JP5097596B2 - Measuring device using line sensor - Google Patents

Description

  The present invention relates to a measuring device using a line sensor. For example, in the field of measuring the contact force generated between a pantograph and a trolley line using image processing, it can be applied to a technique for accurately obtaining the relative displacement between the hull and the sliding plate, which is particularly necessary to obtain the contact force It is a thing.

There is contact force to evaluate the contact performance between the trolley wire and the pantograph, which is an electric railway facility. A pantograph is one of the current collectors installed on the roof of an electric railway vehicle. Since the contact force is a physical quantity, the current collection performance can be objectively evaluated.
Further, since the contact force directly affects trolley wire wear, maintenance such as wear prediction by contact force measurement can be performed.

There are the following methods for measuring the contact force.
1. A method of obtaining a contact force by installing a sensor such as an accelerometer or a strain gauge on a pantograph and measuring a force (cross-sectional force) and an inertial force generated on a cross section of the pantograph (Non-patent Document 1).
2. A method of obtaining a contact force by measuring in advance an impulse response function of a pantograph with respect to contact force fluctuation and performing convolution integration of a time series response of the pantograph while traveling and the impulse response function (Patent Document 1).
3. A method of obtaining contact force by installing two light sources such as LEDs on the pantograph's spring portion, and obtaining the relative displacement by image processing, and obtaining the amount of expansion and contraction of the spring (Patent Document 2).

  In the method described in Patent Document 2, when the contact force is measured by image processing, it is necessary to accurately obtain the relative displacement. For this purpose, it is necessary to consider the inclination and rotation of the pantograph. As a method for obtaining the displacement of the pantograph by image processing, for example, there is a method described in Japanese Patent Application No. 2006-285902.

In this method, as shown in FIG. 11, the marker 4 is attached to the pantograph 3, the marker 4 is photographed by the line sensor 2 installed on the roof of the vehicle 1, and the position of the pantograph 3 is detected by pattern matching. Is a method of measuring.
In other words, the image acquired by the line sensor 2 is subjected to image processing by the image processing unit 6, and the height and acceleration of the pantograph 3 obtained as a result of the image processing are recorded in the recording device 7.

Here, the marker 4 has a striped pattern by disposing a strip-shaped reflective region 4a that easily reflects light and a strip-shaped non-reflective region 4b that hardly reflects light above and below the reflective region 4a. It is. As shown in FIG. 12, the marker 4 has two or more white band-like reflective regions 4a and is complicated in shape, thereby increasing the amount of features used for pattern matching and preventing erroneous detection and the like. It is also possible.
The scanning direction (line sensor imaging surface) of the line sensor 2 is a direction that crosses the reflective area 4a and the non-reflective area 4b of the marker 4, that is, the vertical direction.

FIG. 13 shows a flowchart of a pantograph measuring method by image processing, and FIG. 14 shows a configuration diagram of the pantograph measuring apparatus.
As shown in FIG. 14, the image processing unit 6 includes a space-time image creation unit 60, memories 61 and 62, a pattern matching unit 63, a pantograph position interpolation unit 64, an end process determination unit 65, and a height / acceleration output unit 66. The image processing by these, that is, the pantograph measurement method is processed by the following procedures (a) to (f) as shown in FIG.

(A) Space-Time Image Generation The video of the marker 4 attached to the pantograph 3 is acquired every unit time (every frame rate) by the line sensor 2, and the space-time image generator 60 generates the space-time image shown in FIG. In the space-time image, the horizontal axis represents time, and the vertical axis represents height (position). It is assumed that the marker 4 has one white band-like reflection area 4a. As shown in FIG. 15, according to the shape of the marker 4, a striped pattern composed of a black band A with little light reflection and a white band B with high light reflection is continuously photographed. Further, the background of the pantograph 3 such as a train line facility crossing the pantograph 3 at high speed appears as noise C.

(B) Pattern Matching When the size of the marker 4 and the direction of the camera are taken into consideration, the size of the marker 4 reflected by the position direction on the image is almost determined as shown in FIG. Therefore, a pattern corresponding to the pattern and height of the marker 4 is generated, and the pattern matching unit 63 scans the scanning line from the upper side to the lower side every unit time as shown in FIG. . If the pattern approximates more than a certain level, a plurality of records are recorded with an order from the most approximated position on the scanning line. That is, a plurality of records such as the first candidate and the second candidate are recorded. Also, if there is no approximate location, it is suspended.

(C) Pantograph Position Interpolation In pattern matching, the reserved position D and the non-connected non-connected position E as shown in FIG. 18 are linearly filled by the pantograph position interpolation unit 64 so as to fill the gap from the front and rear points. The position is interpolated.

(D) Process end determination The pattern matching and position interpolation processes described above are repeated for all space-time images. When the end process determination unit 65 determines that the process has ended, the pantograph position is recorded.

(E) Height output The height / acceleration output unit 66 outputs the recorded pantograph position as the height.
(F) Acceleration output The height / acceleration output unit 66 obtains the extreme value of the inflection point convex downward at the recorded pantograph position, and outputs an extreme value equal to or greater than a predetermined value as the acceleration.
Ikeda, Usuda: "Study on contact force measurement method between overhead wire and pantograph", RTRI REPORT Vol.14, No.6, pp7-12,2000.6 JP-A-11-194059 JP-A-200-235310

  As described above, the method for measuring the displacement of the pantograph 3 by attaching the marker 4 to the pantograph 3 and detecting the marker position photographed by the line sensor 2 has the following problems (i) and (ii). .

(I) If the line sensor 2 is tilted with respect to the marker 4 when the marker 4 attached to the pantograph 3 is photographed by the line sensor 2, accurate displacement cannot be measured. This is because when the marker width is used as the calibration value X [mm / pix], an accurate calibration value cannot be obtained if the line sensor 2 is tilted.

  For example, as shown in FIG. 19, consider a case where a calibration value is obtained from a marker 4 having a marker width of 2b [mm] when the scanning direction of the line sensor 2 is tilted from the vertical direction instead of the vertical direction. . Here, it is assumed that an angle φ of a trajectory (hereinafter referred to as an imaging line) that the line sensor 2 scans across the marker 4 is measured from the horizontal direction.

When edge processing is performed and the width of the marker 4 of 2b is measured by the line sensor 2 as x _p [pix], that is, when the number of pixels is detected as x _p , the calibration value is 2b / x _p [ mm / pix].
However, since the photographing line is actually inclined at an angle φ, the correct calibration value is 2b / x _p sinφ [mm / pix], and an error occurs accordingly.

(Ii) When the relative displacement is measured using two or more markers, the error is more complicated than the error of (1), and it is difficult to measure an accurate displacement.
For example, in order to measure the relative displacement of the pantograph, that is, the relative displacement between the boat body and the sliding plate body, markers are respectively attached to the boat body and the sliding plate body, and the distance between the centers of the two markers is measured by the line sensor 2. Shall.
As shown in FIG. 20A, when one marker simply translates without rotational movement with respect to the other marker, the output value x _pr [pix] from the line sensor is added to the calibration value 2b / x _p [mm / pix] is multiplied and measured to be 2bx _pr / x _p [mm], but since the photographing line is actually inclined by an angle φ from the horizontal direction, it is correct 2bx _pr / x _p sin φ [mm]. Also in this case, the error increases as 1 / sinφ increases.

Furthermore, as shown in FIG. 20 (b), the rotation of one marker with respect to the other marker by an angle ψ (in the case of subscripting due to restrictions on the PC application software, substitute the alphabet “psi”) When moving with movement, the measurement error becomes even more complex. The exact center-to-center distance at this time is 2bx ′ _pr / x ′ _p sinφ−l _psi tanψ. l _psi c, the horizontal distance between the two markers after movement.

  The present invention has been made in view of the above-described prior art, and corrects the inclination of the photographing line by using a specially shaped marker even when the photographing line of the line sensor is inclined with respect to the marker. It is another object of the present invention to provide an apparatus capable of accurately calculating the displacement of a measurement object to which a marker is attached.

A measuring apparatus using a line sensor according to claim 1 of the present invention that solves the above-described problem is an image of a marker attached to an object to be measured, a line sensor that images the marker, and an image acquired by the line sensor. In the measurement apparatus including an image processing unit that processes and calculates the position of the object to be measured, the marker has a striped pattern formed by arranging a plurality of adjacent strip-shaped regions of different colors, A straight line that is a boundary between the regions includes at least two parallel straight lines that are parallel to each other; and at least two inclined straight lines that are inclined with respect to the parallel straight line, and the image processing unit includes: with three long less of line segments between a plurality of intersections of said straight line parallel and the inclined straight line and said imaging line line sensor is scanned trajectory so as to cross the region of the marker It obtains the ratio, and obtains at least two coordinates of the gradient or the intersection with respect to the marker of the shooting line based on a ratio of the three lengths.

A measuring apparatus using a line sensor according to a second aspect of the present invention that solves the above-described problem is the measuring apparatus according to the first aspect , wherein the image processing unit is based on at least two coordinates of an inclination of the photographing line or the intersection. Then, an actual distance between the intersections is obtained, and a calibration value is obtained by dividing the distance by the number of pixels between the intersections on the line sensor.

The measurement apparatus using a line sensor according to a third aspect of the present invention for solving the above-described problems is the measurement apparatus according to the first aspect , wherein the image processing unit is configured to detect the inclination of the photographing line or the intersection before and after the measurement object is displaced. The relative displacement of the object to be measured is obtained by comparing two coordinates.

A measuring apparatus using a line sensor according to a fourth aspect of the present invention that solves the above-described problems is the measuring device according to the first aspect , wherein the image processing unit is configured to detect an inclination of the photographing line before and after the measured object is displaced with rotation or By comparing the two coordinates of the intersection, a displacement accompanied by a relative rotation of the object to be measured is obtained.

A measuring apparatus using a line sensor according to a fifth aspect of the present invention for solving the above-mentioned problems is the measuring apparatus according to the first aspect, wherein the object to be measured is held on the vehicle so as to be movable up and down via a link structure frame. A pantograph comprising a boat plate and a sliding plate supported via a spring against the boat plate and in contact with a trolley wire, wherein the markers are respectively attached to the boat plate and the sliding plate. To do.

A measuring apparatus using a line sensor according to a sixth aspect of the present invention that solves the above-described problem is the measurement apparatus according to the fifth aspect , wherein the image processing unit is configured to capture the image of the markers attached to the boat board and the sliding board, respectively. wherein obtaining each inclination of line.

The measuring device using a line sensor according to claim 7 of the present invention that solves the above-described problem is the measuring device according to claim 5 , wherein the image processing unit is connected to the markers respectively attached to the boat board and the sliding board. Each of the coordinates is obtained.

A measuring device using a line sensor according to an eighth aspect of the present invention for solving the above-mentioned problems is the measurement device according to the sixth or seventh aspect , wherein the image processing unit is attached to the boat board and the sliding board, respectively. Based on the inclination of the photographing line with respect to the marker or the coordinates of the intersection point, the relative position of the sliding plate with respect to the boat plate is obtained, and the expansion and contraction of the spring sandwiched between the boat plate and the sliding plate from the relative position it characterized to determine the amount.

A measuring apparatus using a line sensor according to a ninth aspect of the present invention for solving the above-mentioned problems is that, in any one of the first aspect, the width of the measuring device is uniformly expanded to both sides as it goes in one longitudinal direction. The two regions that are in contact with both sides of the region have a shape in which the width is uniformly reduced only inward as it goes in the longitudinal direction of one of the regions, thereby forming a boundary between the region and the two regions. These two straight lines are inclined in opposite directions.

A measuring apparatus using a line sensor according to a tenth aspect of the present invention for solving the above-described problems is the measurement apparatus according to the first aspect, wherein any one of the regions has a parallelogram shape and the other region in contact with both sides. The inclined straight lines forming a boundary are parallel to each other .

The present invention has the following effects by complicating the shape of the marker.
(A) It is possible to correct the inclination of the photographing line of the line sensor, and to calculate the displacement with high accuracy.
(B) Since it can be seen how much the photographing line of the line sensor is inclined, it can be estimated from the photographing line where the origin is.
(C) The displacement can be calculated with high accuracy even if the pantograph moves in the horizontal direction in a state where the photographing line of the line sensor is inclined.
(D) Even if the pantograph moves with rotation while the photographing line of the line sensor is tilted, the displacement can be calculated with high accuracy.
(E) By attaching markers to the pantograph's hull and sliding plate and correcting the inclination of the line sensor, the relative displacement can be accurately calculated, and the amount of expansion / contraction of the pantograph's spring can be obtained with high accuracy.

  The best mode for carrying out the present invention will be described with reference to FIGS. The color of the marker in the present invention includes an achromatic color consisting only of lightness belonging to the white, gray and black systems, and a chromatic color consisting of three elements such as hue, lightness and saturation such as red, yellow and green. Including. Therefore, in the present invention, not only a monochrome CCD but also a color CCD can be used as the line sensor. In the following description, a case where the color of the marker is an achromatic color, that is, a case where regions where light is easily reflected and regions where light is difficult to reflect is alternately arranged will be described as an example.

(1) Inclination Correction of Line Sensor FIG. 1 is an example of an inclination correction marker 14 attached to a measurement target in the present invention. The marker 4 is assumed to be attached horizontally or at a known angle.
As shown in FIG. 1, the inclination correcting marker 14 includes two strip-shaped reflective regions 14b and 14c that are likely to reflect light between three strip-shaped non-reflective regions 14a, 14d, and 14e that are difficult to reflect light. The shape of the non-reflective region 14a disposed in the center is a trapezoidal shape.

  That is, the central non-reflective region 14a has a shape in which the width is uniformly expanded to both sides (+ y-axis direction and -y-axis direction) as it goes in one longitudinal direction (-x-axis direction), and when x = -L. Is larger than the width 2a when x = L (a <c). In the figure, the origin of the xy coordinate system is placed at the center of the central non-reflective region, the x axis is taken in the horizontal direction, and the y axis is taken in the up and down direction.

Therefore, a function of a straight line (hereinafter referred to as a boundary line) that defines the boundary between the central non-reflective region 14a and the reflective region 14b in contact with the upper side thereof is y = (ac) x / 2L + (a + c). Will be given by / 2.
Similarly, the function of the boundary line that defines the boundary between the central non-reflective region 14a and the reflective region 14c in contact therewith is given by y = (c−a) x / 2L + (a + c) / 2. It will be.

Further, the boundary line between the upper reflective region 14b and the non-reflective region 14d in contact with the upper side, and the boundary line between the lower reflective region 14c and the non-reflective region 14e in contact with the lower side are parallel to the x axis. Is separated by a distance b.
Therefore, the function that defines the boundary line between the upper reflective region 14b and the non-reflective region 14d in contact with the upper side is given by y = b, and the lower reflective region 14c and the non-reflective region in contact with the lower side are provided. The function that defines the boundary with 14e is given by y = −b.

  In short, the boundary lines between the three belt-like non-reflective regions 14a, 14d, and 14e and the two belt-like reflective regions 14b and 14c are four as described above, and y = b and y = −b are mutually Parallel parallel lines, and with respect to these parallel lines, y = (ac) x / 2L + (a + c) / 2 and y = (c-a) x / 2L + (a + c) / 2 is an inclined straight line inclined in opposite directions. The quantities of L, a, b, and c are assumed to be known.

As shown in FIG. 1A, the line sensor is caused to scan across the non-reflective areas 14a, 14d, 14e and the reflective areas 14b, 14c with respect to such a specially shaped inclination correcting marker 14. As shown in the drawing, the shooting line that is the locus at the time of tilting is assumed to be inclined from the vertical direction instead of the vertical direction. The inclination angle φ is measured from the horizontal direction.
Therefore, the function that defines the inclined line is given by y = kx + d = xtanφ + d.
Here, by extracting the edge of the line sensor image, the length between the intersection points of the function that defines the inclined line and the function that defines the four boundary lines described above is obtained.

For example, the center of the length m that crossing the non-reflection areas 14a, and the length n across the upper reflective region 14b, a length l across the lower side of the reflection area 14c, the non-reflective region 14d of the uppermost the length p across the length p is required to cross the non-reflection areas 14e of the bottom side. That is, data with a length ratio p: l: m: n: p is obtained.

Here, as shown in FIG. 1C, from the length ratio p: l: m: n and the marker shape, the length ratio n: (l + m−n) / 2: (− l + m + n) / 2 is It is constant regardless of the inclination of the line sensor 2. (L + mn) / 2 and (-1 + m + n) / 2 were calculated from p: l: m: n. These ratios p: l: m: n: p vary with the lengths l _{O ′} and l _{O ″ and} can be expressed by [Equation 1].

That is, based on the length ratio p: l: m: n and the marker shapes L, a, b, and c, as shown in FIG. 1B, the length that is the x coordinate of the intersections O ′ and O ″. l _{O '} and l _{O "} can be recognized. FIG.1 (b) is an enlarged view of the part enclosed with the broken line in Fig.1 (a).
Further, the intersections O ′ and O ″ of the inclined straight lines y = (ac) x / 2L + (a + c) / 2 and y = (c−a) x / 2L + (a + c) / 2 are obtained. By substituting l x _' and l _{o "} which are x coordinates, y coordinates are obtained.
Accordingly, the xy coordinates of the intersections O ′ and O ″ are as shown in [Formula 2].

Further, tanφ of the inclined line is obtained by dividing the difference of the y-coordinates of the intersections O ′ and O ″ by the difference of the x-coordinates of the intersections O ′ and O ″.
Similarly, the y-coordinate (x-axis intercept) at which the inclined line intersects the x-axis is set to x = 0 in the straight line passing through the intersections O ′ and O ″, so that as shown in [Equation 3] Become.

The distance between the two points of intersections O ′ and O ″ is the length that the line sensor crosses the central non-reflective region 14a and is the actual value of m. The distance between the points of intersections O ′ and O ″ is For example, based on [Equation 2] and [Equation 3], the difference between the x coordinates of the intersections O ′ and O ″ is divided by cosφ, or the difference between the y coordinates of the intersections O ′ and O ″ is obtained as sinφ. Or by dividing the square of the difference between the x-coordinates of the intersections O ′ and O ″ and the square of the difference between the y-coordinates to obtain the square root. If this value is L _m [mm], the center If a value obtained by measuring the length of the line sensor crossing the non-reflective region 14a by the line sensor is x _p [pix], L _m / x _p [mm / pix] can be obtained as a calibration value.

Further, since the shooting position of the line sensor 2 is known, it can be estimated from the shooting position where the origin is.
In this way, the relative position and direction of the inclination correction marker 14 with respect to the imaging line are obtained, and therefore, the position and direction of the measurement target to which the inclination correction marker 14 is attached is obtained.

In the above description, the tilt correction marker 14 in which the two belt-like reflection regions 14b and 14c are inserted between the three belt-like non-reflection regions 14a, 14d, and 14e is taken as an example. Instead, what is shown in FIGS. 10 (a) to 10 (d) is conceivable.
That is, FIG. 10A shows a marker having the same shape as the shape shown in FIG. 1 used in the above description, that is, the shape of the central non-reflective region forms a trapezoid.

FIG. 10B shows a case where the width of the central non-reflective area is uniformly reduced from the upper and lower sides to 0 as it goes to the right in the figure, that is, the central non-reflective area becomes an isosceles triangle. It is a thing.
Further, in FIG. 10C, the shape of the central non-reflective region is trapezoidal, and the upper and lower reflective regions are right triangles.
FIG. 10 (d) shows a case where the boundary line between the central non-reflective region and the upper and lower reflective regions are balanced with each other, that is, the central non-reflective region is a parallelogram.

(1.1) Correcting the tilt of the lyche sensor based on specific numerical values The same marker shape as in FIG. 1 is used. Specific values are shown in FIG. 2 below. a = 1 cm, b = 4 cm, c = 3 cm, d = 4 cm, L = 10 cm. The unit of length in FIG. 2 is all [cm]. Note that the L, a, b, and c values of the markers used in the following description remain unchanged from those shown in FIG.

First, the photographing line of the line sensor in FIG. Since the inclination is a straight line passing through φ = 60 ° and (x, y) = (2, 0), it can be expressed as follows.
y = xtan 60 ° -2tan 60 ° = 2x-4
Next, a function of the boundary line (inclination straight line) that defines the boundary between the central non-reflective region and the reflective region in contact with the upper side and the lower side thereof is given by the following expression.
y = (ac) x / 2L + (a + c) / 2 =-(x / 10) +2
y = (c−a) x / 2L + (a + c) / 2 = (x / 10) −2

The intersection points O ′ and O ″ can be obtained by solving simultaneous equations for the function that defines the imaging line of the above-described line sensor and the two inclined straight lines. The coordinates of the intersection points are as follows.
(X _{O '} , y _O' ) = (2.9825 ..., 1.7017 ...)
(X _{O "} , y _O" ) = (0.8970 ..., -1.9102 ...)
In order to obtain l: m: n, the coordinates of the intersections d and e are required. The coordinates of the intersections d and e are obtained by solving the equation of the line sensor and simultaneous equations of y = −4 and y = 4, respectively.
(X _d , y _d ) = (0.3094 ...,-4)
(X _e , y _e ) = (4.3094 ..., 4)

The length of l: m: n is obtained from the above four coordinates.
l = 2.4129 ...
m = 4.1708 ...
n = 2.6537 ...
The lengths of l: m: n were respectively used as ratios and substituted into the following equations. The values shown in FIG. 2 were used for the values of L, a, b, and c.

According to the above [Equation 2], the calculation result is as follows.
(X _{O '} , y _O' ) = (2.9825 ..., 1.7017 ...)
(X _{O "} , y _O" ) = (0.8970 ..., -1.9102 ...)
At this time, the following results are obtained by applying specific numerical values to the variable [Equation 3] for the inclination φ and the x-axis intercept of the line sensor.
φ = 60 °
x = 2

(2) Displacement correction with respect to horizontal movement of pantograph As shown in FIG. 3, it is assumed that the pantograph is moved to +8 (to the right) in the horizontal direction. Before the pantograph moves in the horizontal direction, the photographing line of the line sensor is a straight line with an inclination of φ = 60 ° and (x, y) = (3, 0).
The ratio before horizontal movement is obtained in the same manner as the calculation formula (1.1).

The calculation results are as follows.
l = 2.5355 ...
m = 3.9391 ...
n = 2.7629 ...

Similar to the calculation method in (1.1) above, the inclinations φ and x of the line sensor are as follows.
φ = 60 °
x = 3
Thereafter, it is assumed that the pantograph moves +8 in the horizontal direction.
If the ratio after horizontal movement is l ′: m ′: n ′,
l '= 1.5551 ...
m '= 5.7928 ...
n '= 1.8896 ...
And

The inclinations φ ′ and x ′ of the line sensor after horizontal movement are as follows.
φ '= 60 °
x ′ = − 5
Since (horizontal movement amount) = (x before movement) − (x ′ after movement), the horizontal movement amount is as follows.
Horizontal movement amount = 3-(− 5) = 8

(3) Displacement correction for movement accompanied by rotation of pantograph As shown in FIG. 4, it is assumed that the pantograph moves by +12 (upward) in the vertical direction with rotation of an angle ψ (= 10 °) counterclockwise. To do. ψ should be substituted with the alphabet “psi” when subscripted due to restrictions on PC application software. The coordinate system x′-y ′ and the coordinate center O _psi after the movement are set to the coordinate system xy and the coordinate center O before the movement.

The imaging line of the line sensor is a straight line with an inclination of φ = 80 ° and a point I = (1, 0) in the coordinate system xy before movement. Accordingly, the imaging line of the line sensor is φ ′ = 80 ° −10 ° = 70 ° in the coordinate system x′-y ′ after movement, and intersects the x ′ axis at point I ′.
The lengths I to I ′ can be obtained by examining the number of pixels from the image of the line sensor and multiplying the number of pixels by the calibration value. In order to obtain the length of I to I ′ by desk calculation, since the coordinates of I are known, the coordinates of the remaining I ′ must be obtained. On the xy axis, it can be obtained if the equation of the straight line of the x ′ axis is known.

Here, since the center-to-center distance before and after the rotational movement is set to O to O _psi and the rotation angle is set to ψ = 10 °, the linear expression of the x ′ axis can be defined as follows.
y = xtan10 ° + 12
Further, the straight line expression of the photographing line of the line sensor is expressed as follows.
y = xtan80 ° -tan80 °
When the above equation is solved by simultaneous equations, the xy coordinates of I ′ can be obtained.
(X _{I '} , y _I' ) = (3.2159 ..., 12.5670 ...)
Therefore, the length of I to I ′ is as follows.
I ~ I '= 12.7609 ...

Further, the ratio before rotational movement is obtained in the same manner as the calculation formula (1.1). The calculation results are as follows.
l = 2.0977 ...
m = 3.8598 ...
n = 2.1658…
From this, the inclination φ and the x-axis intercept of the line sensor before movement are obtained.
φ = 80 °
x = 1

Assuming that the ratio after rotational movement is l ′: m ′: n ′, l ′: m ′: n ′ can also be obtained in the same manner as the calculation formula (1.1).
l '= 2.4085 ...
m '= 3.5664 ...
n '= 2.5384 ...
In the coordinate system xy before the movement, the inclination φ ′ and the x ′ axis intercept of the line sensor are as follows.
φ '= 70 °
x '= 3.2655 ...

The relative angle ψ is
ψ = 80 ° −70 ° = 10 °
From this, the difference y _{Opsi-I ′} of the y coordinate between the coordinate center O _psi and the point _{I ′} is y _{Opsi-I ′} = (O _{psi to} I ′) 3.2655sin10 °.
= 0.5670…
Moreover, the y _I 'y-coordinate y _I' of the point I _{'= (I~I')} sinφ
= 12.7609sin80 °
= 12.5670…

Therefore, the center-to-center distance O to O _psi is as follows.
_{O~O psi = y I '- y} Opsi-I' =
= 12.5670-0.5670 = 12

(4) Calculation of relative displacement of two markers As shown in FIG. 5, it is assumed that two markers are attached to a pantograph. That is, the other marker is displaced counterclockwise by an angle ψ (= 10 °) with respect to one marker, and is displaced +12 (upward) in the vertical direction and +2 (rightward) in the horizontal direction. Assuming that ψ should be substituted with the alphabet “psi” when subscripted due to restrictions on PC application software.

For one marker coordinate system xy and coordinate center O, the other marker coordinate system x′-y ′ and coordinate center O _psi are used.
The imaging line of the line sensor is a straight line with an inclination of φ = 80 ° and a point I = (1,0) in the coordinate system xy of one marker. Therefore, the imaging line of the line sensor is φ ′ = 80 ° −10 ° = 70 ° in the coordinate system x′-y ′ of the other marker, and intersects the x ′ axis at point I ′.

The xy coordinates of the center point between the relative displacements are as follows.
(X _O , y _O ) = (0, 0)
(X _Opsi , y _Opsi ) = (2,12)
That is, the center-to-center distance O to O _psi is as follows.
O to O _psi = (2 ² +12 ² ) ^1/2
= 12.1655…
Also, a vertical line is drawn downward from O _psi , the intersection point with the x axis is H, a straight line is drawn from O _{psi in the} horizontal direction, a vertical line is drawn downward from the intersection point between the straight line and the line sensor imaging line, and the perpendicular line and x Let H 'be the intersection with the axis.

The lengths I to I ′ can be obtained by examining the number of pixels from the image of the line sensor and multiplying the number of pixels by the calibration value. In order to obtain the length of I to I ′ by desk calculation, since the coordinates of I are known, the coordinates of the remaining I ′ must be obtained. On the xy axis, it can be obtained if the equation of the straight line of the x ′ axis is known. Here, the coordinates of the center point between the relative displacements are (x _Opsi , y _Opsi ) = (2, 12), and the rotation angle is set to ψ = 10 °. Can be defined as
y = xtan10 ° + (12-2tan10 °)

Further, the straight line expression of the photographing line of the line sensor is expressed as follows.
y = xtan80 ° -tan10 °
By solving the above equation with simultaneous equations, the coordinates of I ′ can be obtained.
Therefore, the length of I to I ′ is as follows.
I ~ I '= 12.3913 ...
The ratio of the shape of the lower marker is obtained in the same manner as the calculation formula (1.1).
The calculation results are as follows.
l = 2.0977 ...
m = 3.8598 ...
n = 2.1658…

From this, the inclination φ and the x-axis intercept of the line sensor with respect to one marker are obtained.
φ = 80 °
x = 1
When the ratio of marker shapes is l ′: m ′: n ′, l ′: m ′: n ′ can also be obtained in the same manner as the calculation formula (1.1).
l '= 2.1771 ...
m '= 4.0131 ...
n '= 2.3231 ...

The inclination φ ′ and the x ′ axis intercept of the line sensor with respect to the other marker are as follows.
φ '= 70 °
x '= 1.1695.
The relative angle ψ is
ψ = 80 ° −70 ° = 10 °
Than this,
y _{Opsi-I ′} = (O _psi ˜I ′) sinψ
= X'sinψ
= 1.1695 sinψ
= 0.2030 ...

From y _{I ′} = ( _{I to I ′} ) sinφ,
y _{I '} = 12.3913 sin10 °
= 12.2030…
Therefore, O to H are as follows.
_{O~H = y I '-y Opsi-} I' = 12.203007030 = 12
To find the center distance O-O _psi , consider the triangle OO _psi H.
Since the base is OH and the right triangle has a height of O _{psi to} H, the center-to-center distance O to O _psi is the hypotenuse of the triangle OO _psi H.
Since O to I = x = 1, the lengths I to H may be obtained.

FIG. 5 shows that I to H = I to H ′ + H to H. First, I to H ′ are obtained. I to H ′ can be obtained by the following equations.
I to H ′ = (O _{psi to} H) / tanφ
= 12 / tan80 °
= 2.1159

Next, H to H ′ are obtained. H to H ′ can be obtained by the following formula.
H to H ′ = (O _{psi to} I ′) cos ψ−y _{Opsi-I ′} / tan ψ
= 1.11695cos10 ° -0.2030 / tan10 °
= 1.1159
Therefore, the lengths I to H are as follows.
I to H = I to H '+ H to H'
= 2.1159−1.5559
= 1
From the above, the center-to-center distance O to O _psi is as follows.
O to O _psi = [(1 + 1) ² +12 ² ] ^1/2
= 12.1655…

Thus far, the center-to-center distance O to O _psi in the initial state has been calculated.
Assuming that the marker has moved from the initial state, the distance between the centers after movement can be obtained by the same method as described above (the calculation is the same, so the calculation of the distance between the centers after movement is omitted).
The relative displacement can be calculated by calculating the difference between the center distances after the movement from the center distance in the initial state.
Since the relative displacement is the same value as the amount of expansion and contraction of the spring, the contact force can be measured by multiplying this value by the spring constant.

Hereinafter, specific examples of the present invention will be described with reference to FIGS.
In this embodiment, in the field of measuring the contact force generated between the pantograph and the trolley line using image processing, the measurement error of the relative displacement between the hull and the sliding plate necessary for obtaining the contact force is corrected. The present invention relates to a method and apparatus.

FIG. 6 is an explanatory view showing a boat body and a sliding plate to which markers are attached, FIG. 7 is a perspective view showing the structure of a pantograph, FIG. 8 is a flowchart showing image processing, and FIG. 9 is a block diagram of an image processing apparatus.
As shown in FIGS. 6 and 7, the pantograph 3 includes a boat plate 3a that is held on a vehicle via a link structure frame 9 so as to be movable up and down, and a spring 8 for the boat plate 3a. The sliding plate 3b is supported and is in contact with a trolley wire (not shown).

  Markers 14 are attached to the boat board 3a and the sliding board 3b at two locations, respectively. As described above, the marker 14 is a striped pattern in which two band-like reflection areas are arranged between three band-like non-reflection areas, and the shape of the non-reflection area arranged in the center is trapezoidal. It has become.

In this embodiment, as shown in FIG. 6, from the initial state, the sliding plate 3b rotates / translates with respect to the boat plate 3a, and as a result, each marker 14 moves relatively, thereby the marker Even when the center-to-center distance is relatively displaced from L _k to L _kr , the center-to-center distance L _kr after movement can be obtained in the same manner.
Then, the relative displacement can be calculated by obtaining the difference between the center distances after movement from the center distance in the initial state. Since the relative displacement is the same value as the amount of expansion and contraction of the spring, the contact force can be measured by multiplying this value by the spring constant.

Each marker 14 attached to the pantograph 3 is photographed by the line sensor 2 (see FIG. 9), and the line sensor 2 scans across the reflective area and the non-reflective area of each marker 14.
The image acquired by the line sensor 2 is subjected to image processing by the image processing unit shown in FIG. The image processing unit includes an image input unit 70, memories 71 and 72, a marker pattern extraction unit 73, a pattern matching 74, a calibration value calculation unit 75 considering inclination, a correction unit 76 for horizontal / rotational movement, and a relative displacement output unit 77. Consists of

Image processing by the image processing unit shown in FIG. 9 is as shown in the flowchart of FIG. 8, and is processed in the following procedures (a) to (e).
(A) The image input unit 70 uses an image obtained by photographing the tilt correction marker 14 attached to the boat board 3a and the sliding board 3b with the line sensor 2 as an input image.
(A) The marker pattern extraction unit 73 captures a portion where the pantograph 3 does not move at all, and extracts the pattern of the inclination correction marker 14 from the portion.

(C) The pattern matching unit 75 detects a part of the input image that matches the pattern of the inclination correction marker 14 by pattern matching.
(D) The calibration value calculation unit 75 calculates a calibration value [mm / pix] taking into account the inclination of the pantograph 3 from the marker pattern, as described in the section of “(1) Line sensor inclination correction”.
(E) The correction unit 76 applies the inclination of the line sensor 2 to the position information of the pantograph 3 obtained by pattern matching, as described in the section “(3) Displacement correction for movement accompanied by rotation of the pantograph”. Is corrected, and an accurate movement amount and rotation amount of the pantograph 3 are calculated.

(F) The relative displacement output unit 77 outputs an accurate relative displacement based on the calibration value [mm / pix] of (d) and the movement amount and rotation amount of (e).
The relative displacement output unit 77 can further accurately determine the amount of expansion / contraction of the spring 8. That is, when the movement in the horizontal direction occurs, the spring extends diagonally, so that the amount of expansion and contraction corresponding to the diagonal extension can be calculated. Further, since the angle of the sliding plate 3b can be obtained, contact force measurement can be performed in consideration of the moment of inertia of the sliding plate 3b.

In the present embodiment, the measurement error of the displacement can be reduced by correcting the inclination of the imaging line of the line sensor 2 using the marker 14, and the inclination angle (imaging surface) of the line sensor 2 during the correction process. Therefore, it can be estimated from the inclination angle where the origin is.
Further, even when the pantograph 3 moves in the horizontal direction or with rotation while the imaging surface of the line sensor 2 is inclined, the displacement can be calculated with high accuracy.

  Furthermore, by attaching two markers 14 to the boat body 3a and the sliding plate 3b of the pantograph 3 and adjusting the inclination of the line sensor 2, the relative displacement can be calculated accurately, and the expansion and contraction of the spring 8 of the pantograph 3 can be calculated. The amount can be obtained with high accuracy.

  In the above-described embodiments, the “measurement error reduction method in pantograph contact force measurement” has been described. However, the present invention is not limited to this, and all measurements using a one-dimensional image by a line sensor such as a CCD are performed. Is applicable.

  The present invention can grasp rotational / translational motion even with one-dimensional measurement by using a special marker shape, and can be applied to image processing, for example, in a test for measuring shaking of a building or a structure against seismic motion. Is.

FIG. 1A is an explanatory diagram of a tilt correction marker used in the present invention, FIG. 1B is an explanatory diagram showing intersections l _{O ′} and l _{O ″} on the tilt correction marker, and FIG. ) Is an explanatory diagram regarding a ratio of lengths n: (l + mn) / 2: (− 1 + m + n) / 2 on the inclination correction marker. It is explanatory drawing of the marker for inclination correction by which the specific shape was specified. It is explanatory drawing regarding the displacement correction | amendment with respect to the horizontal direction movement of a pantograph. It is explanatory drawing regarding the displacement correction | amendment with respect to the movement accompanying rotation of a pantograph. It is explanatory drawing regarding calculation of the relative displacement of two markers. It is a conceptual diagram which shows the boat body and the sliding board to which the marker for inclination correction which concerns on one Example of this invention was attached. It is a perspective view which shows the pantograph with which the marker for inclination correction which concerns on one Example of this invention was attached. It is a flowchart which shows the image processing which concerns on one Example of this invention. It is a block diagram which shows the image process part which concerns on one Example of this invention. It is explanatory drawing which shows the modification of the marker for inclination correction. FIG. 11A is a schematic diagram of a pantograph measuring apparatus according to Japanese Patent Application No. 2006-285902, and FIG. 11B is an enlarged view of a one-dot chain line portion in FIG. It is explanatory drawing which shows the marker of the complicated shape. It is a flowchart which shows the pantograph measuring method based on Japanese Patent Application No. 2006-285902. It is a block diagram which shows the pantograph measuring apparatus based on Japanese Patent Application No. 2006-285902. It is a graph which shows the example of a space-time image. It is a graph which shows the example of search pattern generation. It is a graph which shows the pattern matching with respect to a space-time image. It is a graph which shows interpolation of a pantograph position. It is explanatory drawing of the locus | trajectory (imaging line) of the line sensor which scans a marker. 20A is an explanatory diagram showing the parallel movement of two markers, FIG. 20B is an explanatory diagram of displacement accompanying the rotation of the two markers, and FIG. 20C is an explanatory diagram showing the rotation of the markers. .

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Vehicle 2 Line sensor 3 Pantograph 3a Boat body 3b Slide board 4 Striped pattern marker 4a Non-reflective area 4b Reflective area 5 Illumination 6 Image processing part 7 Recording device 14 Tilt correction markers 14a, 14d, 14e Non-reflective areas 14b, 14c Reflection area 70 Image input unit 71, 72 Memory 73 Marker pattern extraction unit 74 Pattern matching unit 75 Calibration value calculation unit 76 in consideration of inclination Correction unit 77 for horizontal / rotational movement Relative displacement output unit A Black band B White band C Noise D Reserved part E Unconnected part

Claims (10)

A marker attached to the object to be measured;
A line sensor for photographing the marker;
In a measuring apparatus comprising: an image processing unit that performs image processing on an image acquired by the line sensor and calculates a position of the object to be measured;
The marker is a striped pattern formed by arranging a plurality of adjacent strip-shaped regions of different colors,
Parallel straight lines that are at least parallel to each other between the regions; and
A straight line serving as a boundary between the regions includes at least two inclined straight lines inclined with respect to the parallel straight line;
Wherein the image processing unit, the line sensor of the three least one of the line segments between a plurality of intersections of said straight line parallel and the inclined linear imaging line and a scanning trajectory to cross the region of the marker A measuring device using a line sensor, wherein a ratio of lengths is obtained, and at least two coordinates of an inclination of the photographing line with respect to the marker or the intersection are obtained based on the ratio of the three lengths . The image processing unit obtains an actual distance between the intersection points based on at least two coordinates of the photographing line inclination or the intersection point, and the number of pixels between the intersection points on the line sensor 2. The measuring apparatus using a line sensor according to claim 1, wherein the calibration value is obtained by dividing the distance. The image processing unit obtains a relative displacement of the object to be measured by comparing two coordinates of an inclination of the photographing line or the intersection point before and after the object to be measured is displaced. A measuring device using the line sensor described in 1. The image processing unit obtains displacement with relative rotation of the object to be measured by comparing two coordinates of an inclination of the photographing line or the intersection point before and after the object to be measured is displaced with rotation. A measuring device using the line sensor according to claim 1. The object to be measured includes a boat board that is held on a vehicle via a frame of a link structure so as to be movable up and down, and a sliding board that is supported by the boat board via a spring and contacts a trolley wire. 2. The measuring device using a line sensor according to claim 1, wherein the measuring device is a pantograph, and the markers are attached to the boat board and the sliding board, respectively. 6. The measuring apparatus using a line sensor according to claim 5, wherein the image processing unit obtains inclinations of the photographing lines with respect to the markers respectively attached to the boat board and the sliding board. 6. The measuring apparatus using a line sensor according to claim 5, wherein the image processing unit obtains coordinates of the intersection point with respect to the markers respectively attached to the boat board and the sliding board. The image processing unit obtains a relative position of the sliding board with respect to the boat board based on an inclination of the photographing line with respect to the markers attached to the boat board and the sliding board or coordinates of the intersection point, The measuring device using a line sensor according to claim 6 or 7, wherein an expansion / contraction amount of the spring sandwiched between the boat board and the sliding board is obtained from a relative position. Any one of the regions has a shape in which the width is uniformly expanded to both sides as it goes in one longitudinal direction, and the two regions that are in contact with both sides of the region have a width that goes inward as they go in the one longitudinal direction. 2. The measuring device using a line sensor according to claim 1, wherein two straight lines serving as a boundary between the region and the two regions are inclined in opposite directions by forming a shape that is uniformly reduced only. . 2. The line sensor according to claim 1, wherein any one of the regions has a parallelogram shape, and the inclined straight lines forming a boundary with the other region in contact with both sides are parallel to each other. Measuring device.

Images