Device and method for measuring stream water depth in real-time through positioning data filtering

Abstract

Disclosed is a device and a method for measuring stream water depth in real-time through positioning data filtering to ensure the reliability of the measured water depth data even when the stream water depth measurement data is filtered and applied to a small stream having a small basin area and a steep slope. The device for measuring stream water depth in real-time through positioning data filtering includes: a measuring part for measuring the water depth of a stream to be measured; a positioning data filtering part for filtering the water depth data measured by the measuring means by a local linear regression-based bivariate scatterplot smoothing technique through elastic bandwidth application; and a water depth calculating part for calculating a water depth of the stream to be measured by using the positioning data filtered by the positioning data filtering part, and minimizing the uncertainty of the water depth measurement.

Claims

1. A real-time stream water depth measurement device using positioning data filtering, the real-time stream water depth measurement device comprising: a measuring unit configured to measure a water depth of a stream to be measured; a positioning data filtering unit configured to filter water depth data measured by the measuring unit using a local linear regression-based bivariate scatterplot smoothing technique through adaptive bandwidth application; and a water depth calculating unit configured to calculate the water depth of the stream to be measured using the positioning data, which is filtered by the positioning data filtering unit and in which uncertainty of the water depth measurement is minimized, wherein the positioning data filtering unit includes: a smoothed value estimating unit configured to estimate a smoothed value at an arbitrary position in the measured positioning data; a smoothed most probable value calculating unit configured to calculate a plurality of smoothed most probable values for an initial fixed bandwidth such that a square of an estimated error is minimum; a cross-validated residual calculating unit configured to calculate a cross-validated residual for each bandwidth; an optimal bandwidth calculating unit configured to calculate an optimal bandwidth for each point; a bandwidth selecting unit configured to select a bandwidth close to an optimal bandwidth from among initially input bandwidths; and a final smoothed most probable value calculating unit configured to calculate a final smoothed most probable value by linear interpolation using the smoothed most probable value.

2. The real-time stream water depth measurement device of claim 1, wherein the measuring unit is an ultrasonic sensor which transmits ultrasonic waves to a surface of a measurement target, receives reflected waves formed by the transmitted ultrasonic waves being reflected from the surface of the measurement target and being returned, and measures a distance from the surface of the measurement target on the basis of a time difference between an ultrasonic transmission time and a reflected wave reception time, an ultrasonic transmission velocity, and a reflection velocity of the reflected wave.

3. The real-time stream water depth measurement device according to claim 1, further comprising a communication unit configured to transmit the calculated water depth information calculated by the water depth calculating unit to a stream information management server that comprehensively manages stream information.

4. The real-time stream water depth measurement device according to claim 1, further comprising a Web server configured to receive the calculated water depth information by the water depth calculating unit through wired and/or wireless communication, store the calculated water depth information in a database using a Web service, and provide the corresponding result through a Web page.

5. A real-time stream water depth measurement device using positioning data filtering, comprising: a Global Positioning System (GPS) reception unit configured to receive GPS information according to an installation position of the stream water depth measurement device from a GPS satellite; an ultrasonic sensor unit configured to transmit ultrasonic waves to a surface of a measurement target, receive reflected waves formed by the transmitted ultrasonic waves being reflected from the surface of the measurement target and being returned, and measure a distance from the surface of the measurement target on the basis of a time difference between an ultrasonic transmission time and a reflected wave reception time, an ultrasonic transmission velocity, and a reflection velocity of the reflected wave; a positioning data filtering unit configured to filter the positioning data of the ultrasonic sensor unit and minimize the uncertainty of the water depth measurement due to rough water surface ripples of a small stream; and a water depth calculating unit configured to calculate a water depth, which is a distance from a bottom of the measurement target to the water surface of the measurement target, using altitude information included in the GPS information received from the GPS reception unit, distance information filtered by the positioning data filtering unit and measured by the ultrasonic sensor unit, and distance information between a structure and the bottom of the measurement target, wherein the positioning data filtering unit includes: a smoothed value estimating unit configured to estimate a smoothed value at an arbitrary position in the measured positioning data; a smoothed most probable value calculating unit configured to calculate a plurality of smoothed most probable values for an initial fixed bandwidth such that a square of an estimated error is minimum; a cross-validated residual calculating unit configured to calculate a cross-validated residual for each bandwidth; an optimal bandwidth calculating unit configured to calculate an optimal bandwidth for each point; a bandwidth selecting unit configured to select a bandwidth close to an optimal bandwidth from among initially input bandwidths; and a final smoothed most probable value calculating unit configured to calculate a final smoothed most probable value by linear interpolation using the smoothed most probable value.

6. The real-time stream water depth measurement device according to claim 5, further comprising a Web server configured to receive the calculated water depth information by the water depth calculating unit through wired and/or wireless communication, store the calculated water depth information in a database using a Web service, and provide the corresponding result through a Web page.

7. The real-time stream water depth measurement device of claim 5, wherein; the GPS information received by the GPS reception unit includes position information composed of latitude and longitude coordinates values, altitude information representing an altitude, and time information composed of date and time; and the GPS reception unit receives position information about a position at which the stream water depth measurement device is installed and provides the position information about the corresponding position when the water depth information of the measurement target is provided at the corresponding position.

8. The real-time stream water depth measurement device according to claim 5, further comprising a communication unit configured to transmit the calculated water depth information calculated by the water depth calculating unit to a stream information management server that comprehensively manages stream information.

9. A real-time stream water depth measurement method using positioning data filtering, the real-time stream water depth measurement method comprising: transmitting ultrasonic waves to a surface of a measurement target and receiving reflected waves formed by the transmitted ultrasonic waves being reflected from the surface of the measurement target and returned; estimating a smoothed value at an arbitrary position to filter the positioning data received by the ultrasonic sensor, calculating a plurality of smoothed most probable values for an initial fixed bandwidth such that a square of an estimated error is minimum, and calculating a cross-validated residual for each bandwidth; calculating an optimal bandwidth for each point and selecting a bandwidth close to the optimal bandwidth from among initially input bandwidths; calculating a final smoothed most probable value by linear interpolation using the smoothed most probable value; and calculating a distance from the surface of the measurement target using the filtered positioning data on the basis of a time difference between an ultrasonic transmission time and a reflected wave reception time, an ultrasonic transmission velocity, and a reflection velocity of the reflected wave and calculating a water depth.

10. The real-time stream water depth measurement method of claim 9, further comprising transmitting water depth information calculated in the calculating of the water depth to a system controller and a server PC in a wired or wireless manner to be utilized as input data of an automatic discharge measurement program, and checking the water depth information in real time using a real-time water depth and discharge display Web service.

11. The real-time stream water depth measurement method of claim 9, wherein, in order to estimate a smoothed value at an arbitrary position for filtering the positioning data, {circumflex over (f)}(x.sub.i), which is a smoothed value at an arbitrary position x.sub.i in scatterplots, is estimated by local regression using a least-squares method using an equation, Ê[Y|x.sub.i]={circumflex over (α)}+{circumflex over (β)}x.sub.j, x.sub.j∈N.sub.i, here, N denotes a local bandwidth around x.sub.i and is optimized for each position, and when a specific bandwidth J is given and the number of parameters in the bandwidth is N, a local linear regression tracker is obtained using an equation, ŷ.sub.k={circumflex over (α)}+{circumflex over (β)}x.sub.k, k=1, . . . , N, here, {circumflex over (α)} and {circumflex over (β)} may be obtained by linear regression of local data which is present in i.sub.−J/2, . . . , i.sub.+J/2, and ŷ.sub.i′ becomes a smoothed most probable value corresponding to the position of x.sub.i.

12. The real-time stream water depth measurement method of claim 11, wherein, in order to calculate a smoothed most probable value that is linked to a locally adaptive bandwidth, when an observation value of y=f(x) and a bandwidth of J(x) at the x position are given, J(x) and f(x) are calculated such that the square of an estimation error is minimized, using an equation, e.sup.2(f,J)=E[Y−f(X|J(x))].sup.2.

13. The real-time stream water depth measurement method of claim 12, wherein, in order to minimize, when the number of pieces of whole data is n, a plurality of smoothed most probable values f(x) are calculated for initial fixed bandwidths J of various sizes between 0<j<n, and a cross-validated residual is calculated for each bandwidth J by a cross-validated residual calculating unit using an equation, $r_{\left(i\right)\left(J\right)}=\left[y_i-\hat{f}\left(x_i|J\right)\right]/\left(1-\frac{1}{J}-\frac{{\left(x_i-{\stackrel{\_}{x}}_J\right)}^2}{V_J}\right),$ such that the square of an estimation error is minimized, wherein equations, $V_J=\frac{1}{n}\underset{j=i-J/2}{\overset{i+J/2}{.Math.}}{\left(x_i-{\stackrel{\_}{x}}_J\right)}^2,{\stackrel{\_}{x}}_J=\frac{1}{J}\underset{j=i-J/2}{\overset{i+J/2}{.Math.}}{x}_i$ are satisfied.

14. The real-time stream water depth measurement method of claim 13, wherein, in order to derive ê(f,J|x.sub.i) as a result of smoothing |r.sub.(i)(J)| for x.sub.i having a bandwidth of J=0.2n and to calculate an optimal bandwidth J.sub.cv(x.sub.i) at each point, the optimal bandwidth calculated using an equation, ê(f,J.sub.cv|x.sub.i)=min ê(f,J|x.sub.i), and a value of J corresponding to ê.sub.min among derived most probable errors ê for the plurality of bandwidths J in a specific position x.sub.i becomes the optimal bandwidth J.sub.cv.

15. The real-time stream water depth measurement method of claim 14, wherein J.sub.cv(x.sub.i) for each is re-smoothed with the bandwidth of J=0.2n, and then a bandwidth close to J.sub.cv(x.sub.i) among the bandwidths initially input by a bandwidth selecting unit is selected using an equation, J.sub.i1≤J.sub.cv(x.sub.i)≤J.sub.i2, when smoothed most probable values corresponding to J.sub.i1, J.sub.i2 which are selected for the obtained x.sub.i are y.sub.i1*, y.sub.i2*, respectively, y.sub.i* is obtained using the two values by linear interpolation using an equation, y.sub.i*=(y.sub.i1*−y.sub.i2*)/(J.sub.i1−J.sub.i2)(J.sub.cv*(x.sub.i)−J.sub.i2)+y.sub.i2*, and the final smoothed most probable value, y.sub.i*, corresponding to x.sub.i reuses the value smoothed with a bandwidth of J=0.5n.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) FIG. 1A is a photograph showing a state of a rough water surface of a small stream during flooding.

(2) FIG. 1B is a graph of a result of measuring a water depth of a small stream during flooding.

(3) FIG. 2 is a configuration diagram of a real-time stream water depth measurement device using positioning data filtering according to an embodiment of the present invention.

(4) FIG. 3A is a configuration diagram illustrating an example of a position at which a stream water depth measurement device is installed.

(5) FIG. 3B is a configuration diagram of an ultrasonic water depth measurement device based on Arduino according to an embodiment of the present invention.

(6) FIG. 4 is a detailed configuration diagram of a positioning data filtering unit according to the present invention.

(7) FIG. 5 is a configuration diagram of bandwidth setting when positioning data is filtered according to the present invention.

(8) FIG. 6 is a graph showing a difference between results of smoothing application according to a bandwidth.

(9) FIG. 7 is a flowchart illustrating a real-time stream water depth measurement method using positioning data filtering according to the present invention.

DETAILED DESCRIPTION

(10) Hereinafter, exemplary embodiments of a device and a method for measuring a stream water depth in real time using positioning data filtering according to the present invention will be described in detail as follows.

(11) Features and advantages of the device and the method for measuring the stream water depth in real time using positioning data filtering according to the present invention will become clear from the detailed description of each embodiment below.

(12) FIG. 2 is a configuration diagram of a real-time stream water depth measurement device using positioning data filtering according to an embodiment of the present invention.

(13) FIG. 3A is a configuration diagram illustrating an example of a position at which a stream water depth measurement device is installed, and FIG. 3B is a configuration diagram of an ultrasonic water depth measurement device based on Arduino according to an embodiment of the present invention.

(14) In the following description, a stream water depth measurement device using an ultrasonic sensor is described as an example of a stream water depth measuring unit to which the present invention is applied, but the present invention is not limited thereto, and the present invention may also be applied to a stream water depth measurement device using spatiotemporal image analysis or a combination thereof.

(15) In addition, a positioning data filtering unit according to the present invention is described as being included in the stream water depth measurement device, but the present invention is not limited thereto, and a location at which the positioning data filtering unit is installed may be included in a stream information management server that comprehensively manages stream information.

(16) In addition, the present invention includes an algorithm for filtering water depth data of a small stream during flooding which is measured using the stream water depth measurement device and includes a water depth data correction algorithm for minimizing uncertainty of water depth measurement due to rough water surface ripples.

(17) In particular, the measured water depth data may be transmitted to a system controller and a server PC in a wired or wireless manner to be utilized as input data of an automatic discharge measurement program, and the measured water depth data may be checked in real time by a real-time water depth and discharge display Web service.

(18) In addition, a stream water depth measurement device based on Arduino using an open source based single board microcontroller may be included so that development and application environments may be improved.

(19) The device and the method for measuring the stream water depth in real time using positioning data filtering according to the present invention include a positioning data filtering unit for increasing the accuracy by filtering the positioning data received from the water depth measuring unit using a local linear regression-based bivariate scatterplot smoothing technique through adaptive bandwidth application.

(20) The real-time stream water depth measurement device using positioning data filtering according to the embodiment of the present invention includes a Global Positioning System (GPS) reception unit 10 which receives GPS information according to an installation position of the stream water depth measurement device from a GPS satellite, an ultrasonic sensor unit 20 which transmits ultrasonic waves to a surface of a measurement target, receives reflected waves formed by the transmitted ultrasonic waves being reflected from the surface of the measurement target and being returned, and measures a distance from the surface of the measurement target on the basis of a time difference between an ultrasonic transmission time and a reflected wave reception time, an ultrasonic transmission velocity, a reflection velocity of the reflected wave, and the like, a positioning data filtering unit 30 which filters the positioning data of the ultrasonic sensor unit 20 to minimize the uncertainty of the water depth measurement due to the rough water surface ripples of the small stream, a water depth calculating unit 40 which calculates a water depth, which is a distance from a bottom of the measurement target to the water surface of the measurement target, using altitude information included in the GPS information received from the GPS reception unit 10, distance information filtered by the positioning data filtering unit 30 and measured by the ultrasonic sensor unit 20, and distance information between a structure and the bottom of the measurement target, a wireless communication unit 50 which enables wireless communication with an external communication device to transmit the water depth information of the measurement target calculated by the water depth calculating unit 40 and the GPS information according to the measurement position received by the GPS reception unit 10 to the external communication device in a wireless manner, and a Web server 60 which receives the calculated water depth information through the wireless communication unit 50, stores the calculated water depth information in a database using a Web service, and provides the corresponding result through a Web page, as illustrated in FIG. 2.

(21) Here, the GPS reception unit 10 receives position information about a position at which the stream water depth measurement device is installed and provides the position information about the corresponding position when the water depth information of the measurement target is provided at the corresponding position.

(22) The GPS information includes all of position information composed of latitude and longitude coordinates values, altitude information representing an altitude, time information composed of date and time, and the like.

(23) The Web service of the Web server 60 includes an application programming interface (API) that can retrieve and provide the water depth information.

(24) The distance information between the structure and the bottom of the measurement target may be replaced with any information that can calculate the water depth of the measurement target when only the altitude information received from the GPS reception unit 10 and the distance information measured by the ultrasonic sensor unit 20 are present.

(25) The wireless communication unit 50 may transmit the water depth information of the measurement target and the GPS information corresponding to the water depth information to an external communication device such as a laptop computer, a smart phone, or a tablet PC using Wi-Fi communication, mobile communication, near field communication, or wireless Internet communication so as to separately store or analyze the water depth information.

(26) The external communication device may be a stream information management server that comprehensively manages the stream information and may use wired communication to increase stability rather than wireless communication.

(27) The real-time stream water depth measurement device using positioning data filtering according to the present invention is installed in the middle of a bridge of the stream, as illustrated in FIG. 3A, and may be installed to transmit data by connecting continuous power and the Internet.

(28) A measurement range of the ultrasonic water depth sensor of the ultrasonic sensor unit 20 may be 5 m or more in consideration of a height from the bridge to the stream bed, a minimum measurement time interval may be less than 10 seconds, and the resolution may be less than 1 mm.

(29) The stream water depth measurement device installed as described above may include a stream water depth measurement device based on Arduino using an open source based single board microcontroller, as illustrated in FIG. 3B, so that the development and application environments may be improved.

(30) In addition, the real-time stream water depth measurement device using positioning data filtering according to the present invention should operate stably even under sudden heavy rains and thus should have a structure that is highly waterproof and dustproof.

(31) In particular, the stream water depth measurement device should be durable such that it can be operated for real-time small stream measurement without interruption and may be firmly installed to ensure normal operation even under extreme conditions such as lightning, typhoon, or heavy rain and thus all devices or facilities may be arranged for maximum performance to minimize obstacles.

(32) The detailed configuration of the positioning data filtering unit according to the present invention will be described as follows.

(33) FIG. 4 is a detailed configuration diagram of the positioning data filtering unit according to the present invention.

(34) The positioning data filtering unit 30 according to the present invention includes a smoothed value estimating unit 31 which estimates a smoothed value at an arbitrary position in the positioning data received by the ultrasonic sensor, a smoothed most probable value calculating unit 32 which calculates a plurality of smoothed most probable values for an initial fixed bandwidth such that a square of the estimated error is minimum, a cross-validated residual calculating unit 33 which calculates a cross-validated residual for each bandwidth, an optimal bandwidth calculating unit 34 which calculates an optimal bandwidth for each point, a bandwidth selecting unit 35 which selects a bandwidth close to the optimal bandwidth among initially input bandwidths, and a final smoothed most probable value calculating unit 36 which calculates a final smoothed most probable value by linear interpolation using the smoothed most probable value.

(35) The positioning data filtering according to the present invention will be described in detail as follows.

(36) Generally, a large amount of measurement data measured and inspected is represented as scatterplots with (x, y) type bivariate parameters, and the qualitative trends are analyzed by deriving mathematical stroke lines through linear or nonlinear regression.

(37) However, it is often impossible to derive quantitative mathematical stroke lines due to a lack of correlation between a large amount of time series measurement data or parameters outside a local domain.

(38) In this case, smoothing techniques have been mainly applied in which the fluctuations caused by random behavior are removed, local trends are reflected, and then overall strokes are derived.

(39) These techniques may basically use a least-squares method to estimate a most probable value on the basis of surrounding local data (local linear regression) and then link the estimated value to smooth scatterplots and track trends (scatterplot smoothing).

(40) As a representative scattering smoothing technique, there is locally weighted scatterplot smoothing (LOWESS) or locally estimated scatterplot smoothing (LOESS), which uses locally divided data based on a fixed bandwidth (or window).

(41) In the present invention, in order to improve the reliability of positioning data filtering, an adaptive bandwidth is calculated by reflecting the characteristics of the local data, thereby improving accuracy and processing efficiency, which is a disadvantage of the LOWESS.

(42) FIG. 5 is a configuration diagram of bandwidth setting when positioning data is filtered according to the present invention.

(43) The positioning data filtering unit 30 of the present invention divides bivariate data into adaptive bandwidths and calculates most probable values by applying weighted linear regression in each section.

(44) The positioning data filtering unit 30 estimates the most probable values by repeating the same task at all x positions or given spans and connects the x positions to obtain a smoothed stroke curve.

(45) The basis of scatterplot smoothing is the bandwidth width setting and the weighted least-squares method.

(46) As illustrated in FIG. 5, the bandwidth refers to a region for designating only a part of the scatterplot, a certain range (width) of data is selected and used for calculation based on a value of an x parameter of one object, and another bandwidth is set based on a value of an x parameter of a next object.

(47) That is, the bandwidth is used as a means for estimating a local pattern.

(48) FIG. 6 is a graph showing a difference between results of smoothing application according to a bandwidth.

(49) In the case of the LOWESS method, a bandwidth represented by a size of a fixed bandwidth is used as a number between ⅓ and ⅔ of all data at a corresponding position.

(50) In this case, as illustrated in FIG. 6, when the bandwidth is too large, a regression function appears as a flat curve similar to a straight line (under-fitting), and when the bandwidth is small, the regression function appears as a curve having a large degree of bending (over-fitting).

(51) Therefore, it is important to specify the appropriate bandwidth for the purpose of use.

(52) A filtering algorithm applied to the positioning data filtering unit 30 according to the present invention will be described in detail as follows.

(53) First, in the smoothed value estimating unit 31, {circumflex over (f)}(x.sub.i) which is a smoothed value (or a most probable value) at an arbitrary position x.sub.i in scatterplots is estimated by local regression using a least-squares method as follows.
Ê[Y|x.sub.i]={circumflex over (α)}+{circumflex over (β)}x.sub.j,x.sub.j∈N.sub.i  [Equation 1]

(54) Here, N denotes a local bandwidth around x.sub.i and is optimized for each position.

(55) When a specific bandwidth J is given and the number of parameters in the bandwidth is N, a local linear regression tracker is obtained by Equation 2 below.
ŷ.sub.k={circumflex over (α)}+{circumflex over (β)}x.sub.k,k=1, . . . ,N  [Equation 2]

(56) Here, {circumflex over (α)} and {circumflex over (β)} may be obtained by linear regression of local data which is present in i.sub.−J/2, . . . , i.sub.+J/2, and ŷ.sub.i′ becomes a smoothed most probable value corresponding to the position of x.sub.i.

(57) The bandwidth J is not fixed and an additional task proceeds as follows.

(58) In order to calculate a smoothed most probable value that is linked to a locally adaptive bandwidth, in a filtering algorithm (Friedman's Super Smoother) according to the present invention, when an observation value of y=f(x) and a bandwidth of J(x) are given at the x position, a basic principle is to calculate J(x) and f(x) such that a square of an estimation error is minimized, as shown in Equation 3.
e.sup.2(f,J)=E[Y−f(X|J(x))].sup.2  [Equation 3]

(59) In order to minimize Equation 3 in the smoothed most probable value calculating unit 32, when the number of pieces of whole data is n, a plurality of smoothed most probable values f(x) are calculated by applying Equation 2 to initial fixed bandwidths J, of various sizes between 0<j<n, and a cross-validated residual is calculated for each bandwidth, J, by the cross-validated residual calculating unit 33 as follows.

(60) $\begin{array}{cc}{r}_{\left(i\right)\left(J\right)}=\left[y_i-\hat{f}\left(x_i|J\right)\right]/\left(1-\frac{1}{J}-\frac{{\left(x_i-{\stackrel{\_}{x}}_J\right)}^2}{V_J}\right)& \left[\mathrm{Equation}4\right]\end{array}$

(61) In this case, equations,

(62) $V_J=\frac{1}{n}\underset{j=i-J/2}{\overset{i+J/2}{.Math.}}{\left(x_i-{\stackrel{\_}{x}}_J\right)}^2,{\stackrel{\_}{x}}_J=\frac{1}{J}\underset{j=i-J/2}{\overset{i+J/2}{.Math.}}{x}_i,$
are satisfied.

(63) Equation 5 below is used to derive ê(f,J|x.sub.i) as a result of smoothing |r.sub.(i)(J)| derived using Equation 4 for x.sub.i having a bandwidth of J=0.2n and to calculate an optimal bandwidth J.sub.cv(x.sub.i) at each point in the optimal bandwidth calculating unit 34.

(64) Equation 5 means that a value of J corresponding to ê.sub.min among derived most probable errors ê for a plurality of bandwidths J at a specific position x.sub.i becomes the optimal bandwidth J.sub.cv.
ê(f,J.sub.cv|x.sub.i)=minê(f,J|x.sub.i)  [Equation 5]

(65) J.sub.cv(x.sub.i) for each position x.sub.i is re-smoothed with a bandwidth of J=0.2n, and then a bandwidth close to J.sub.cv(x.sub.i) among the bandwidths initially input by the bandwidth selecting unit 35 is selected as in Equation 6.
J.sub.i1≤J.sub.cv(x.sub.i)≤J.sub.i2  [Equation 6]

(66) When smoothed most probable values corresponding to J.sub.i1, J.sub.i2 which are selected for x.sub.i previously obtained in the final smoothed most probable value calculating unit 36 are y.sub.i1*, y.sub.i2*, respectively, y.sub.i* is obtained using the two values by linear interpolation as follows.
y.sub.i*=(y.sub.i1*−y.sub.i2*)/(J.sub.i1−J.sub.i2)(J.sub.cv*(x.sub.i)−J.sub.i2)+y.sub.i2*  [Equation 7]

(67) The final smoothed most probable value y.sub.i* corresponding to x.sub.i reuses the value smoothed with a bandwidth of J=0.5n.

(68) Using such a positioning data filtering algorithm, the real-time stream water depth measurement device using positioning data filtering according to the present invention may filter measured stream water depth data to ensure the reliability of the measured water depth data even when the measured stream water depth data is applied to a small stream having a small basin area and a steep slope.

(69) A real-time stream water depth measurement method using positioning data filtering according to the present invention will be described in detail as follows.

(70) FIG. 7 is a flowchart illustrating a real-time stream water depth measurement method using positioning data filtering according to the present invention.

(71) First, ultrasonic waves are transmitted to a surface of a measurement target and reflected waves formed by the transmitted ultrasonic waves being reflected from the surface of the measurement target and being returned are received (S601).

(72) Subsequently, a smoothed value at an arbitrary position in positioning data received by an ultrasonic sensor is estimated (S602), a plurality of smoothed most probable values for an initial fixed bandwidth are calculated such that a square of an estimated error is minimum (S603), and a cross-validated residual is calculated for each bandwidth (S604).

(73) Subsequently, an optimal bandwidth is calculated for each point (S605), and a bandwidth close to an optimal bandwidth is selected from among initially input bandwidths (S606).

(74) A final smoothed most probable value is calculated by linear interpolation using the smoothed most probable value (S607).

(75) Subsequently, a distance from a surface of a measurement target is calculated using the filtered positioning data on the basis of a time difference between an ultrasonic transmission time and a reflected wave reception time, an ultrasonic transmission velocity, a reflection velocity of the reflected wave, and the like, and a water depth is calculated (S608).

(76) The calculated water depth information is stored in a database using a Web service of a Web server and the corresponding result is provided through a Web page of the Web server (S609).

(77) The water depth information calculated by filtering may be transmitted to a system controller and a server PC in a wired or wireless manner to be utilized as input data of an automatic discharge measurement program and may be used to check the water depth information in real time using a real-time water depth and discharge display Web service.

(78) The device and the method for measuring the stream water depth in real time using positioning data filtering according to the present invention described above may filter positioning data of a water depth measuring unit using a local linear regression-based bivariate scatterplot smoothing technique through adaptive bandwidth application so that water depth information with increased accuracy may be provided in real time.

(79) Therefore, it is possible to ensure the reliability of the measured water depth data even when the measured stream water depth data is applied to a small stream having a small basin area and a steep slope.

(80) While the present invention has been particularly described with reference to exemplary embodiments, it will be understood by those skilled in the art that various changes in form and details may be made without departing from the spirit and scope of the present invention.

(81) Therefore, the exemplary embodiments should be considered in a descriptive sense only and not for purposes of limitation. The scope of the invention is defined not by the detailed description of the invention but by the appended claims and encompasses all modifications and equivalents that fall within the scope of the appended claims and will be construed as being included in the present invention.

(82) The present invention relates to a device and a method for measuring a stream water depth in real time using positioning data filtering, which filters measured stream water depth data to ensure the reliability of the measured water depth data even when the measurement data for the stream water depth is applied to a small stream having a small basin area and a steep slope.