period shorter than the drop period of the droplet.

Furthermore, the volume calculation unit 102 may calculate the volume of the droplet reflected in each of the plurality of images associated with the entire drop cycle (third period) in the process of deriving the correction coefficient, which will be described later.

The volume storage unit 103 stores the calculation result of the droplet volume by the volume calculation unit 102 . Note that each volume may be stored in association with information that explicitly or implicitly indicates the time corresponding to the volume. Such information may be, for example, simply the value of the frame number of the image, or the value of the imaging time of the image, the calculation time of the volume, or the storage time of the volume.

The approximation calculation unit 104 reads the volume in the unit period from storage in the volume storage unit 103, and approximates the time change of the volume in the unit period with a straight line (first straight line) based on the volume. Here, as exemplified in FIG. 4, the volume changes with time in a nonlinear manner (for example, in the form of a cubic curve) over the fall period, so the slope of the approximate straight line of the time change of the volume in a unit period also changes with time. do. The approximation calculator 104 outputs the slope of the straight line to the flow rate calculator 107 .

Specifically, the approximation calculation unit 104 arranges volumes (hereinafter also referred to as samples) in a unit period in chronological order. Since the volume of the droplet increases substantially monotonically in the unit period, the temporal change in the volume of the droplet in the unit period can be approximated by a straight line l(t) shown in Equation (1) below.

l(t) represents the volume of the droplet at time t (t represents the frame number, for example). a represents the slope of the straight line l(t) and takes a positive value. b represents the intercept of the straight line l(t) with the volume axis. The approximation calculation unit 104 can calculate a and b by applying, for example, the least-squares method to the samples within the unit period.

Furthermore, in the process of deriving the correction coefficient, the approximation calculation unit 104 calculates the time change of the volume over the entire fall cycle (third period) by a straight line (second straight line: l _e (t) = a _e t + b _e ). approximation (see the dotted line in FIG. 4), and a straight line (third straight line) approximating the time change of the volume in the target period (fourth period) for deriving the correction coefficient, which is part of the fall cycle. The slope (a _e ) of this second line represents the estimated drop flow rate based on observations over the entire drop cycle. Approximation calculation section 104 outputs the slope of each straight line to correction coefficient derivation section 105 . By plotting the slope of the approximation line for multiple periods of interest, the time (frame) versus slope relationship as exemplified in FIG. 5 can be obtained.

Alternatively, the approximation calculation unit 104 approximates the time change in the volume of the droplet over the drop period by a function f(t) (a cubic function in the example of FIG. 4), and the first derivative of this function f' ( By calculating t) (which is a quadratic function in the example of FIG. 5), the slope in each frame can also be derived collectively. Therefore, the approximation calculation unit 104 may output the first derivative to the correction coefficient derivation unit 105 instead of the slope.

Note that the approximation calculation unit 104 sets a plurality of target periods for one fall cycle, approximates the change in volume over time with a straight line for each target period, and outputs the slope to the correction coefficient derivation unit 105. good too.

The flow rate calculation unit 107 receives from the approximation calculation unit 104 the slope of the straight line approximating the time change of the droplet volume in the unit period. Further, the flow rate calculation unit 107 determines a correction coefficient corresponding to any one of droplet volumes in a unit period according to a predetermined relationship between the volume and the correction coefficient. Such a relationship may be defined, for example, by a function or table (for example, lookup table (LUT)) stored in correction coefficient storage unit 106 . In this case, the flow rate calculation unit 107 reads a correction coefficient corresponding to one of the droplet volumes in the unit period from the correction coefficient storage unit 106, or calculates a function value corresponding to the volume. , may determine the correction factor.

Note that the volume used to determine the correction coefficient may be, for example, the volume near the center of the unit period. That is, if the total number of images associated with a unit period is an odd number, the volume calculated from the one image that is chronologically centered among them may be used to determine the correction factor. On the other hand, if the total number of images associated with the unit period is even, then the mean of the volumes calculated from the two images occupying the middle in their chronological order may be used to determine the correction factor. good.

The flow rate calculator 107 calculates the (estimated) flow rate of droplets based on observations over a unit period of flow rate estimation by applying the determined correction factor to the slope of the straight line. For example, the flow rate calculator 107 may multiply the slope of the straight line by a correction factor or its reciprocal. The flow rate is typically the volume of droplets flowing through the ejection port cross section per unit time (for example, ml/H), but the flow rate may be the mass of droplets flowing through the ejection port cross section per unit time. good. Volume can be converted to mass if the density of the droplet is known.

The flow estimator of FIG. 1 operates as illustrated in FIG. The process of FIG. 2 starts from step S201. In step S201, the image capturing unit 101 captures an image of the vicinity of the nozzle that ejects droplets (specifically, a range that can cover the entire droplet even if the volume of the droplet has grown to the maximum).

The volume calculation unit 102 calculates the volume of the droplet captured in the image captured in step S201 (step S202), and stores the volume in the volume storage unit 103 (step S203).

The imaging unit 101 continues imaging (loop from step S201 to step S204) until the unit period for flow rate estimation ends (for example, a predetermined number of images are completed). On the other hand, when the photographing of a plurality of images associated with the unit period, the calculation of the volume captured in the image, and the storage of the volume are completed, the process proceeds to step S205.

In step S205, the approximation calculation unit 104 reads the volume stored in step S203, and approximates the time change of the droplet volume over the unit period with a straight line. The flow rate calculator 107 determines a correction coefficient corresponding to one of the droplet volumes in a unit period according to a predetermined relationship between the volume and the correction coefficient (step S206). Then, the flow rate calculator 107 calculates the flow rate by applying the correction coefficient determined in step S206 to the slope of the straight line calculated in step S205 (step S207), and the process of FIG. 2 ends.

As described above, the flow rate calculator 107 calculates the flow rate of droplets based on the predetermined relationship between the volume and the correction coefficient. Therefore, the process of deriving the correction coefficient is performed prior to the flow rate calculation.

In the process of deriving the correction coefficient, the correction coefficient derivation unit 105 obtains from the approximation calculation unit 104 the slope _ae of the straight line that approximates the time change of the droplet volume in the entire drop cycle, and the part of the drop cycle. Receive the slope of a straight line that approximates the period of interest for correction factor derivation (or the first derivative f'(t) of the function f(t) that approximates the time variation of droplet volume over the drop period). It should be noted that, as described above, there is a possibility that a plurality of target periods may be set for one fall cycle, not just one. In this case, the approximation calculation unit 104 receives the slope for each of the plurality of target periods.

The correction coefficient deriving unit 105 associates the slope of the target period for deriving the correction coefficient with any of the volumes in the target period, thereby obtaining the relationship between volume and slope as illustrated in FIG. 6 .

When the correction coefficient derivation unit 105 receives the first derivative f′(t), the inverse function f ⁻¹ (v) of the function f(t) (this inverse function is the time corresponding to the volume v ) can be used to associate the slope with the volume v as follows. g(v) in equation (2) (see, eg, FIG. 6) returns the slope corresponding to volume v.

Then, the correction coefficient deriving unit 105 calculates the ratio between the slope _ae for the entire fall cycle and the slope associated with the volume. That is, the correction coefficient derivation unit 105 may divide the former slope _ae by the latter slope, or may divide the latter slope by the former slope _ae . A correction coefficient derivation unit 105 derives this ratio as a correction coefficient corresponding to the volume. In the following description, the ratio obtained by dividing the slope _ae for the entire drop cycle by the slope associated with the volume is taken as the correction coefficient corresponding to the volume. The correction coefficient derivation unit 105 can derive correction coefficients corresponding to a plurality of volumes by repeating the same processing for a plurality of target periods. In order to derive more correction coefficients, the correction coefficient deriving unit 105 may repeat the same process for not only one falling cycle but also a plurality of falling cycles.

Note that the volume associated with the slope for the target period for deriving the correction coefficient may be, for example, the volume near the center of the target period. That is, if the total number of images associated with a period of interest is an odd number, then the volume computed from the one central image in their chronological order may be associated with the slope for that period of interest. On the other hand, if the total number of images associated with the period of interest is even, then the mean of the volume calculated from the two central images in their chronological order is associated with the slope for that period of interest. good too.

Alternatively, the correction coefficient deriving unit 105 may divide the function g(v) of formula (2) by the slope _ae for the entire fall cycle, or divide the slope _ae by the function g(v) of formula (2). ). According to such calculation, correction coefficients corresponding to each volume v can be collectively derived.

As described above, the slope _ae for the entire drop cycle represents the flow rate (that is, the volume of droplets dropped per unit time) in the drop cycle. It has been empirically found that, if the flow rate does not change, the shapes of the curves representing the change in volume over time for different drop periods are similar. On the other hand, when there is a change in the flow rate, the shape of the curve after the change is similar to the shape obtained by compressing or extending the curve in the fall cycle before the change in the time direction according to the change rate of the flow rate. clearly evident. That is, the slope of the approximation straight line near a given volume in the curve after the change in flow rate is approximately similar to the value obtained by multiplying the slope of the approximation straight line near that volume in the curve before the change by the change rate of the flow rate. Therefore, the following formula (3) holds.

In equation (3), a represents the slope of a straight line approximating the time change of the droplet volume in the unit period of flow rate estimation (see equation (1)), v is the droplet near the center of the unit period, for example and g(v) is a function that returns the slope corresponding to that volume v (see equation (2). flow rate is the observed flow rate over this unit period, and a _e is the slope of the second straight line as before and represents the estimated flow rate of the drop based on observations over the (at least one) drop period, thus flow rate/a _e is the correction factor Denotes the rate of change in observed flow over a unit period relative to the estimated flow in the derivation process, and a _e /g(v) is the correction factor (c(v)) corresponding to volume v. means.

As can be read from the formula (3), the slope a of the straight line that approximates the time change of the droplet volume in the unit period of the flow rate estimation has a correction coefficient c ( By multiplying v), it is possible to estimate the observed flow rate over the unit period. That is, the flow rate can be estimated using the periodicity of the shape of the growth curve of the volume of the droplet without waiting for the drop of the droplet, as long as the local time change of the volume of the droplet is analyzed.

The correction coefficient derivation unit 105 stores the derived correction coefficient in the correction coefficient storage unit 106 in association with the volume. The accumulation of pairs of volumes and correction coefficients stored in the correction coefficient storage unit 106 corresponds to information representing the relationship between volumes and correction coefficients. This information may be stored in the correction coefficient storage unit 106 in the form of a lookup table, for example. Alternatively, the correction coefficient derivation unit 105 may obtain a function representing the above relationship (for example, a quartic function of volume as shown in FIG. 7) by approximating the derived correction coefficients as a volume polynomial. . In this case, the correction coefficient derivation unit 105 may store the coefficient of each term of the function in the correction coefficient storage unit 106, for example.

The flow rate estimation device of FIG. 1 operates as illustrated in FIG. 8 in the process of deriving the correction coefficient. The process of FIG. 8 starts from step S301. In step S301, the photographing unit 101 photographs the vicinity of the nozzles that eject droplets.

The volume calculation unit 102 calculates the volume of the droplet captured in the image captured in step S301 (step S302), and stores the volume in the volume storage unit 103 (step S303).

The imaging unit 101 continues imaging until at least one drop cycle ends (for example, the droplet adheres to the tip of the nozzle, gradually increases in volume, and finally falls) (steps S301 to S304). loop until). On the other hand, when the shooting of a plurality of images associated with the fall cycle, the calculation of the volume reflected in the images, and the storage of the volume are completed, the process proceeds to step S305.

In step S305, the approximation calculation unit 104 reads out the volume stored in step S303, approximates the time change of the droplet volume over the entire drop cycle with a straight line, and further calculates a correction coefficient that is a part of the drop cycle. A straight line approximates the time variation of the droplet volume over the period of interest of the derivation. The correction coefficient derivation unit 105 derives the ratio of the slope for the entire fall cycle and the slope for the target period as a correction coefficient corresponding to any of the volumes in the target period (step S306). The correction coefficient derivation unit 105 associates the correction coefficient derived in step S306 with the volume and stores it in the correction coefficient storage unit 106, and the processing in FIG. 8 ends. Steps S306 to S307 may be repeatedly performed for a plurality of target periods belonging to the same fall cycle, or may be repeatedly performed for a plurality of target periods belonging to different fall cycles.

As described above, the flow rate estimation device according to the first embodiment approximates the time change of the droplet volume over a unit period, which is a part of the droplet falling period, with a straight line, and The slope of the straight line is converted into the flow rate using the correction coefficient associated with the volume. Here, the correction coefficient represents the ratio between the slope of a straight line approximating the time change of the droplet volume over the entire drop cycle calculated in the past and the slope of the local straight line near the volume. Therefore, according to this flow rate estimating apparatus, the flow rate can be estimated for each unit period without waiting for the drop of the droplet by using the periodicity of the shape of the growth curve of the volume of the droplet. In addition, according to this flow rate estimation device, it is not necessary to capture the image at the moment the droplet falls, but it is sufficient to capture the temporal change in the volume of the droplet attached to the nozzle. Droplet flow rates can be estimated without the need for equipment.

Various functional units described in the above embodiments may be realized by using circuits. A circuit may be a dedicated circuit that implements a specific function, or it may be a general-purpose circuit such as a processor.

At least part of the processing in each of the above embodiments can also be implemented using a general-purpose computer as basic hardware. A program that implements the above process may be provided by being stored in a computer-readable recording medium. The program is stored in the recording medium as an installable format file or an executable format file. Recording media include magnetic disks, optical disks (CD-ROM, CD-R, DVD, etc.), magneto-optical disks (MO, etc.), semiconductor memories, and the like. Any recording medium may be used as long as it can store the program and is readable by a computer. Alternatively, the program that implements the above processing may be stored on a computer (server) connected to a network such as the Internet, and downloaded to the computer (client) via the network.

While several embodiments of the invention have been described, these embodiments have been presented by way of example and are not intended to limit the scope of the invention. These novel embodiments can be implemented in various other forms, and various omissions, replacements, and modifications can be made without departing from the scope of the invention. These embodiments and modifications thereof are included in the scope and gist of the invention, and are included in the scope of the invention described in the claims and equivalents thereof.

101... Imaging unit 102... Volume calculation unit 103... Volume storage unit 104... Approximation calculation unit 105... Correction coefficient derivation unit 106... Correction coefficient storage unit 107... Flow rate calculation unit

Claims (1)

an image capturing unit configured to generate a series of images by continuously capturing images of an ejection port that ejects at least one droplet including the first droplet;
Each of the plurality of images in the series of images is associated with a second period that is a part of the first period from when the first droplet adheres to the tip of the ejection port until it falls. a volume calculation unit that calculates the volume of the reflected first droplet;
an approximation calculation unit that approximates a time change in the volume of the first droplet during the second period with a first straight line;
determining a correction coefficient corresponding to one of the volumes of the first droplet in the second period according to a predetermined relationship between the volume and the correction coefficient, and applying the correction coefficient to the slope of the first straight line; and a flow rate calculation unit adapted to calculate an estimated flow rate based on observations of said first droplet over said second period of time.

Images