JP3032776B2 - Passage position detection method and passage position detection device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for determining whether a magnetic target whose magnetic condition is unknown has passed a fixed range from a magnetic detector whose position or state is stationary and fixed. The present invention relates to a passing position detecting method and a passing position detecting device which can surely detect a passing position regardless of the size.

[0002]

2. Description of the Related Art A magnetic body in the terrestrial magnetism is magnetized by its own magnetization or terrestrial magnetism. Therefore, the presence or absence of the magnetic material can be detected by detecting the magnetism from the magnetic material.However, when the degree of magnetization of the magnetic material is various, the degree of the side of the magnetic detector, that is, It was impossible to detect whether the vehicle had passed the distance.

[0003] To solve this problem, Japanese Patent Application Laid-Open No. Sho 55-1
No. 62078 has been proposed. That is, this magnetic body detection method using a three-axis magnetic detector is based on a three-axis magnetic detector of a rectangular coordinate system (X, Y, Z) having the Z axis as a vertical axis.
The ratio C = Hv / Hh of the maximum magnetic component (absolute value) Hv in the Z-axis direction of the magnetic target moving on the XY plane having a constant axial direction and the magnetic component Hh on the XY plane at that time is calculated. By comparing with a threshold value, it is detected how far the side of the magnetic detector has passed.

[0004]

However, the maximum magnetic component (absolute value) Hv in the Z-axis direction of the magnetic material and XY at that time
In a conventional method for detecting a magnetic substance, a ratio C = Hv / Hh with respect to a plane magnetic component Hh is taken, and the ratio C is compared with a threshold value.
There is a drawback that the magnetic measurement at the time of passing right above, for example, the determination of a reference value by prior measurement is required, and when the magnetic component Hh on the XY plane is 0, the ratio C = Hv / Hh becomes infinite. However, there is a disadvantage that the magnetic conditions, that is, the inclination and the declination of the magnetic efficiency of the magnetic target affect the accuracy of the detection result.

SUMMARY OF THE INVENTION An object of the present invention is to solve the above-mentioned drawbacks, does not require a prior magnetic measurement, has no processing in which infinity occurs in the course of arithmetic processing, and has a magnetic property such as the size of a hull and geomagnetism. It is an object of the present invention to provide a method and a device for detecting a passing position of a magnetic target which is hardly affected by magnetic conditions of a body.

[0006]

In order to solve the above-mentioned object, a passing position detecting method according to the present invention is directed to a method of detecting a passing position on each axis of a rectangular coordinate system (X, Y, Z) having a Z axis as a vertical axis. A magnetic target having a fixed position and a three-axis magnetic detector having a fixed position, and a magnetic target moving in an XY plane having a constant Z-axis direction with respect to the three-axis magnetic detector. In a passing position detection method for detecting passing within a given range, a combined magnetic signal value of a magnetic target is obtained from each sampled detection signal of the three-axis magnetic detector, and a timing at which the maximum value is obtained, From the absolute value | Hz | of the detection signal of the Z-axis magnetic sensor for each sampling signal,
A value Hn ₂ (i) obtained by subtracting the horizontal component value √ (Hx ² + Hy ² ) obtained from the detection signals of the X-axis magnetic sensor and the Y-axis magnetic sensor.
= | Hz | −√ (Hx ² + Hy ² ) (i = 1, 2,...,
n) is calculated and the result of the calculation is stored in the storage means. When the composite magnetic signal value of the magnetic target becomes maximum, the reversely stored data is stored in the storage means in order from the most recently stored in the storage means. A value S (i) = a predetermined number of stored values Hn ₂ divided by the maximum value H _t0 of the composite magnetic signal, respectively.
Seeking Hn ₂ / H _t0, this values _{S (i) = Hn 2 /} H t0,
Performing neural network arithmetic processing based on a predetermined weight value and performing waveform identification, and detecting the passage of a magnetic target within a predetermined range from the arithmetic value of the waveform identification. Features.

In the passing position detecting device of the present invention, the position at which the magnetic sensors having directivity are arranged on each axis of the orthogonal coordinate system (X, Y, Z) having the Z axis as the vertical axis is determined. A passing position detecting device including a fixed three-axis magnetic detector and detecting the passage of a magnetic target that moves on an XY plane in which the Z-axis direction is constant with respect to the three-axis magnetic detector within a predetermined range. A magnetic level detector that generates an output when detecting a magnetic signal of a predetermined level or more,
A sampling circuit for sampling each detection signal of the three-axis magnetic detector with an output of the magnetic level detector; and for each sampling signal output by the sampling circuit, a composite magnetism of a magnetic target object from each detection signal of the three-axis magnetic detector. A magnetic signal synthesizing means for obtaining a signal, a maximum value detecting means for detecting a maximum value H _t0 of the synthetic magnetic signal based on the synthetic magnetic signal obtained by the magnetic signal synthesizing means, and a sampling signal output by the sampling circuit. From the absolute value | Hz | of the detection signal of the Z-axis magnetic sensor, the horizontal component value √ (Hx ² + Hy ² ) obtained from the detection signals of the X-axis magnetic sensor and the Y-axis magnetic sensor
Hn ₂ (i) = | Hz | −√ (Hx ² + Hy ² )
(I = 1, 2,..., N), storage means for storing the calculation result of the calculation means, and when the maximum value detection means detects the maximum value H _t0 of the composite magnetic signal, A predetermined number of each value Hn ₂ stored in the storage unit in reverse order is read out in order from the most recently stored in the storage unit, and this value Hn ₂ is read out of the composite magnetic signal detected by the maximum value detection unit. Calculating means for executing the calculation Hn ₂ / H _t0 divided by the maximum value H _t0 , and each calculated value S (i) = Hn ₂ calculated by the calculating means
/ H _t0 and a waveform identification processing means for performing a neural network arithmetic processing based on a predetermined weighting value to identify a waveform, and comparing the arithmetic value processed by the waveform identification processing means with a predetermined reference value And comparing means.

[0008] Here, referring to FIG.
17 (I), the course of the magnetic target 15 is in the north direction, and (II) is the direction of the magnetic target 15. The upper part of each of (I) and (II) shows a case where the course is heading in the south direction, and the lower part of FIG. Represents the magnetic substance target 15 when viewed from the side.

The permanent component H of the hull magnetism of the magnetic target 15
_P ((a) in FIG. 17 (I)), and the induction component guided to the magnetic target 15 when the direction and the magnitude of the geomagnetism are represented by dotted lines are H _I (FIG. 17 (I)). (B))
When the hull magnetic magnetic body target 15 is determined as the synthetic components, FIG. 17 (I) (iii) H _P +
H _I (H _P and H _I are both vectors).

On the other hand, the course of the magnetic target 15 is shown in FIG.
If toward the opposite direction of the south direction (I), permanently component hull magnetic magnetic body target 15 H _P (FIG. 17 (II)
Of (b)), since the geomagnetic direction and its magnitude at that position is not changed as dotted lines shown, the hull magnetic magnetic body target 15 at this time, the permanent components of the hull magnetic H _P
((A) of FIG. 17 (II)) and its induction component H _I ((b) of FIG. 17 (II)) induced to the magnetic substance target 15 and become a composite component. (C)
The H _P + H _I.

In general, the magnitude of the hull magnetism of the magnetic target 15 is proportional to the size of the magnetic target 15, that is, the size of the hull. However, FIG.
As is apparent from (II), the magnitude of the Z-direction component becomes the same regardless of the direction of travel, and the travel direction of the magnetic target 15 remains unchanged if the Z component is replaced. ₂ = | Hz | -√ (Hx ² + H
y ² ) has no relation to the course direction of the magnetic target 15.

Therefore, in the present invention, the hull magnetism is displayed as a magnetic dipole moment (magnetic moment), and when each detection signal detected by the three-axis magnetic detector is processed, it is independent of the size of the hull. The value Hn ₂ (i) = | Hz | −√
(Hx ² + Hy ² ) (i = 1, 2,..., N) is divided by the maximum value H _t0 of the composite magnetic signal to make the dimensionless, and then the waveform is determined by the waveform identification processing using the neural network operation. Since the target is identified and compared with a reference value that is within a certain range from the three-axis magnetic detector, a magnetic target whose magnetic condition is unknown passes within a certain range from the three-axis magnetic detector. Regardless of the size of the magnetic material,
Further, even if the magnetic condition is unknown, it can be reliably detected.

[0013]

FIG. 1 shows the configuration of an embodiment of a passing position detecting device according to the present invention. In FIG. 1, a three-axis magnetic detector 1 has a rectangular coordinate system (X,
X-axis magnetic sensor having directivity on each of Y, Z) axes 2,
A Y-axis magnetic sensor 3 and a Z-axis magnetic sensor 4 are provided. As these magnetic sensors, for example, a flux gate type or a ring core type is used.

When the three-axis magnetic detector 1 detects a magnetic signal of a predetermined level or more, a magnetic level detector 5 for generating an output is provided, and sampling is performed when the magnetic level detector 5 generates an output. The circuit 6 is driven, and the sampling circuit 6 outputs a sampling signal.

The X-axis magnetic sensor 2, the Y-axis magnetic sensor 3, and the Z-axis magnetic sensor 4 are provided with corresponding A / D converters 7 for digitizing the respective detected signals. X-axis magnetic sensor 2, Y-axis magnetic sensor 3, and Z-axis magnetic sensor 4 digitized by A / D converter 7
Is input to the magnetic signal synthesizing means 8 and the calculating means 9.

Now, as shown in FIG. 3, the magnetic moment of the magnetic target 15 is represented by M, and the inclination and declination of the magnetic moment M are represented by β and γ, respectively, as shown in FIG. Detector 1
On the other hand, the detection signal detected by the X-axis magnetic sensor 2 is Hx, and the detection signal detected by the Y-axis magnetic sensor 3 is Hx. Assuming that the detection signal detected by the Hy and Z axis magnetic sensors 4 is Hz, the magnetic signal synthesizing means 8 generates a composite magnetic signal value √ (Hx ^{2) of the} magnetic target 15 based on the detection signals of the respective magnetic sensors. + Hy ² + Hz ² ).

When the output signal of the magnetic level detector 5 is received by the sampling circuit 6, the sampling circuit 6 sends the sampling signal to the magnetic signal synthesizing means 8 and the arithmetic means 9, and the magnetic signal synthesizing means 8 and the arithmetic means 9 The above A / D
Each detection signal Hx, Hy, digitized by the converter 7
Hz.

The magnetic signal synthesizing means 8 calculates, for each sampling signal output from the sampling circuit 6, the above-mentioned synthesized magnetic signal value か ら from each of the captured detection signals Hx, Hy and Hz.
(Hx ² + Hy ² + Hz ² ) is calculated, and the maximum value detection means 10 provided at the next stage is used to calculate the composite magnetic signal value √ (Hx ² + Hy ² +
Hz ² ).

The calculating means 9 is provided for each sampling signal output from the sampling circuit 6 and for each detection signal H taken in.
The horizontal component value √ (Hx ² + H) obtained from the detection signals of the X-axis magnetic sensor 2 and the Y-axis magnetic sensor 3 from the absolute value | Hz | of the detection signal of the Z-axis magnetic sensor 4 based on x, Hy, and Hz.
The value Hn ₂ obtained by subtracting the y ^2), i.e. _{Hn 2 (i) = | Hz} | -
The operation of √ (Hx ² + Hy ² ) (i = 1, 2,..., N) is performed, and the operation result Hn ₂ (i) is sequentially stored in the memory 11.

The maximum value detecting means 10 generates a composite magnetic signal value Ht (k) = √ (H
x ² + Hy ² + Hz ² ) and the previously captured combined magnetic signal value Ht (k−1) = √ (Hx ² + Hy ² + Hz ² ) are sequentially compared, and the combined magnetic signal value √ (Hx ² + Hy) is compared. and it obtains the maximum value H _t0 of ² + Hz ^2).

The maximum value detecting means 10 outputs the composite magnetic signal
Value √ (Hx^Two+ Hy^Two+ Hz^Two) Maximum value H_t0When asked for
That is, the maximum value detecting means 10 determines the maximum value of the composite magnetic signal.
Value H _t0Is detected, the maximum value detecting means 10
Maximum value of magnetic signal H_t0Signal of the detection of
Maximum value of air signal H_t0Are sent to the calculating means 12.

The calculating means 12 reverses the values Hn _{2 of the} result of the calculation by the calculating means 9 stored in the memory 11 by a predetermined number N in order from the most recently stored one to the previous one. The values are read out in order and
n ₂ the turn performs Hn ₂ / H _t0 calculation dividing the maximum value H _t0 of the detected resultant magnetic signal with a maximum value detecting means 10, the respective calculated values _{S (i) = Hn 2 /} H t0 (i = ,..., N) are input to the waveform identification processing means 13.

The respective values Hn ₂ are converted to the maximum value H _{t0 of the} composite magnetic signal.
By dividing by, the dimensionless magnetic material model can be obtained, and the magnetic material target 15 can be handled regardless of the magnitude of the magnetic efficiency M.

In the waveform discriminating means 13, the calculating means 12
Based on a predetermined number N values Hn ₂ entered the predetermined weighting value performs neural network processing from the waveform discrimination processing for waveform identification operation consisting of product and sum. The calculation result value processed by the waveform identification processing means 13 is output to the comparison means 14. This waveform identification processing means 13 will be described in detail with reference to FIG.

The comparison means 14 determines how far the side of the three-axis magnetic detector 1 should pass to reach the magnetic target 1
5 is set in advance as a reference value, and the calculated value processed by the waveform identification processing unit 13 is compared with the comparison value by the comparison unit 14.
Are compared. When the calculated value processed by the waveform identification processing means 13 is larger than the reference value, “1” meaning that the magnetic substance target 15 passes within a predetermined range is set to the comparison means 14.
When the calculated value output from the waveform identification processing means 13 is smaller than the reference value, “0” indicating that the magnetic target 15 passes outside a predetermined range is set to the comparison means 14.
Output from

FIG. 2 is a conceptual explanatory diagram of one embodiment of the waveform discrimination processing means, in which a neural network is used as the waveform discrimination processing means 13. In the figure, the neural network of the waveform identification processing means 13 is composed of N input units 21-1 to 21-N, N hidden units 22-1 to 22-N, and one output unit 23. I have. And input unit 2
1-1 to 21-N, when the maximum value detecting means 10 detects the maximum value _Ht0 of the composite magnetic signal as described with reference to FIG. Going back up in order, a value S (i) = a predetermined number N of values Hn ₂ divided by the maximum value H _t0 of the composite magnetic signal =
Hn ₂ / H _t0 (i = 1, 2,..., N) are input, respectively.

When the maximum value detecting means 10 detects the maximum value H _t0 of the composite magnetic signal, and the value of the operation result most recently stored in the memory 11 is Hn ₂ (t ₁ ), the calculating means 12 The calculated value S (t ₁ ) = H
n ₂ (t ₁ ) / H _t0 is input to the input unit 21-1;
The next newest operation result value S (t ₂ ) in the memory 11 =
Hn ₂ (t ₂ ) / H _t0 is inputted to the input unit 21-2, and the next new operation result value S in the memory 11 is obtained.
(T ₃ ) = Hn ₂ (t ₃ ) / H _t0 is the input unit 21-3
Is input, the following Similarly the operation result value S (t _N) =
Hn ₂ (t _N ) / H _t0 is input to the input unit 21-N.

The input units 21-1 to 21-N are hidden.
Between the units 22-1 to 22-N as shown in the figure.
A₁, A_Two, A_Three............ B₁, B_Two, B_Three……,
C₁, C _Two, C_ThreeWeighted by weighting values such as
Each is linked, and the hidden unit 22-1
, And between the output unit 23 and the output unit 23 as shown in FIG.
K₁, M_Two, M_ThreeWeighted by the weight value of ......
Linked.

The values S (t ₁ ) to S
By performing weighting processing using a neural network having (t _N ) as an input, that is, performing a neural network operation consisting of a product and a sum, an output value of a certain value is output from the output unit 23.

The above used in the neural network
Weight A of₁, A_Two, A_Three............ B ₁, B_Two, B_Three…
…, C₁, C_Two, C_Three…… etc and M₁, M_Two, M_Three......
Are typical magnetic models as shown in FIGS.
Hn in le_Two/ H_t0Is determined from the waveform pattern.
You.

FIGS. 5 to 14 show Hn ₂ / _Ht0 waveform diagrams simulated by calculation with changing conditions. In FIGS. 5 to 14, the horizontal axis represents the distance from the position at which the composite magnetic signal value Ht = √ (Hx ² + Hy ² + Hz ² ) of the magnetic substance target becomes the maximum, that is, from the position of the closest point P in FIG. Distance opposite to the direction of travel of the body target 15 (−
L) represents the vertical axis represents the ratio of H _t0 of Hn _2. 5 to 9 show that the lateral distance W when the approach angle α = 0 ° and the magnetic conditions of the inclination (β) and the declination (γ) of the magnetic moment of the magnetic substance target 15 are the same and the approach angle α = 0 ° is 0. , 10, 2
FIGS. 10 to 14 show that the magnetic condition of the magnetic efficiency of the magnetic target 15 is equal to the inclination (β) and the angle of deviation (γ), and the approach angle α = 0 °
(Α is shown in FIG. 15), the horizontal distance W is 0, 10, 2
It was changed to 0, 30, and 40 m. 5 to 9
Correspond to FIGS. 10 to 14 in a state where the inclination (β) is changed.

FIG. 5 shows the case where the side (lateral distance W) of the three-axis magnetic detector 1 is 0 m, that is, the magnetic target 1.
5 when passing just above the three-axis magnetic detector 1,
FIG. 6 shows the case where the side (lateral distance W) of the three-axis magnetic detector 1 is 10 m, and FIG. 7 shows the side (lateral distance W) of the three-axis magnetic detector 1.
FIG. 8 shows the case when the side (lateral distance W) of the three-axis magnetic detector 1 is 30 m, and FIG. 9 shows the case when the side (lateral distance W) is 40 m of the three-axis magnetic detector 1. Things.

FIG. 10 shows a case where the side (lateral distance W) of the three-axis magnetic detector 1 is 0 m, that is, when the magnetic target 15 passes right above the three-axis magnetic detector 1. FIG. 11 shows the side (lateral distance W) 10 of the three-axis magnetic detector 1.
m, FIG. 12 shows a case where the three-axis magnetic detector 1 is at a side (lateral distance W) of 20 m, and FIG. 13 shows a case where the three-axis magnetic detector 1 is at a side (lateral distance W) of 30 m. 14 in FIG.
This is the case when the side (lateral distance W) of the axial magnetic detector 1 is 40 m.

FIG. 16 is a flowchart of one embodiment of the present invention.
The operation of FIG. 1 is described with reference to FIG.
This will be described next. The waveform identification processing means 13 includes
Hn in typical magnetic model_Two/ H_t0From the waveform pattern
Each weight A₁, A_Two, A_Three............ B₁, B_Two, B_Three…
…, C ₁, C_Two, C_Three…… etc and M₁, M_Two, M_Three……But
It is set in advance. The comparison means 14 has a three-axis magnetic detection
The distance up to the side of the vessel 1
Determines whether to pass the target 15 within a predetermined range
The reference value is set in advance.

Under these conditions, the three-axis magnetic detector 1
In contrast, the magnetic target 15 moving at a constant speed on the XY plane in which the Z-axis direction is constant approaches the three-axis magnetic detector 1 (step 1), and the magnetic level detector 5 moves the X of the three-axis magnetic detector 1 to X. When a magnetic signal of a predetermined level or more is detected in any of the detection signals of the axis magnetic sensor 2, the Y axis magnetic sensor 3, and the Z axis magnetic sensor 4 (Step 2), the magnetic level detector 5 Generate output. The sampling circuit 6 is driven by the output of the magnetic level detector 5, and the sampling circuit 6 outputs a sampling signal.

Each detection signal detected by the X-axis magnetic sensor 2, the Y-axis magnetic sensor 3, and the Z-axis magnetic sensor 4 is
The detection signals of the X-axis magnetic sensor 2, the Y-axis magnetic sensor 3, and the Z-axis magnetic sensor 4 are digitized by corresponding A / D converters 7, respectively. Each time, it is taken into the magnetic signal synthesizing means 8 and the arithmetic means 9 (step 3).

In the magnetic signal synthesizing means 8, based on the detection signals Hx, Hy, and Hz of the respective magnetic sensors, the magnetic target 15
The composite magnetic signal value √ (Hx ² + Hy ² + Hz ² ) of
In the arithmetic means 9, each detection signal Hx, Hy,
Hz, the absolute value of the detection signal of the Z-axis magnetic sensor 4 | H
A value Hn ₂ obtained by subtracting the horizontal component value √ (Hx ² + Hy ² ) obtained from the detection signals of the X-axis magnetic sensor 2 and the Y-axis magnetic sensor 3 from z |, that is, Hn ₂ = | Hz | −√ (Hx ² + H
y ² ), and stores the result Hn ₂ of the result in the memory 11
(Step 4).

The maximum value detecting means 10 generates a composite magnetic signal value Ht (k) = √ (H
x ² + Hy ² + Hz ² ) is compared with the previously acquired combined magnetic signal value Ht (k−1) = Ｈ (Hx ² + Hy ² + Hz ² ), and the combined magnetic signal value √ (Hx ² + Hy ^2). + Hz ² ) maximum value H
Perform processing to find _t0 . When the maximum value detecting means 10 detects the maximum value _Ht0 of the composite magnetic signal (step 5),
The maximum value detection means 10 sends to the calculation means 12 a signal indicating that the maximum value H _t0 of the composite magnetic signal has been detected and the maximum value H _t0 of the composite magnetic signal. The calculating means 12 reads the value Hn ₂ of the result of the calculation by the calculating means 9 stored in the memory 11 in reverse order from the most recently stored one by a predetermined number N from the memory 11 in order (step 6). ), And dividing the N values Hn ₂ (t ₁ ) to Hn ₂ (t _N ) by the maximum value H _t0 of the composite magnetic signal to obtain S (i) = Hn ₂ / H
_{The value of t0} (i = 1, 2,..., N) is determined (step 7), and the N calculated values S (i) are obtained by the waveform identification processing means 1.
3 is input.

In the waveform discriminating means 13, the calculating means 12
N values S (t ₁ ) to S (t _N ) calculated by
And the weighting values A ₁ , A ₂ , A ₃
.., B ₁ , B ₂ , B _3, ..., C ₁ , C ₂ , C _3, etc. and M ₁ , M ₂ , M ₃ ,. (Step 8).

Here, the measured value S (t ₁ ) to the value S inputted from the calculating means 12 to the waveform identification processing means 13 are shown.
When the waveform discriminating means 13 performs the waveform discriminating process of the neural network operation based on (t _N ), the learning skill previously learned before the three-axis magnetic detector 1 is arranged at a predetermined position on the sea floor. Is performed to predict what kind of waveform is formed, and the resulting lateral distance W of the three-axis magnetic detector 1 is output in a dimensionless form.

The operation value processed by the waveform discrimination processing means 13 is a dimensionless one, so that an operation for returning to the original size is executed. That is, based on the calculated value processed by the waveform identification processing means 13, Hn ₂ = η × _Ht0 is calculated from the dispersion ratio η = calculated value / sensor measurement range, and is returned to the original size.

On the other hand, when the speed of the magnetic target 15 obtained by another measuring means is V and the measuring time is t, the distance L from the point P shown in FIG. 15 is L = V × t From the measurement time t, the time (timing) at which the magnetic substance target 15 passes the side of the three-axis magnetic detector 1 is obtained, and the side of the three-axis magnetic detector 1 at that time is obtained. Is determined.

Then, the comparison means 14 sets a reference value for determining how far the side of the three-axis magnetic detector 1 should pass through within a predetermined range of the magnetic target 15. Is compared with (step 9). When the calculated value processed by the waveform discrimination processing means 13 is larger than the reference value, "1" indicating that the magnetic target 15 passes within a predetermined range is output from the comparison means 14 (step 10). If the calculated value processed by the waveform identification processing means 13 is smaller than the reference value, "0" indicating that the magnetic target 15 passes outside a predetermined range is output from the comparison means 14 (step). 11).

In step 5, the composite magnetic signal value √ (Hx ²
+ Hy ² + Hz ²⁾ is at not the maximum value H _t0, the step 3 to step 5 to a maximum value H _t0 of the resultant magnetic signal is detected is repeated.

In addition, the magnetic level detector 5 detects a magnetic signal of a predetermined level or more even when the detection signal of any of the X-axis magnetic sensor 2, the Y-axis magnetic sensor 3, and the Z-axis magnetic sensor 4 of the three-axis magnetic detector 1 is detected. When the signal cannot be detected (step 2), the three-axis magnetic detector 1 in which the magnetic target is 15 fixed points
Is processed as if it does not pass within a certain range from, and the process ends.

In the above description, since each of the calculated values S (t ₁ ) to S (t _N ) processed by the waveform discriminating means 13 is made dimensionless, it is reduced to the original size. Although the return operation is performed, the following operation can be performed under the dimensionless state as it is.

That is, the waveform discriminating process is performed by the calculating means 12.
The measured value S (t ₁) Or value S (t
_N), The waveform identification processing means 13
3-axis magnetic detector when performing waveform discrimination processing of work calculation
Learned before 1 was placed in place on the sea floor
What kind of waveform is formed by learning based on learning skill
Or three-axis magnetic detection as a result.
The lateral distance W on the side of the output unit 1 is output in a dimensionless form.
You.

The output and the distance to the side of the three-axis magnetic detector 1 at the dimensionless ratio within a predetermined range of the magnetic target 15 are determined. The reference value corresponding to the non-dimensionalization that determines whether or not to pass (the reference value corresponding to the non-dimensionalization at this time is a value obtained by dividing the reference value of the non-dimensionalization ratio by the maximum value _Ht0 ) is compared with the comparison means. A comparison is made at 14 (step 9). Similarly, when the above calculated value processed by the waveform identification processing means 13 is larger than the reference value corresponding to the non-dimensionalization, "1" meaning that the magnetic substance target 15 passes through a predetermined range. Is output from the comparison means 14 (step 10), and when the operation value processed by the waveform identification processing means 13 is smaller than the reference value, "0" meaning that the magnetic target 15 passes outside a predetermined range.
Is output from the comparison means 14 (step 11).

[0049]

As described above, according to the present invention, a magnetic target is displayed as a magnetic dipole moment, and then the detected magnetic signal is subjected to arithmetic processing to perform dimensionless processing. An Hn ₂ / _Ht0 waveform, which is irrespective of the size of the target and eliminates the influence of the difference in the direction of approach of the magnetic target to the magnetic sensor on the detection of the passing position, as an element for waveform identification; create different magnetic conditions of magnetic pattern in a magnetic model assuming the (dip and declination of magnetic moment) and the depth, i.e. Hn ₂ / H _t0 waveform, using a weight determined from this, and a product and sum The calculated value is obtained by a neural network calculation and compared with a reference value, so that it is determined whether a magnetic target whose magnetic condition is unknown has passed within a certain range from a fixed point magnetic detector. Magnetic dipole Regardless of the moment, also the magnetic condition can be detected reliably even if unknown. Then, it is possible to detect a passing position without requiring a prior magnetic measurement and hardly affected by magnetic conditions of the magnetic target.

[Brief description of the drawings]

FIG. 1 is a configuration of an embodiment of a passage position detecting device according to the present invention.

FIG. 2 is a conceptual explanatory diagram of one embodiment of a waveform identification processing means.

FIG. 3 is an explanatory view of a dip angle and a declination angle of a magnetic dipole moment M of a magnetic target.

FIG. 4 is an explanatory diagram of a positional relationship between a magnetic target and a three-axis magnetic detector in a rectangular coordinate system (X, Y, Z).

FIG. 5 is a Hn ₂ / _Ht0 waveform diagram simulated by calculation.

FIG. 6: Hn ₂ simulated by calculation with changing conditions
FIG. _{14 is a} / _Ht0 waveform diagram.

FIG. 7: Hn ₂ simulated by calculation with changing conditions
FIG. _{14 is a} / _Ht0 waveform diagram.

FIG. 8: Hn ₂ simulated by calculation with changing conditions
FIG. _{14 is a} / _Ht0 waveform diagram.

FIG. 9: Hn ₂ simulated by calculation with changing conditions
FIG. _{14 is a} / _Ht0 waveform diagram.

FIG. 10 shows H simulated by calculation with changing conditions.
It is n ₂ / H _t0 waveform.

FIG. 11 is a graph simulating H by changing conditions.
It is n ₂ / H _t0 waveform.

FIG. 12 is a graph simulating H by changing conditions.
It is n ₂ / H _t0 waveform.

FIG. 13 is a graph simulating H by changing conditions and calculating.
It is n ₂ / H _t0 waveform.

FIG. 14 is a graph simulating H by changing conditions.
It is n ₂ / H _t0 waveform.

FIG. 15 is an XY plane explanatory view illustrating a traveling direction and coordinates of a magnetic target on the XY plane.

FIG. 16 is a flowchart of one embodiment of the present invention.

FIG. 17 is a diagram illustrating a Z-axis direction component of the magnetic efficiency between the traveling direction of the magnetic target and the hull magnetism.

[Explanation of symbols]

 Reference Signs List 1 3-axis magnetic detector 2 X-axis magnetic sensor 3 Y-axis magnetic sensor 4 Z-axis magnetic sensor 5 Magnetic level detector 6 Sampling circuit 7 A / D converter 8 Magnetic signal synthesizing means 9 Computing means 10 Maximum value detecting means 11 Memory 12 Calculation means 13 Waveform identification processing means 14 Comparison means 15 Magnetic target

────────────────────────────────────────────────── ─── Continuing on the front page (51) Int.Cl. ⁷ Identification symbol FI // G01R 33/02 G01R 33/02 Q (72) Inventor Kunihiko Yoshimi 2-29 Heiancho, Tsurumi-ku, Yokohama-shi, Kanagawa 1 Kyosan Seisakusho Co., Ltd. (72) Inventor Haruyuki Morishita 2-29-2 Heian-cho, Tsurumi-ku, Yokohama-shi, Kanagawa 1 Kyosan Seisakusho Co., Ltd. (56) References JP-A-2-74892 (JP, A JP-A-56-162078 (JP, A) JP-A-60-71910 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) G01B 7/00 G01C 15/00 G01V 3 / 08 G01V 3/10 G01R 33/02 G01C 21/00

Claims (2)

(57) [Claims] 1. An orthogonal coordinate system (X,
A magnetic sensor having directivity on each axis of (Y, Z) is provided with a three-axis magnetic detector whose position is fixed, and an XY plane having a fixed Z-axis direction with respect to the three-axis magnetic detector; A passage position detecting method for detecting passage of a moving magnetic target within a predetermined range, wherein a combined magnetic signal value of the magnetic target is obtained from each sampled detection signal of the three-axis magnetic detector. , The timing at which the maximum value is obtained, and the horizontal component value √ () obtained from the detection signals of the X-axis magnetic sensor and the Y-axis magnetic sensor from the absolute value | Hz | of the detection signal of the Z-axis magnetic sensor for each sampling signal. Hx ² + Hy ² ) Hn ₂ (i) = | Hz | − (Hx ² + Hy ² ) (i
= 1, 2,..., N) and stores the result of the calculation in the storage means. When the composite magnetic signal value of the magnetic target reaches the maximum value, the storage means starts with the most recently stored value. A predetermined number of values Hn ₂ previously stored in the reverse storage means in order.
_Are divided by the maximum value H _t0 of the composite magnetic signal, respectively.
(I) = seeking Hn ₂ / H _t0, the each value _{S (i) = Hn 2 /} H t0, performs waveform identification by performing a neural network processing based on the weighting value set in advance, the waveform A passing position detecting method, wherein a passing of a magnetic target within a predetermined range is detected from a calculated value of the identification. 2. An orthogonal coordinate system (X,
A magnetic sensor having directivity on each axis of (Y, Z) is provided with a fixed three-axis magnetic detector at a fixed position, and an XY plane having a fixed Z-axis direction with respect to the three-axis magnetic detector; A passage level detection device that detects passage of a moving magnetic target within a predetermined range, a magnetic level detector that generates an output when a magnetic signal of a predetermined level or more is detected, and a magnetic level. A sampling circuit for sampling each detection signal of the three-axis magnetic detector at the output of the detector;
Magnetic signal synthesizing means for obtaining a synthetic magnetic signal of a magnetic target from each detection signal of the axial magnetism detector; and a maximum value _Ht0 of the synthetic magnetic signal based on the synthetic magnetic signal obtained by the magnetic signal synthesizing means. A maximum value detecting means for detecting, and a Z signal for each sampling signal output from the sampling circuit.
Hn ₂ (i) = | obtained by subtracting the horizontal component value √ (Hx ² + Hy ² ) obtained from the detection signals of the X-axis magnetic sensor and the Y-axis magnetic sensor from the absolute value | Hz | H
z |-を (Hx ² + Hy ² ) (i = 1, 2,..., n); a storage unit for storing the calculation result of the calculation unit; upon detection of the maximum value H _t0, reads out the storage unit most recently stored values Hn number of predetermined which are sequentially before the reverse uplink storage from those in _2, the maximum value the value Hn ₂ Calculation Hn ₂ / H _t0 divided by the maximum value H _t0 of the composite magnetic signal detected by the detection means
Means for performing a neural network operation process based on each of the calculated values S (i) = Hn ₂ / _Ht0 and a predetermined weight value, and performing waveform identification for performing waveform identification. A passing position detecting device comprising: a processing unit; and a comparing unit that compares a calculated value processed by the waveform identification processing unit with a predetermined reference value.