JP6148760B2 - Object feature extraction system - Google Patents

Description

The present invention relates to a system for extracting features of an object, and more particularly to a system for extracting features of an object having a circular or elliptical outline using a beam sensor.
Here, the term beam sensor is broadly interpreted to mean a photoelectric sensor, a magnetic beam sensor, a particle beam sensor, or the like. The term beam is used for sensing beams such as light, magnetic beams, and particle beams.

An object moved by a belt conveyor or the like has many members having a surface having a circular or elliptical shape, and there is an object feature extraction system for extracting the feature. Expensive.
In addition, a system for discriminating between a motorcycle and a bicycle using a transmission / light-shielding pattern caused by wheel characteristics formed by the contour of the wheel using a photoelectric sensor has been put to practical use in discriminating between a motorcycle and a bicycle. .
Hereinafter, for the sake of brevity, motorcycles and bicycles are collectively referred to as two-wheeled vehicles. The identification of the type of motorcycle is required in a bicycle parking lot that accepts motorcycles and bicycles. Conventionally, there has been a wheel discrimination system that uses a photoelectric sensor for vehicle type discrimination.
However, a function that satisfies the manager of the bicycle parking area has not been achieved due to the problem that the bicycle can be completed easily by shielding the photoelectric sensor with a foot or the like.

Japanese Patent Application Laid-Open No. H10-228561 describes vehicle type discrimination to which a wheel discrimination system improved from the simple photoelectric sensor arrangement method is applied. This is a method paying attention to the side shape of the wheel of a two-wheeled vehicle. The light shielding state (showing the detected state, referred to as the on state in the following description) and the transmitting state (showing the undetected state) in the beam of the plurality of photoelectric sensors arranged on the side surface. The wheel is discriminated based on the pattern.
It is composed of photoelectric sensors installed on both sides of the passage, and five beams are radiated in a direction crossing the passage. The vehicle detection device includes a light projecting unit that projects five beams and a light receiving unit that receives each beam.
This configuration is an indirect method depending on the light shielding / transmission pattern of the beam generated by the characteristics of the wheel. Therefore, there is a problem that does not function sufficiently.

JP 2007-304725 A

In an indirect method as described in the prior art document, the feature of the wheel is judged indirectly by discriminating not the wheel feature itself but the transmission / shading pattern generated by the feature. there were. For this reason, the accuracy of discriminating the vehicle type has not realized the discriminating accuracy that sufficiently satisfies the bicycle parking manager.
The present invention adopts a photoelectric sensor arrangement pattern different from the prior art and the prior art, and directly extracts features such as the outer diameter and inner diameter of a wheel having a circular or elliptical outline, so that noise can be accurately detected. It is an object to provide an object feature extraction system that contributes to realizing strong wheel discrimination.
Also, an object feature extraction system that extracts features such as diameter, major axis, and minor axis of an object that is translated by a belt conveyor or the like and has an outline of a circle or an ellipse when observed from the side in the translation direction. Providing is also an issue.
Furthermore, it is also an object to provide a robust object feature extraction system that contributes to dealing with frauds such as reduced recognition and impersonation by verifying, correcting, and recorrecting the extracted object features. Become.

According to the first aspect of the present invention, a relative parallel movement is performed with respect to the beam sensor by a conveying means including a belt conveyor , an object having a circular outline is set as an object, and the emitted beam is the relative parallel movement. A system in which a plurality of beam sensors are arranged so as to proceed perpendicular to a moving path to form a group of beam sensors, and a feature of the object is extracted from a plurality of detection patterns of the beams, and a) the beam sensor At least three beam sensors in the group are arranged at predetermined vertical intervals in the direction perpendicular to the movement path plane, and b) the beam sensors arranged in the vertical direction are the object. The time difference between the detection times when the front side of the traveling direction is detected, the moving speed of the object, the vertical distance of the existing value, and the three moving paths on the contour relating to the detection of the three beams The horizontal interval is the product of the time difference between the detection times of each point and the moving speed, and the vertical distance in the vertical direction of the three moving paths and the vertical direction is an existing value. A position calculating means for calculating the relative positions of the three points on the front contour of the object, and c) by means of a predetermined geometric calculation means, the circular trajectory through which the three points pass as the contour of the object, An object feature extraction system comprising a feature extraction means for calculating a feature is provided. According to the second aspect of the present invention, relative parallel movement is performed with respect to the beam sensor by a conveying unit including a belt conveyor , an object having an elliptical contour is set as an object , and the emitted beam is the relative parallel movement. A system in which a plurality of beam sensors are arranged so as to proceed perpendicular to a moving path to form a group of beam sensors, and a feature of the object is extracted from a plurality of detection patterns of the beams, and a) the beam sensor At least four beam sensors in the group are arranged at predetermined vertical intervals in the direction perpendicular to the movement path plane, and b) the beam sensors arranged in the vertical direction are the object. The time difference between the detection times when the front side of the traveling direction is detected, the moving speed of the object, the vertical distance of the existing value, and the directions of the four moving paths on the contour relating to the detection of the four beams The horizontal interval is the product of the time difference between the detection times for each point and the moving speed, and the vertical distance in the vertical direction of the four points of the moving path is an existing value. C) position calculation means for calculating the position of four points on the front contour of the object, and c) by means of a predetermined geometric calculation means, the elliptic trajectory through which the four points pass as the contour of the object, And d) at least one beam sensor of the group of beam sensors is parallel to the moving path surface from the beam sensor at the uppermost position in the vertical direction and is longer than the vertical interval. E) a detection time when the uppermost beam sensor detects the front side of the object in the traveling direction, and a beacon disposed in the traveling direction in parallel with the beam sensor at the uppermost position. The moving speed calculating means for calculating the moving speed of the object from the time difference from the detection time when the sensor detects the front side of the moving direction of the object is provided, and the moving speed of the object used by the position calculating means is the moving speed There is provided an object feature extraction system characterized by the moving speed calculated by the speed calculating means. Here, the contour may be not only the outer periphery of the object but also the inner periphery when the shape of the object is a donut shape.

In the object feature extraction system according to the first aspect, further, a) at least one of the beam sensor groups is parallel to the moving path plane from the uppermost beam sensor in the vertical direction in the traveling direction; B) arranged at a parallel interval shorter than the vertical interval; b) a detection time when the uppermost beam sensor detects the front side of the object in the direction of travel; and a parallel time direction from the beam sensor at the uppermost position. A moving speed calculating means for calculating a moving speed of the object from a time difference from a detection time when the beam sensor detected the front side in the traveling direction of the object is provided, and the moving speed of the object used by the position calculating means is: The moving speed calculated by the moving speed calculating means may be used.

Here, the relative translation may be generated by the photoelectric sensor group that translates with respect to the fixed object.

The uppermost position is set such that the height from the movement path plane is equal to or less than half of the diameter of the circle or ellipse formed by the assumed contour in the vertical direction of the movement path plane. It is good.

From the photoelectric sensors arranged in the traveling direction in parallel, further arranged photoelectric sensors arranged in parallel to the movement path surface and in parallel with the traveling direction longer than the short parallel spacing,
From the time difference between the time when the uppermost photoelectric sensor detects the front side in the direction of travel of the object and the time when the photoelectric sensors arranged at a long parallel interval detect the front side in the direction of travel of the object An average moving speed calculating means for calculating an average moving speed of a long section of the object, and calculating a trend of acceleration / deceleration from a comparison between the average moving speed and the moving speed;
A speed correction means for correcting the moving speed for calculating the positions of the three points or the four points according to the acceleration / deceleration trend may be provided.

The at least three or four photoelectric sensors including the at least three or at least four photoelectric sensors and the photoelectric sensors arranged in the traveling direction parallel to the movement path surface and at a long parallel interval are the movement path surface. Are arranged in the vertical direction with the above-mentioned vertical interval,
Furthermore, one photoelectric sensor is arranged at the short parallel interval in the traveling direction parallel to the movement path surface from the photoelectric sensor arranged at the long parallel interval,
a) The time when the photoelectric sensor arranged at the long parallel interval detects the front side in the traveling direction of the object, and the photoelectric sensor arranged at the long parallel interval in parallel with the photoelectric sensor. Furthermore, it comprises second movement speed calculation means for calculating the second movement speed of the object from the time difference from the time when one photoelectric sensor detects the front side in the traveling direction of the object,
b) Calculated from the time difference between the times when at least three or four photoelectric sensors including the photoelectric sensors arranged at long parallel intervals detect the front side in the traveling direction of the object and the moving speed By comparing the positional relationship of 3 or 4 points on the front contour of the object with the positional relationship of the 3 or 4 points calculated by the position calculating means, a deviation exceeding a predetermined ratio was detected. In some cases, a verification unit that executes a notification operation may be provided.

The detection delay time of the photoelectric sensor that changes in accordance with the inclination of the tangent of the contour trajectory formed by the front side in the traveling direction is estimated from the circular or elliptical trajectory calculated by the feature extraction unit, and constitutes the photoelectric sensor group Time correction means for correcting the detection time of the photoelectric sensor may be provided.

There may be provided time correction repeating means for calculating the detection time by repeating the time correction using the circular or elliptical trajectory recalculated by the feature extraction means using the correction detection time corrected by the time correction means. Good.

The detection time may be calculated by repeating the time correction a plurality of times.

(Delete)

FIG. 1 is an installation image diagram of a wheel feature extraction system according to a first embodiment introduced into an entrance / exit passage of a bicycle parking lot. FIG. 2 is a diagram showing the positional relationship between the wheel and the photoelectric sensor group at time t1 at the moment when the beam of the photoelectric sensor 1001 changes to the ON state. FIG. 3 is a diagram showing the positional relationship between the wheel and the photoelectric sensor group at time t2 at the moment when the beam of the photoelectric sensor 1002 is turned on by the wheel. FIG. 4 is a diagram showing the positional relationship between the wheel and the photoelectric sensor group at time t3 at the moment when the beam of the photoelectric sensor 1003 transitions to the ON state by the wheel. FIG. 5 is a diagram showing the positional relationship between the wheel and the photoelectric sensor group at time t3 at the moment when the beam of the photoelectric sensor 1004 is turned on by the wheel. FIG. 6 is a timing chart showing the off state, the on state, and the state transition of the photoelectric sensors 1001 to 1004. FIG. 7 is a functional configuration diagram of the wheel feature extraction computer 7000. FIG. 8 is a contrast diagram showing the principle by which the positions of the three points are calculated. FIG. 9 is a flowchart from when a discrimination target moving toward the settlement machine is detected after moving through the bicycle parking lot until the wheel feature is output. FIG. 10 is an installation image diagram of the object feature extraction system according to the second embodiment that is introduced into a belt conveyor that carries a member having an elliptical outline. FIG. 11 is a diagram showing the positional relationship between the member and the magnetic beam sensor group at time t1 at the moment when the beam of the magnetic beam sensor 10001 is turned on by the member 10007. FIG. 12 is a diagram showing the positional relationship between the member and the magnetic beam sensor group at time t2 at the moment when the beam of the magnetic beam sensor 10002 is turned on by the member 10007. FIG. 13 is a diagram showing the positional relationship between the member contour and the magnetic beam sensor group at time t3 at the moment when the beam of the magnetic beam sensor 10003 is turned on by the member 10007. FIG. 14 is a diagram showing the positional relationship between the member contour and the magnetic beam sensor group at time t4 at the moment when the beam of the magnetic beam sensor 10004 is turned on by the member 10007. FIG. 15 is a diagram showing the positional relationship between the member contour and the magnetic beam sensor group at time t5 at the moment when the beam of the magnetic beam sensor 10005 is turned on by the member 10007. FIG. 16 is a timing chart showing the off state, the on state, and the state transition of the magnetic beam sensors 10001 to 10005. FIG. 17 is a functional configuration diagram of the object feature extraction computer 17000. FIG. 18 is a diagram illustrating the principle that the delay time occurs. FIG. 19 is a block diagram showing a data structure of a delay time table indicating a delay time for one magnetic beam sensor. FIG. 20 is a flowchart of the operation of the object feature extraction computer from detection of a member transported on a belt conveyor to parallel translation until output. FIG. 21 is an installation image diagram of the wheel discriminating system of the first embodiment introduced into the entrance / exit passage of the bicycle parking lot. FIG. 22 is a diagram showing the positional relationship between the wheels and the photoelectric sensor group at time t1 at the moment when the beam of the photoelectric sensor 1001 transitions to the ON state. FIG. 23 is a diagram showing a positional relationship between the wheels and the photoelectric sensor group at time t2 at the moment when the beam of the photoelectric sensor 1002 is turned on by the wheel 1008. FIG. 24 is a diagram showing a positional relationship between the wheels and the photoelectric sensor group at time t3 at the moment when the beam of the photoelectric sensor 1003 is turned on by the wheel 1008. FIG. 25 is a diagram showing a positional relationship between the wheels and the photoelectric sensor group at time t4 at the moment when the beam of the photoelectric sensor 1004 is turned on by the wheel 1008. FIG. 26 is a diagram illustrating the positional relationship between the wheels and the photoelectric sensor group at time t5 at the moment when the beam of the photoelectric sensor 21001 is turned on by the wheel 1008. FIG. 27 is a timing chart showing the off state, the on state, and the state transition of the photoelectric sensors 1001 to 1004 and 21001. FIG. 28 is a functional block diagram of the wheel feature extraction computer 28000. FIG. 29 is a contrast diagram showing the principle by which the positions of the three points are calculated. FIG. 30 is a flowchart from when a discrimination target moving toward the checkout machine is detected by moving through the bicycle parking lot to output to the wheel feature extraction computer. FIG. 31 is an installation image diagram of the wheel feature extraction system according to the fourth embodiment introduced into the entrance / exit passage of the bicycle parking lot. FIG. 32 is a diagram showing the positional relationship between the wheel and the photoelectric sensor group at time t1 when the wheel 1008 transitions to the ON state of the photoelectric sensor 1001 beam. FIG. 33 is a diagram showing the positional relationship between the wheels and the photoelectric sensor group at time t2 at the moment when the beam of the photoelectric sensor 1002 is turned on by the wheel 1008. FIG. 34 is a diagram showing the positional relationship between the wheels and the photoelectric sensor group at time t3 at the moment when the beam of the photoelectric sensor 1003 is changed to the ON state by the wheel 1008. FIG. 35 is a diagram showing a positional relationship between the wheels and the photoelectric sensor group at time t4 at the moment when the beam of the photoelectric sensor 1004 is turned on by the wheel 1008. FIG. 36 is a diagram showing the positional relationship between the wheels and the photoelectric sensor group at time t5 at the moment when the beam of the photoelectric sensor 31001 changes to the ON state. FIG. 37 is a diagram showing the positional relationship between the wheels and the photoelectric sensor group at time t6 at the moment when the beam of the photoelectric sensor 31002 is turned on by the wheel 1008. FIG. 38 is a diagram showing a positional relationship between the wheels and the photoelectric sensor group at time t7 at the moment when the beam of the photoelectric sensor 31003 is turned on by the wheel 1008. FIG. 39 is a diagram illustrating the positional relationship between the wheels and the photoelectric sensor group at time t8 at the moment when the beam of the photoelectric sensor 31004 is turned on by the wheel 1008. FIG. 40 is a timing chart showing the off state, the on state, and the state transition of the photoelectric sensors 1001 to 1004 and 31001 to 10004. FIG. 41 is a functional configuration diagram of the wheel feature extraction computer 41000. FIG. 42 shows the positions (relative positions) of three points on the trajectory to be discriminated from the time differences obtained from t5, t7 and t8, the moving speed V2, and the distances Y1 and Y2 in the vertical direction of the passage surface. This shows the principle of calculating.

Hereinafter, specific examples of the present invention will be described with reference to the drawings.
In the embodiment, each functional component that realizes the function of the present invention executes a control program such as firmware incorporated in advance by a processor such as a computer or a circuit, and cooperates with a device such as a photoelectric sensor or a magnetic beam sensor. It is realized by doing. These programs are recorded on a computer-readable recording medium, read from the recording medium by the processor, and executed by a system user operating or receiving a signal from a device constituting the system. Is done.

FIG. 1 is an installation image diagram of a wheel feature extraction system according to a first embodiment introduced into an entrance / exit passage of a bicycle parking lot. The upper stage is a plan view and the lower stage is a side view.
Here, an object feature extraction system used for discriminating the type of a bicycle parking lot is referred to as a wheel feature extraction system.
A plurality of photoelectric sensors 1001 to 1004 are arranged on both sides of a passage 1005 serving as a moving path of the object. In each photoelectric sensor, a light-emitting device that emits a beam from one side and a light-receiving device that receives the beam on the other side are paired so as to face each other.
The photoelectric sensors 1001, 1003 and 1004 are arranged in a line in the direction perpendicular to the passage surface. The interval between the photoelectric sensor 1001 and the photoelectric sensor 1003 is Y1, and the interval between the photoelectric sensor 1003 and the photoelectric sensor 1004 is Y2.
The photoelectric sensor 1002 is installed at a position parallel to the passage surface from the photoelectric sensor 1001 by a short parallel distance XS in the traveling direction of the two-wheeled vehicle 1007. Here, XS is a photoelectric sensor 1002 at a timing earlier than the photoelectric sensor 1003 arranged at the interval Y1 from the photoelectric sensor 1001 after the beam of the photoelectric sensor 1001 detects the wheel when the assumed motorcycle passes. Are set to a short enough length to detect the wheel.
Since there is a settlement machine 1006 at the side of the passage at a predetermined distance from the photoelectric sensor groups 1001 to 1004 in the traveling direction of the two-wheeled vehicle 1007, the two-wheeled vehicle using the bicycle parking lot always passes here.
Reference numeral 1008 denotes a wheel of the motorcycle 1007. The photoelectric sensor 1001 arranged at the highest position is provided to have a height equal to or less than the radius of the smallest wheel among the assumed wheels 1008.
Here, although a bicycle is described as a two-wheeled vehicle, a motorcycle or a three-wheeled vehicle can also be included in the discrimination target vehicle. In the present invention, a two-wheeled vehicle is a vehicle to be stored in a bicycle parking lot. The number of wheels of the discrimination target vehicle is not interpreted as being limited to two.
In addition, a photoelectric sensor 1002 is installed to calculate a moving speed V1, which will be described later. If the moving speed is known in advance, a sensor group is configured by only three photoelectric sensors 1001, 1003, and 1004. It may be a thing. In this case, a wheel feature extraction computer that does not include the moving speed calculation unit 7002 may be used.

  The photoelectric sensors 1001 to 1004 are in an on state (a state in which an object shields light beyond a predetermined transition threshold area of the beam section) and an off state (even if the beam section is not shielded or shielded). Is transmitted to a wheel feature extraction computer (not shown).

FIG. 2 is a diagram illustrating the positional relationship between the wheels and the photoelectric sensor group at time t1 at the moment when the beam of the photoelectric sensor 1001 changes to the ON state. Reference numeral 2001 denotes a contact point between the beam of the photoelectric sensor 1001 and a wheel.
Here, although expressed as a contact, strictly speaking, it is the center point of the beam cross section at the moment when the member covers the beam cross section having an area exceeding the transition threshold area of light transmission / shielding such as 50% of the beam cross section.

  FIG. 3 is a diagram showing the positional relationship between the wheels and the photoelectric sensor group at the time t2 at the moment when the beam of the photoelectric sensor 1002 is turned on by the wheel 1008. Reference numeral 3001 denotes a contact point between the beam of the photoelectric sensor 1002 and the wheel.

  FIG. 4 is a diagram showing a positional relationship between the wheels and the photoelectric sensor group at time t3 at the moment when the beam of the photoelectric sensor 1003 is turned on by the wheel 1008. Reference numeral 4001 denotes a contact point between the beam of the photoelectric sensor 1003 and the wheel.

  FIG. 5 is a diagram showing the positional relationship between the wheels and the photoelectric sensor group at time t4 at the moment when the beam of the photoelectric sensor 1004 is turned on by the wheel 1008. Reference numeral 5001 denotes a contact point between the beam of the photoelectric sensor 1004 and the wheel.

(Timing chart)
FIG. 6 is a timing chart showing the off state, the on state, and the state transition of the photoelectric sensors 1001 to 1004. The horizontal axis 6005 indicates the progress of time.
Reference numeral 6001 is a chart showing the off state of the photoelectric sensor 1001 at a high level and the on state at a low level. Reference numeral 6006 indicated by a dashed arrow indicates a time t1 at which the photoelectric sensor 1001 transitions from the off state to the on state.
Reference numeral 6002 is a chart showing the off state of the photoelectric sensor 1002 at a high level and the on state at a low level. Reference numeral 6007 indicated by a dashed arrow indicates a time t2 at which the photoelectric sensor 1002 transitions from the off state to the on state.
Reference numeral 6003 is a chart showing the off state of the photoelectric sensor 1003 at a high level and the on state at a low level. Reference numeral 6008 indicated by a dashed arrow indicates a time t3 when the photoelectric sensor 1003 transitions from the off state to the on state.
Reference numeral 6004 is a chart showing the off state of the photoelectric sensor 1004 at a high level and the on state at a low level. Reference numeral 6009 indicated by a dashed arrow indicates a time t4 when the photoelectric sensor 1004 transitions from the off state to the on state.

(Functional configuration of wheel feature extraction computer)
FIG. 7 is a functional configuration diagram of the wheel feature extraction computer 7000.
The beam state receiving unit 7001 receives an on state signal and an off state signal transmitted from the photoelectric sensors 1001 to 1004. The time zone in which the on-state signal is received is indicated at a low level on the timing chart. On the other hand, the time zone in which the off-state signal is received is indicated by a high level.
The moving speed calculation unit 7002 calculates the moving speed V1 to be discriminated according to a predetermined speed calculation formula from the time difference between the time t1 and the time t2 and the existing value XS.
The position calculation unit 7003 calculates the front contour trajectory to be determined from the time difference obtained from time t1, time t3, and time t4, the moving speed V1, and the distances Y1 and Y2 in the vertical direction of the passage surface. The position (relative position) of three points on the orbit symmetrical about the path vertical axis is calculated. Further, it moves symmetrically with respect to the vertical axis of the passage, and calculates the positions of three points on the front contour trajectory to be discriminated.
Here, the position calculation program and the intervals Y1 and Y2 are stored in the storage unit 7004.

FIG. 8 is a contrast diagram showing the principle by which the positions of the three points are calculated.
The upper diagram (1) is the timing chart itself of FIG. 8005 is the timing at which the photoelectric sensor 1001 first detects the discrimination target. Similarly, 8006 and 8007 are timings when the photoelectric sensors 1003 and 1004 first detect the discrimination target.
The lower figure (2) is a diagram in which the timings indicated by 8001, 8002 and 8003 are converted into coordinate positions 8004, 8005 and 8006 taken on the X axis indicating the passage parallel direction and the Y axis indicating the passage vertical direction. is there.
Y1 shown in the diagram (2) indicates a distance in the Y-axis direction between the photoelectric sensor 1001 and the photoelectric sensor 1003 (hereinafter referred to as a height distance), and is an existing value. Similarly, Y2 indicates the height interval between the photoelectric sensors 1003 and 1004 and is an existing value.
X1 is calculated by multiplying the time difference between the time t1 and the time t3 by the moving speed V1, and a point on the contour trajectory on the front side of the height of the photoelectric sensor 1001 and a contour on the front side of the height of the photoelectric sensor 1003. A distance in the X-axis direction from a point on the orbit is shown. Similarly, x2 is calculated by multiplying the time difference between the time t3 and the time t4 by the moving speed V1, the point on the contour trajectory on the front side of the height of the photoelectric sensor 1003, and the contour on the front side of the height of the photoelectric sensor 1004. A distance in the X-axis direction from a point on the orbit is shown.
Therefore, 8004, 8005, and 8006 indicate the relative positions of the three points on the contour contour on the front side of the discrimination target indicated by the graph 8007 shown on the XY coordinates (however, they are symmetric with respect to the Y axis). For example, if a position (X0, Y0) indicated by 8004 is assumed, 8005 is expressed as (X0 + (t3-t1) V1, Y0-Y1). Similarly, 8006 is represented as (X0 + (t4-t3) V1, Y0-Y1-Y2).

Returning to the description of FIG.
The wheel feature extraction unit 7005 performs a known geometric operation on the relative positions (8004, 8005, and 8006) of the three points on the front contour trajectory to be discriminated calculated by the position calculation unit 7003 to obtain a circular trajectory. The diameter is calculated, and the wheel features are extracted as the size of the track. A geometric calculation program including circular or elliptical orbital equations for executing the geometric calculation is stored in the storage device 7006.
Here, the diameter is extracted with the wheel feature as the size of the wheel, but the present invention is not limited to this. As wheel characteristics adopted in the present invention, a center position, a radius, an outer peripheral length, a wheel thickness (wheel height), and the like can be appropriately employed.

  The output unit 7007 outputs a feature indicating the size of the wheel to a vehicle discrimination computer (not shown).

FIG. 9 is a flowchart from when a discrimination target moving toward the settlement machine is detected after moving through the bicycle parking lot until the wheel feature is output.
In the beam state reception step S9001, the beam state reception unit receives the on state signal and the off state signal of the photoelectric sensor groups 1001 to 1004.
In the movement speed calculation step S9002, the movement speed calculation unit calculates the movement speed to be determined.
In position calculation step S9003, the position calculation unit calculates the positions of three points on the front contour trajectory to be determined by the calculation method described above.
In wheel feature calculation step S9004, the wheel feature calculation unit extracts a diameter as a feature indicating the size of the wheel.
In output step S9005, the output unit outputs a feature indicating the size of the wheel to another system such as a vehicle type discrimination system incorporating the wheel feature extraction system of the present embodiment.

Here, it is assumed that the wheels of the two-wheeled vehicle enter the passage straight without tilting, and the contour can be regarded as a circle. When it is assumed that the contour is an ellipse due to inclination or the like, the number of photoelectric sensors arranged in a line in the vertical direction of the passage surface is set to four or more, and the position on the front contour trajectory calculated by the position calculation unit 7003 is set to 4 As a point, the wheel feature extraction unit 7005 may perform a known geometric calculation to calculate a short diameter or a long diameter of an elliptical orbit as appropriate.

FIG. 10 is an installation image diagram of the object feature extraction system according to the second embodiment, which is introduced into a belt conveyor that carries a member having an elliptical outline. The upper stage is a plan view and the lower stage is a side view. The member is provided so that the diameter (minor axis or major axis) indicated by the contour thereof is fixed and translated so as to be perpendicular to the belt conveyor surface.
Here, the magnetic beam sensor group detects the member when it is translated by the belt conveyor, but the translation of the present invention is a relative translation, and the direction of the diameter of the contour of the fixed member. A method of generating relative translation of the member with respect to the magnetic beam sensor group by a method in which the magnetic beam sensor group moves in a direction perpendicular to the magnetic beam sensor group may be used.
In the following description, a state where a magnetic beam is approaching and detecting an object is an on state, and a state where the object is not detected far is an off state. Describe.
A plurality of magnetic beam sensors indicated by 10001 to 10005 are arranged beside a belt bear 10006 that conveys a member 10007 having an elliptical outline. The beams emitted from the magnetic beam sensors are all installed so as to intersect perpendicularly with respect to the translation of the member.
Magnetic beam sensors indicated by 10001, 10003, 10004, and 10005 are arranged in a line in the direction perpendicular to the belt conveyor surface. The interval between the magnetic beam sensor 10001 and the magnetic beam sensor 10003 is Y1, and the interval between the magnetic beam sensor 10003 and the magnetic beam sensor 10004 is Y2. Further, the interval between the magnetic beam sensor 10004 and the magnetic beam sensor 10005 is Y3.
The magnetic beam sensor 10002 is installed at a position parallel to the belt conveyor surface and separated from the magnetic beam sensor 10001 by a short parallel interval XS in the traveling direction of the belt conveyor 10006. Here, XS is a timing earlier than that of the magnetic beam sensor 10001 arranged at the interval Y1 from the magnetic beam sensor 10001 after the beam of the magnetic beam sensor 10001 detects the contour when the assumed member 10007 passes. Thus, the beam of the magnetic beam sensor 10002 is set to a short enough length to detect the contour. This is advantageous for calculating the average moving speed at an early stage, but the short length of the present invention is not limited to this, and may be shorter than XL described later.
Here, the magnetic beam sensor 10001 arranged at the highest position is provided so as to have a height equal to or less than the minor axis of the smallest member among the assumed members 10007.

  The magnetic beam sensors 10001 to 10005 transmit the on state and off state of the beam crossing the belt conveyor 10006 to an object feature extraction computer (not shown).

FIG. 11 is a diagram showing the positional relationship between the member and the magnetic beam sensor group at time t1 at the moment when the beam of the magnetic beam sensor 10001 is turned on by the member 10007. 11001 is a contact point between the beam of the magnetic beam sensor 10001 and the member contour.
Although expressed here as a contact point, logically 50% of the beam cross section is the transition threshold area and is the moment when the member passes the center point of the beam cross section, but the actual sensitivity of the sensor is not limited to this. .

  FIG. 12 is a diagram showing the positional relationship between the member and the magnetic beam sensor group at time t2 at the moment when the beam of the magnetic beam sensor 10002 is turned on by the member 10007. Reference numeral 12001 denotes a contact between the beam of the magnetic beam sensor 10002 and the member contour.

  FIG. 13 is a diagram showing the positional relationship between the member contour and the magnetic beam sensor group at time t3 at the moment when the beam of the magnetic beam sensor 10003 is turned on by the member 10007. Reference numeral 13001 denotes a contact between the beam of the magnetic beam sensor 10003 and a member contour.

  FIG. 14 is a diagram showing the positional relationship between the member contour and the magnetic beam sensor group at time t4 at the moment when the beam of the magnetic beam sensor 10004 is turned on by the member 10007. Reference numeral 14001 denotes a contact between the beam of the magnetic beam sensor 10004 and the member contour.

  FIG. 15 is a diagram showing the positional relationship between the member contour and the magnetic beam sensor group at time t5 at the moment when the beam of the magnetic beam sensor 10005 is turned on by the member 10007. Reference numeral 15001 denotes a contact point between the beam of the magnetic beam sensor 10005 and the member contour.

(Timing chart)
FIG. 16 is a timing chart showing the off state, the on state, and the state transition of the magnetic beam sensors 10001 to 10005. The horizontal axis 16000 shows the progress of time.
16001 is a chart showing the off state of the magnetic beam sensor 10001 at a high level and the on state at a low level. 16006 indicated by a broken line arrow indicates a time t1 at which the magnetic beam sensor 10001 transitions from the off state to the on state. 16014 indicated by a broken-line arrow indicates a time t9 at which the on state is changed to the off state.
16002 is a chart showing the off state of the magnetic beam sensor 10002 at a high level and the on state at a low level. 16007 indicated by a broken line arrow indicates a time t2 when the magnetic beam sensor 10002 transitions from the off state to the on state. Reference numeral 16015 indicated by a broken line arrow indicates a time t10 at which the transition from the on state to the off state occurs.
16003 is a chart showing the off state of the magnetic beam sensor 1003 at a high level and the on state at a low level. 16008 indicated by a broken line arrow indicates a time t3 when the magnetic beam sensor 10003 transitions from the off state to the on state. A reference numeral 16013 indicated by a dashed arrow indicates a time t8 when the on state is changed to the off state.
Reference numeral 16004 is a chart showing the off state of the magnetic beam sensor 10004 at a high level and the on state at a low level. 16009 indicated by a broken-line arrow indicates a time t4 when the magnetic beam sensor 10004 transitions from the off state to the on state. A reference numeral 16012 indicated by a broken line arrow indicates a time t7 at which the on state is changed to the off state.
Reference numeral 16005 is a chart showing the off state of the magnetic beam sensor 10005 at a high level and the on state at a low level. 16010 indicated by a dashed arrow indicates a time t5 when the magnetic beam sensor 10005 transitions from the off state to the on state. Reference numeral 16011 indicated by a broken line arrow indicates a time t6 when the on state is changed to the off state.
The time when each chart transitions from the off state to the on state indicates the time when the front contour of the member 10007 is detected by the magnetic beam sensor group.
In addition, the time at which the transition from the on state to the off state indicates the time at which the contour rear side of the member is detected by the magnetic beam sensor group.

(Functional structure of object feature extraction computer)
FIG. 17 is a functional configuration diagram of the object feature extraction computer 17000.
The beam state receiving unit 17001 receives the on state signal and the off state signal transmitted from the magnetic beam sensors 10001 to 10005. The time zone in which the on-state signal is received is indicated at a low level on the timing chart. On the other hand, the time zone in which the off-state signal is received is indicated by a high level.
The moving speed calculation unit 17002 calculates the moving speed V1 of the member according to a predetermined speed calculation formula from the time difference between the time t1 and the time t2 and the existing value XS.
The position calculator 17003 calculates the time difference obtained from the time t1, the time t3, the time t4, and the time t5, the moving speed V1, and the distances Y1 and Y2 in the belt conveyor surface vertical direction, which are the existing values, on the front side of the member. The position (relative position) of four points on the symmetrical track with respect to the contour track and the belt conveyor vertical axis is calculated. Further, the position of four points on the contour track on the front side of the member is calculated by moving symmetrically with respect to the belt conveyor vertical axis.

The principle for calculating these four points is the same as that described in the first embodiment with reference to FIG. 8, and can be extracted from the timing chart shown in FIG.
Here, the position calculation program and the intervals Y1 and Y2 are stored in the storage unit 17004.

Returning to the description of FIG.
The object feature extraction unit 17005 performs a known geometric operation on the relative positions of the four points on the contour trajectory on the front side of the member calculated by the position calculation unit 17003, calculates the major axis and minor axis of the elliptical trajectory, Extract as object features. Here, a program including an elliptic equation as a geometric calculation for executing the geometric calculation is stored in the storage device 17006.
Here, the major axis and the minor axis are extracted using the object feature as the size of the contour trajectory, but the present invention is not limited to this. As the object features that can be employed in the present invention, a center position, an outer peripheral length, and the like can be appropriately employed.

The detection time correction unit 17007 calculates a delay time from the installation height and the moving speed V1 for each magnetic beam sensor, and executes time correction of the detection time.
Here, the delay time is calculated by referring to a delay time table 17008 in which a delay time caused by the height of the magnetic beam sensor from the belt conveyor surface, the size of the track, and the moving speed is previously created for each magnetic beam sensor. Done.

FIG. 18 is a diagram illustrating the principle that the delay time occurs.
Reference numeral 18001 denotes a beam cross section of the magnetic beam sensor at a high position. When the front side of the contour advances to position 18002 at time t0, more than half of the area of the beam cross section is shielded and detection is performed (here, the member moves from left to right). In the case of this height, since the angle between the inclination of the contour and the moving direction of the member is close to 90 degrees, the beam sensitivity is good. If the front side of the contour advances to the position 18003, detection is complete, and there is little occurrence of a delay time that causes the time t1 to pass through the center 18004. Therefore, the delay time may be set to be short or short.
on the other hand. 18005 is a beam cross section of the magnetic beam sensor in a low position. In the case of this height, the beam sensitivity is poor because the angle between the inclination of the contour and the moving direction of the member is shallow. At time t0, when the front side of the contour advances to a position indicated by 18007, more than half of the area of the beam cross section should be shielded and detection should occur. Here, the contour front surface 18008 is in the vicinity of the center 18006 even at a time t1 slightly advanced from 18007 passing through the center 18006 of the beam cross section. Accordingly, the occurrence of a delay time in which the time when the front side of the contour passes the center of the beam cross section is t1 increases.
However, here, the calculation is performed with reference to a preset delay time table for each of the moving speed range and the diameter range divided into a predetermined number.

FIG. 19 is a block diagram showing a data structure of a delay time table indicating a delay time for one magnetic beam sensor. In this embodiment, five tables are prepared. When ignoring the delay times for the highest positions 10001 and 10002, three tables are prepared.
The moving speed area 19001 is divided into three sections: 0 <f ≦ f1, f1 <f ≦ f2, and f2 <f ≦ fmax, where f is the moving speed. Here, fmax is an assumed maximum moving speed.
The size range of the diameter in the vertical direction is divided into three sections: 0 <r ≦ r1, r1 <r ≦ r2, and r2 <r ≦ rmax, where r is the diameter. Here, rmax is the assumed maximum member diameter.
As an example, a delay time t31 indicated by 19003 indicates a delay time in the case where the moving speed is in the range of 0 <f ≦ f1 and the member diameter corresponds to 0 <r ≦ r1.
This table stores the results of experiments on delay times, etc., and the results of rounding of numerical values such as average of delay times and mode values created by a predetermined delay time calculation method.

Returning to the description of FIG.
The time correction repeater 17209 uses the detection time generated by the time correction, and uses the size of the contour trajectory extracted by repeating the operations of the moving speed calculator, the position calculator, and the object feature extractor. The operation of the time correction unit is performed again.
Such a repetition is repeated twice to calculate a highly accurate detection time.
The repetition of the present invention is not limited to the number of times of the present embodiment, and a plurality of times such as three times can be adopted. However, it is necessary to set the number of times so as not to diverge.

The output unit 17010 outputs a feature indicating the size of the object to an object feature extraction computer (not shown).
The feature output here is a feature extracted using the highly accurate detection time.

FIG. 20 is a flowchart of the operation of the object feature extraction computer from detection to output of a member that is transported by the belt conveyor and moves in parallel.
In the beam state reception step S20001, the beam state reception unit receives the on state signal and the off state signal of the magnetic beam sensor groups 10001-10005.
In movement speed calculation step S20002, the movement speed calculation unit calculates the movement speed of the member.
In the position calculation step S20003, the position calculation unit calculates the positions of three points on the contour trajectory on the front side of the member by the calculation method described above.
In the object feature calculation step S20004, the object feature calculation unit extracts a diameter (including a short diameter and a long diameter in the case of an ellipse) as a characteristic indicating the size of the object.
In time correction step S20005, the detection time correction unit corrects the detection time of each height sensor.
In time correction repeat step S20006, the time correction repeater calculates the detection time by repeating the operations of the moving speed calculator, the position calculator, and the object feature extractor twice using the detection time generated by the time correction. To do.
Using the size of the extracted contour trajectory, the operation of the time correction unit In the final feature extraction step S20007, the object feature extraction unit uses the highly accurate detection time by the time correction repetition unit as the object feature. Calculate the major and minor axis of the elliptical orbit.
In the output step S20008, the output unit outputs a feature indicating the size of the object to another system such as a member discrimination system incorporating the object feature extraction system of the present embodiment.
Here, the photoelectric sensor 10002 is installed to calculate the moving speed V1, but when the moving speed is known in advance, the sensor group is composed of only four photoelectric sensors 10001, 10003, 10004, and 10005. It may be. In this case, an object feature extraction computer that does not include the moving speed calculation unit 20002 may be used.

The third embodiment is a modified example of the first embodiment. This is a wheel feature extraction system in which the moving speed of the discrimination target is calculated more precisely and the wheel feature extraction accuracy is higher than that in the first embodiment. Portions that are common to the first embodiment will be described using the same reference numerals as much as possible, omitting redundant descriptions, and focusing on portions that are unique to the third embodiment.

FIG. 21 is an installation image diagram of the vehicle type identification system according to the first embodiment that is introduced into the entrance / exit passage of the bicycle parking lot. The upper stage is a plan view and the lower stage is a side view.
Similar to the first embodiment, a plurality of photoelectric sensors 1001 to 1004 are arranged on both sides of a passage 1005 serving as a moving path of an object. The difference from the first embodiment is that the photoelectric sensor 21001 is installed at a position separated by XL in the traveling direction of the two-wheeled vehicle 1007 parallel to the photoelectric sensor 1001 arranged at the highest position and the passage surface.
The photoelectric sensors 1001, 1003 and 1004 are arranged in a line in the direction perpendicular to the passage surface. The interval between the photoelectric sensor 1001 and the photoelectric sensor 1003 is Y1, and the interval between the photoelectric sensor 1003 and the photoelectric sensor 1004 is Y2.
The photoelectric sensor 1002 is installed at a position parallel to the passage surface from the photoelectric sensor 1001 and separated by a short parallel distance XS in the traveling direction of the two-wheeled vehicle 1007, and XS is set shorter than XL.
Reference numeral 1008 denotes a wheel of the motorcycle 1007. The photoelectric sensor 1001 arranged at the highest position is provided to have a height equal to or less than the radius of the smallest wheel among the assumed wheels 1008.

  The photoelectric sensors 1001 to 1004 and 21001 transmit the on state and off state of the beam crossing the passage to a wheel feature extraction computer (not shown).

  FIG. 22 is a diagram showing the positional relationship between the wheels and the photoelectric sensor group at time t1 at the moment when the beam of the photoelectric sensor 1001 transitions to the ON state. Reference numeral 2001 denotes a contact point between the beam of the photoelectric sensor 1001 and a wheel.

  FIG. 23 is a diagram showing a positional relationship between the wheels and the photoelectric sensor group at time t2 at the moment when the beam of the photoelectric sensor 1002 is turned on by the wheel 1008. Reference numeral 3001 denotes a contact point between the beam of the photoelectric sensor 1002 and the wheel.

  FIG. 24 is a diagram showing a positional relationship between the wheels and the photoelectric sensor group at time t3 at the moment when the beam of the photoelectric sensor 1003 is turned on by the wheel 1008. Reference numeral 4001 denotes a contact point between the beam of the photoelectric sensor 1003 and the wheel.

  FIG. 25 is a diagram showing a positional relationship between the wheels and the photoelectric sensor group at time t4 at the moment when the beam of the photoelectric sensor 1004 is turned on by the wheel 1008. Reference numeral 5001 denotes a contact point between the beam of the photoelectric sensor 1004 and the wheel.

  FIG. 26 is a diagram illustrating the positional relationship between the wheels and the photoelectric sensor group at time t5 at the moment when the beam of the photoelectric sensor 21001 is turned on by the wheel 1008. Reference numeral 26001 denotes a contact point between the beam of the photoelectric sensor 21001 and the wheel.

(Timing chart)
FIG. 27 is a timing chart showing the off state, the on state, and the state transition of the photoelectric sensors 1001 to 1004 and 21001. The horizontal axis 6005 indicates the progress of time.
Reference numeral 6001 is a chart showing the off state of the photoelectric sensor 1001 at a high level and the on state at a low level. Reference numeral 6006 indicated by a dashed arrow indicates a time t1 at which the photoelectric sensor 1001 transitions from the off state to the on state.
Reference numeral 6002 is a chart showing the off state of the photoelectric sensor 1002 at a high level and the on state at a low level. Reference numeral 6007 indicated by a dashed arrow indicates a time t2 at which the photoelectric sensor 1002 transitions from the off state to the on state.
Reference numeral 6003 is a chart showing the off state of the photoelectric sensor 1003 at a high level and the on state at a low level. Reference numeral 6008 indicated by a dashed arrow indicates a time t3 when the photoelectric sensor 1003 transitions from the off state to the on state.
Reference numeral 6004 is a chart showing the off state of the photoelectric sensor 1004 at a high level and the on state at a low level. Reference numeral 6009 indicated by a dashed arrow indicates a time t4 when the photoelectric sensor 1004 transitions from the off state to the on state.
27001 is a chart showing the off state of the photoelectric sensor 21001 at a high level and the on state at a low level. Reference numeral 27002 indicated by a broken line arrow indicates a time t5 when the photoelectric sensor 21001 transitions from the off state to the on state.

(Functional configuration of wheel feature extraction computer)
FIG. 28 is a functional block diagram of the wheel feature extraction computer 28000.
The beam state receiving unit 28001 receives an on state signal and an off state signal transmitted from the photoelectric sensors 1001 to 1004 and 21001. The time zone in which the on-state signal is received is indicated at a low level on the timing chart. On the other hand, the time zone in which the off-state signal is received is indicated by a high level.
The first moving speed calculation unit 28002 calculates the short section moving speed V1 to be determined from the time difference between the time t1 and the time t2 and the existing value XS according to a predetermined speed calculation formula.
The second moving speed calculation unit 28003 calculates the long-term average moving speed VA to be discriminated from the time difference between the time t1 and the time t5 and the existing value XL according to a predetermined speed calculation formula.
The speed correction unit 28004 calculates the acceleration trend α indicating the acceleration or deceleration trend from the difference between VA and V1, using the formula (VA−V1) / (t5−t2), and the photoelectric sensors 1001 to 1004 are calculated. The passing speed is corrected to generate a corrected moving speed for each photoelectric sensor.
The correction formula adopted here is as follows.
The corrected moving speed from the passage of the photoelectric sensor 1001 to the detection of 1003 is
V1 + 1 / 2α (t3−t1),
The corrected movement speed from the passage through the photoelectric sensor 1003 to the detection of 1004 is
V1 + 1 / 2α (t4-t3).
The correction movement speed calculation method shown here is an expression for obtaining an example of the correction movement speed of the present invention, and is not limited to this. The correction transition speed calculation method can be appropriately changed according to the gist of the present invention.

The position calculation unit 28005 calculates the front contour trajectory to be determined from the time difference obtained from the time t1, the time t3, and the time t4, the corrected moving speed, and the distances Y1 and Y2 in the vertical direction of the passage surface. The position (relative position) of three points on the orbit symmetrical about the path vertical axis is calculated. Further, it moves symmetrically with respect to the vertical axis of the passage, and calculates the positions of three points on the front contour trajectory to be discriminated.
Here, the position calculation program and the intervals Y1 and Y2 are stored in the storage unit 28006.

FIG. 29 is a contrast diagram showing the principle by which the positions of the three points are calculated.
The upper diagram (1) is the timing chart itself of FIG. 8001 is a timing at which the photoelectric sensor 1001 first detects the discrimination target. Similarly, 8002 and 8003 are timings when the photoelectric sensors 1003 and 1004 first detect the discrimination target.
29001 is the timing at which the photoelectric sensor 21001 first detects the discrimination target.
The lower figure (2) is a diagram in which the timings indicated by 8001, 8002 and 8003 are converted into coordinate positions 8004, 8005 and 8006 taken on the X axis indicating the passage parallel direction and the Y axis indicating the passage vertical direction. is there.
Y1 shown in the diagram (2) indicates a distance in the Y-axis direction between the photoelectric sensor 1001 and the photoelectric sensor 1003 (hereinafter referred to as a height distance), and is an existing value. Similarly, Y2 indicates the height interval between the photoelectric sensors 1003 and 1004 and is an existing value.
X1 is calculated by multiplying the time difference between the time t1 and the time t3 by the corrected moving speed V1 + 1 / 2α (t3−t1), the point on the contour trajectory on the front side of the height of the photoelectric sensor 1001, and the photoelectric sensor. A distance in the X-axis direction from a point on the contour track on the front side having a height of 1003 is shown.
Similarly, x2 is calculated by multiplying the time difference between time t3 and time t4 by the corrected movement speed V1 + 1 / 2α (t4−t3), the point on the contour trajectory on the front side of the height of the photoelectric sensor 1003, and the photoelectric sensor. A distance in the X-axis direction from a point on the contour track on the front side having a height of 1004 is shown.
Therefore, 8004, 8005, and 8006 indicate the relative positions of the three points on the contour contour on the front side of the discrimination target indicated by the graph 8007 shown on the XY coordinates (however, they are symmetric with respect to the Y axis). For example, if a position (X0, Y0) indicated by 8004 is assumed, 8005 is represented as (X0 + (t3-t1) · [V1 + 1 / 2α (t3-t1)], Y0-Y1). Similarly, 8006 is represented as (X0 + (t4−t3) [V1 + ½α (t4−t3)], Y0−Y1−Y2).
The relative position calculation method shown here is an expression for obtaining an example of the relative position of the present invention, and is not limited to this. The relative position calculation method can be appropriately changed according to the gist of the present invention.

Returning to the description of FIG.
The wheel feature extraction unit 28007 performs a known geometric operation on the relative positions (8008, 8009, and 8010) of the three points on the front contour trajectory to be discriminated calculated by the position calculation unit 28005 to obtain a circular trajectory. The diameter is calculated, and the wheel features are extracted as the size of the track.
Here, a geometric calculation program including an equation of a circular orbit, an elliptical orbit, and the like is stored in the storage device 28008.

  The output unit 28809 outputs a feature indicating the size of the wheel to a vehicle discrimination computer (not shown).

FIG. 30 is a flowchart from when a discrimination target moving toward the checkout machine is detected by moving through the bicycle parking lot to output to the wheel feature extraction computer.
In the beam state reception step S30001, the beam state reception unit receives the ON state signal and the OFF state signal of the photoelectric sensor groups 1001 to 1004 and 21001.
In the movement speed calculation step S30002, the first movement speed calculation unit calculates the movement speed V1 to be determined.
In the long section average moving speed calculation step S30003, the long section average moving speed calculation unit calculates the long section average moving speed VA to be discriminated.
In the speed correction step S30004, the speed correction unit calculates an acceleration trend α indicating a trend of acceleration or deceleration from the difference between VA and V1, and corrects the speed passing through each of the photoelectric sensors 1001 to 1004, and each photoelectric sensor A corrected moving speed is generated.
In the position calculation step S30005, the position calculation unit calculates the positions of three points on the front contour trajectory to be determined by the calculation method described above.
In the wheel feature calculation step S30006, the wheel feature calculation unit extracts the diameter as a feature indicating the size of the wheel.
In the output step S30007, the output unit outputs a feature indicating the size of the wheel to another system such as a vehicle discrimination system incorporating the wheel feature extraction system of this embodiment.

Here, it is assumed that the wheels of the two-wheeled vehicle enter the passage straight without tilting, and the contour can be regarded as a circle. When it is assumed that the contour is an ellipse due to an inclination or the like, the number of photoelectric sensors arranged in a line in the vertical direction of the passage surface is set to four or more, and the position on the front contour trajectory calculated by the position calculation unit 28005 is set to 4 The wheel feature extraction unit 28007 may perform a known geometric operation and calculate the minor axis or the major axis of the elliptical trajectory as appropriate.

The fourth embodiment is also a modified example of the first embodiment. This is a wheel feature extraction system in which the points on the front side of the contour to be discriminated are calculated more precisely and the wheel feature extraction accuracy is higher than that in the first embodiment. Portions that are common to the first embodiment will be described using the same reference numerals as much as possible, omitting redundant descriptions, and focusing on portions that are unique to the third embodiment.
FIG. 31 is an installation image diagram of the wheel feature extraction system according to the fourth embodiment introduced into the entrance / exit passage of the bicycle parking lot. The upper stage is a plan view and the lower stage is a side view.

FIG. 31 is an installation image diagram of the wheel feature extraction system of the third embodiment introduced into the entrance / exit passage of the bicycle parking lot. The upper stage is a plan view and the lower stage is a side view.
A plurality of photoelectric sensors 1001 to 1004 and 31001 to 10004 are arranged on both sides of a passage 1005 serving as a movement path of the object. In each photoelectric sensor, a light-emitting device that emits a beam from one side and a light-receiving device that receives the beam on the other side are paired so as to face each other.
The photoelectric sensors 1001, 1003 and 1004 are arranged in a line in the direction perpendicular to the passage surface. Similarly, photoelectric sensors indicated by 31001, 31003, and 31004 are also arranged in a line in the direction perpendicular to the passage surface.
Here, the photoelectric sensors 1001 and 31001 are set at an interval XL in the direction parallel to the passage surface. XL is set longer than XS described later.
The interval between the photoelectric sensor 1001 and the photoelectric sensor 1003 is Y1, and the interval between the photoelectric sensor 1003 and the photoelectric sensor 1004 is Y2. The interval between the photoelectric sensor 31001 and the photoelectric sensor 31003 is also Y1, and the interval between the photoelectric sensor 31003 and the photoelectric sensor 31004 is also Y2.
The photoelectric sensor 1002 is installed at a position parallel to the passage surface from the photoelectric sensor 1001 by a short parallel distance XS in the traveling direction of the two-wheeled vehicle 1007. Similarly, the photoelectric sensor 31002 is installed at a position parallel to the passage surface and separated from the photoelectric sensor 31001 by a short parallel interval XS in the traveling direction of the two-wheeled vehicle 1007.
Reference numeral 1008 denotes a wheel of the motorcycle 1007. The photoelectric sensors 1001, 1002, 31001, and 31002 arranged at the highest position are provided to have a height equal to or less than the radius of the smallest wheel among the assumed wheels 1008.
Here, photoelectric sensors 1002 and 31002 are installed to calculate movement speeds V1 and V2, which will be described later. If the movement speed is known in advance, 1001, 1003 and 1004 and 31001, 31002 and 31004 A sensor group may be constituted by only six photoelectric sensors. In that case, it is good also as a wheel feature extraction computer which does not include the 1st movement speed calculation part and 2nd movement speed calculation part which are demonstrated later.

  The photoelectric sensors 1001 to 1004 and 31001 to 10004 transmit the on state and off state of the beam crossing the passage to a wheel feature extraction computer (not shown).

  FIG. 32 is a diagram showing the positional relationship between the wheel and the photoelectric sensor group at time t1 when the wheel 1008 transitions to the ON state of the photoelectric sensor 1001 beam. Reference numeral 2001 denotes a contact point between the beam of the photoelectric sensor 1001 and a wheel.

  FIG. 33 is a diagram showing the positional relationship between the wheels and the photoelectric sensor group at time t2 at the moment when the beam of the photoelectric sensor 1002 is turned on by the wheel 1008. Reference numeral 3001 denotes a contact point between the beam of the photoelectric sensor 1002 and the wheel.

  FIG. 34 is a diagram showing the positional relationship between the wheels and the photoelectric sensor group at time t3 at the moment when the beam of the photoelectric sensor 1003 is changed to the ON state by the wheel 1008. Reference numeral 4001 denotes a contact point between the beam of the photoelectric sensor 1003 and the wheel.

  FIG. 35 is a diagram showing a positional relationship between the wheels and the photoelectric sensor group at time t4 at the moment when the beam of the photoelectric sensor 1004 is turned on by the wheel 1008. Reference numeral 5001 denotes a contact point between the beam of the photoelectric sensor 1004 and the wheel.

  FIG. 36 is a diagram showing the positional relationship between the wheels and the photoelectric sensor group at time t5 at the moment when the beam of the photoelectric sensor 31001 changes to the ON state. Reference numeral 36001 denotes a contact point between the beam of the photoelectric sensor 31001 and the wheel.

  FIG. 37 is a diagram showing the positional relationship between the wheels and the photoelectric sensor group at time t6 at the moment when the beam of the photoelectric sensor 31002 is turned on by the wheel 1008. Reference numeral 37001 denotes a contact point between the beam of the photoelectric sensor 31002 and the wheel.

  FIG. 38 is a diagram showing a positional relationship between the wheels and the photoelectric sensor group at time t7 at the moment when the beam of the photoelectric sensor 31003 is turned on by the wheel 1008. Reference numeral 38001 denotes a contact point between the beam of the photoelectric sensor 31003 and the wheel.

  FIG. 39 is a diagram illustrating the positional relationship between the wheels and the photoelectric sensor group at time t8 at the moment when the beam of the photoelectric sensor 31004 is turned on by the wheel 1008. Reference numeral 39001 denotes a contact point between the beam of the photoelectric sensor 31004 and the wheel.

(Timing chart)
FIG. 40 is a timing chart showing the off state, the on state, and the state transition of the photoelectric sensors 1001 to 1004 and 31001 to 10004. The horizontal axis 40005 indicates the progress of time.
Reference numeral 6001 is a chart showing the off state of the photoelectric sensor 1001 at a high level and the on state at a low level. Reference numeral 6006 indicated by a dashed arrow indicates a time t1 at which the photoelectric sensor 1001 transitions from the off state to the on state.
Reference numeral 6002 is a chart showing the off state of the photoelectric sensor 1002 at a high level and the on state at a low level. Reference numeral 6007 indicated by a dashed arrow indicates a time t2 at which the photoelectric sensor 1002 transitions from the off state to the on state.
Reference numeral 6003 is a chart showing the off state of the photoelectric sensor 1003 at a high level and the on state at a low level. Reference numeral 6008 indicated by a dashed arrow indicates a time t3 when the photoelectric sensor 1003 transitions from the off state to the on state.
Reference numeral 6004 is a chart showing the off state of the photoelectric sensor 1004 at a high level and the on state at a low level. Reference numeral 6009 indicated by a dashed arrow indicates a time t4 when the photoelectric sensor 1004 transitions from the off state to the on state.
Reference numeral 40001 is a chart showing the off state of the photoelectric sensor 31001 at a high level and the on state at a low level. Reference numeral 40006 indicated by a dashed arrow indicates a time t5 when the photoelectric sensor 31001 transitions from the off state to the on state.
Reference numeral 40002 is a chart showing the off state of the photoelectric sensor 31002 at a high level and the on state at a low level. Reference numeral 40007 shown by a broken line arrow indicates a time t6 when the photoelectric sensor 31002 transitions from the off state to the on state.
40003 is a chart showing the off state of the photoelectric sensor 31003 at a high level and the on state at a low level. Reference numeral 40008 indicated by a broken line arrow indicates a time t7 at which the photoelectric sensor 31003 transitions from the off state to the on state.
40004 is a chart showing the off state of the photoelectric sensor 31004 at a high level and the on state at a low level. Reference numeral 40009 indicated by a dashed arrow indicates a time t8 when the photoelectric sensor 1004 transitions from the off state to the on state.

(Functional configuration of wheel feature extraction computer)
FIG. 41 is a functional configuration diagram of the wheel feature extraction computer 41000.
The beam state reception unit 41001 receives the on state signal and the off state signal transmitted from the photoelectric sensors 1001 to 1004 and 31001 to 10004. The time zone in which the on-state signal is received is indicated at a low level on the timing chart. On the other hand, the time zone in which the off-state signal is received is indicated by a high level.
The first moving speed calculation unit 41002 calculates a moving speed V1 to be determined from a time difference between time t1 and time t2 and an existing value XS according to a predetermined speed calculation formula.
The second moving speed calculation unit 41003 calculates the moving speed V2 to be determined from the time difference between the time t5 and the time t6 and the existing value XS according to a predetermined speed calculation formula.
The position calculation unit 41004 calculates the front contour trajectory to be determined from the time difference obtained from the time t1, the time t3, and the time t4, the moving speed V1, and the gaps Y1 and Y2 in the vertical direction of the passage surface. The position (relative position) of three points on the orbit symmetrical about the path vertical axis is calculated. Further, it moves symmetrically with respect to the vertical axis of the passage, and calculates the positions of three points on the front contour trajectory to be determined.
Here, the position calculation program and the intervals Y1 and Y2 are stored in the storage unit 41005.

  The principle that the positions of the three points are calculated is the same as in the first embodiment.

The wheel feature extraction unit 41006 performs a known geometric operation on the relative positions of the three points on the front contour trajectory to be discriminated calculated by the position calculation unit 41004, calculates the diameter of the circular trajectory, Extract features as trajectory size. A geometric calculation program including circular or elliptical orbital equations for executing the geometric calculation is stored in the storage device 41007.

The verification unit 41008 relates to the front contour trajectory and the path vertical axis to be discriminated from the time difference obtained from t5, t7 and t8, the moving speed V2, and the distances Y1 and Y2 in the path surface vertical direction which are the existing values. The positions (relative positions) of the three points on the symmetrical trajectory are calculated, and the diameter of the circle formed by these three points is extracted.
The verification unit 41008 continues the operation to calculate and compare the ratios of the diameters related to t5, t7, and t8 and the size of the diameter extracted by the wheel feature extraction unit.
Here, when the ratio between the two is 120% or more or 80% or less, a signal indicating abnormality is sent to the output unit 41209. When it falls within the range of 80% to 120%, the value of the diameter calculated by the wheel feature extraction means is sent to the output unit.
Here, the abnormality is detected when the fluctuation of the ratio of the diameter values is outside the range of ± 20%, but the abnormality detection of the verification unit of the present invention is not limited to this. For example, by appropriately adopting characteristics such as the circumference, the abnormal range can also be changed as appropriate, such as changing the numerical value to ± 10% or changing the actual size difference to ± 5 cm.

FIG. 42 is symmetrical with respect to the front contour trajectory to be discriminated and the path vertical axis from the time difference obtained from t5, t7 and t8, the moving speed V2, and the distances Y1 and Y2 in the path surface vertical direction which are the existing values. This shows the principle of calculating the position (relative position) of three points on the orbit of
The upper diagram (1) is an excerpt of the portion related to times t5 to t8 from the timing chart of FIG. 4001 is the timing at which the photoelectric sensor 31001 first detects the discrimination target. Similarly, 4002 and 4003 are timings when the photoelectric sensors 31003 and 31004 first detect the discrimination target.
The lower figure (2) is a diagram obtained by converting the timings indicated by 42001, 42002 and 42003 into coordinate positions 42004, 42005 and 42006 taken on the X axis indicating the passage parallel direction and the Y axis indicating the passage vertical direction. is there.
Y1 shown in FIG. 2 indicates the Y-axis direction interval (hereinafter referred to as height interval) between the photoelectric sensor 31001 and the photoelectric sensor 31003, and is an existing value. Similarly, Y2 indicates the height interval between the photoelectric sensors 31003 and 31004 and is an existing value.
Further, x′1 is calculated by multiplying the time difference between time t5 and time t7 by the moving speed V2, and a point on the contour trajectory on the front side of the height of the photoelectric sensor 31001 and the front side of the height of the photoelectric sensor 31003. The distance in the X-axis direction with respect to the points on the contour trajectory is shown. Similarly, x′2 is calculated by multiplying the time difference between the time t7 and the time t8 by the moving speed V2, the point on the contour trajectory on the front side of the height of the photoelectric sensor 31003, and the front side of the height of the photoelectric sensor 31004. The distance in the X-axis direction with respect to the points on the contour trajectory is shown.
Therefore, 42004, 4005, and 42006 indicate the relative positions of the three points on the contour trajectory on the front side of the discrimination target indicated by the graph 42007 shown on the XY coordinates (however, they are symmetric with respect to the Y axis). For example, if the position (X0, Y0) shown at 4004 is shown, 42005 is represented as (X0 + (t7-t5) V2, Y0-Y1). Similarly, 42006 is represented as (X0 + (t8-t7) V2, Y0-Y1-Y2).
The relative position shown here is an expression for obtaining an example of the relative position of the present invention, and is not limited to this. The relative position can be appropriately changed according to the gist of the present invention.

Returning to the description of FIG.
The output unit 41209 outputs a feature or abnormality occurrence indicating the size of the wheel received from the verification unit to a vehicle discrimination computer (not shown).

Here, it is assumed that the wheels of the two-wheeled vehicle enter the passage straight without tilting, and the contour can be regarded as a circle. When it is assumed that the contour is an ellipse due to an inclination or the like, the number of photoelectric sensors arranged in a line in the vertical direction of the passage surface is set to four or more, and the position on the front contour trajectory calculated by the position calculation unit 41004 is 4 In this case, the wheel feature extraction unit 41006 may perform a change such as performing a known geometric calculation and calculating the minor axis or major axis of the elliptical orbit.

By adopting the wheel feature extraction system (object feature extraction system used for vehicle type discrimination) with such a configuration, it is possible to detect a fraud such as impersonation on the wheel as an abnormality, and to achieve a robust system with few malfunctions. is there.

  The invention disclosed in the present invention can be used for discriminating vehicle types at bicycle parking lots, discriminating members in factories and the like. In particular, it is expected to be used for discriminating vehicle types in bicycle parking lots where motorcycles and bicycles are mixed and used to store motorcycles.

1001 Photoelectric sensor 1002 Photoelectric sensor 1003 Photoelectric sensor 1004 Photoelectric sensor 1005 Movement path 1008 Wheel

Claims (13)

A plurality of beams are moved relative to the beam sensor by a conveying means including a belt conveyor so that an object having a circular outline is an object, and the emitted beam travels perpendicularly to the movement path of the relative translation. A system in which beam sensors are arranged to form a group of beam sensors, and the characteristics of the object are extracted from a plurality of detection patterns of the beam, and a) at least three of the beam sensor groups move. Arrangement is set in the vertical direction with respect to the route plane with a predetermined vertical interval, and b) between detection times when the beam sensors arranged in the vertical direction detect the front side in the traveling direction of the object. The time difference, the moving speed of the object, the vertical distance of the existing value, and the horizontal distance in the direction of the moving path of the three points on the contour relating to the detection of the three beams are at each point. 3 points on the front contour of the object from the relationship that the vertical difference in the vertical direction in the vertical direction of the movement path of the three points is a product of the time difference between the detection times and the moving speed. A position calculating means for calculating the relative position of the object, and c) a feature extracting means for calculating the characteristics of the object by using a circular trajectory passing through the three points as the contour of the object by a predetermined geometric calculating means. An object feature extraction system characterized by comprising. The beam sensor is moved relative to the beam sensor by a conveying means including a belt conveyor, and an object having an elliptical contour is set as a target, and a plurality of beams are emitted so as to travel perpendicularly to the movement path of the relative translation. A system in which beam sensors are arranged to form a group of beam sensors, and features of the object are extracted from a plurality of detection patterns of the beam, and a) at least four of the beam sensor groups move Arrangement is set in the vertical direction with respect to the route plane with a predetermined vertical interval, and b) between detection times when the beam sensors arranged in the vertical direction detect the front side in the traveling direction of the object. The time difference, the moving speed of the object, the vertical distance of the existing value, and the horizontal distance in the direction of the moving path of the four points on the contour relating to the detection of the four beams are at each point. 4 points on the contour on the front side of the object from the relationship that the vertical distance in the vertical direction of the movement path of the four points is an existing value. A position calculation means for calculating the position of the object, and c) a feature extraction means for calculating a feature of the object by using a predetermined geometric calculation means, with the elliptical trajectory passing through the four points as the outline of the object, D) At least one beam sensor in the group of beam sensors is arranged at a parallel interval shorter than the vertical interval in the traveling direction parallel to the movement path plane from the uppermost beam sensor in the vertical direction. E) a detection time when the uppermost beam sensor detects the front side of the object in the traveling direction, and a beam sensor arranged in parallel with the traveling direction from the uppermost beam sensor travels the object. A moving speed calculating means for calculating the moving speed of the object from the time difference with the detection time when the front side is detected is provided, and the moving speed of the object used by the position calculating means is calculated by the moving speed calculating means. An object feature extraction system characterized by speed. In the object feature extraction system, further, a) at least one beam sensor in the group of beam sensors is shorter than the vertical interval in a traveling direction parallel to the movement path plane from the uppermost beam sensor in the vertical direction. B) a detection time when the uppermost beam sensor detects the front side of the object in the traveling direction, and a beam sensor disposed in parallel with the traveling direction from the uppermost beam sensor. The moving speed calculating means for calculating the moving speed of the object from the time difference with the detection time when the front side of the moving object is detected is provided, and the moving speed of the object used by the position calculating means is the moving speed calculation. The system according to claim 1, wherein the means is a calculated moving speed. The system according to claim 1 or 2, wherein the relative translation is generated by the group of beam sensors that translate with respect to the fixed object. The uppermost position is set such that the height from the movement path plane is equal to or less than half of the diameter of the circle formed by the contour of the object in the vertical direction of the movement path plane. The system according to claim 1 The uppermost position is set such that the height from the moving path plane is equal to or less than half of the diameter of the ellipse formed by the contour of the object in the direction perpendicular to the moving path plane. The system according to claim 2 At least one beam sensor arranged in parallel in the traveling direction and further arranged in parallel with the movement path plane and at a parallel interval longer than the short parallel interval in the traveling direction; From the time difference between the detection time when the uppermost beam sensor detects the front side in the traveling direction of the object and the detection time when the beam sensors arranged at a long parallel interval detect the front side in the traveling direction of the object An average moving speed calculating means for calculating an average moving speed of a long section of the object is calculated, a trend of acceleration / deceleration is calculated from a comparison between the one or more average moving speeds and the moving speed, and the acceleration / deceleration The system according to claim 3, further comprising a speed correcting unit that corrects the moving speed based on a trend. At least one beam sensor arranged in parallel in the traveling direction and further arranged in parallel with the movement path plane and at a parallel interval longer than the short parallel interval in the traveling direction; From the time difference between the detection time when the uppermost beam sensor detects the front side in the traveling direction of the object and the detection time when the beam sensors arranged at a long parallel interval detect the front side in the traveling direction of the object An average moving speed calculating means for calculating an average moving speed of a long section of the object is calculated, a trend of acceleration / deceleration is calculated from a comparison between the one or more average moving speeds and the moving speed, and the acceleration / deceleration The system according to claim 2, further comprising a speed correction unit that corrects the moving speed based on a trend. At least three beam sensors including the beam sensors arranged in the traveling direction parallel to the movement path surface from the at least three beam sensors and spaced apart from each other by the long parallel distance are Further, one beam sensor is arranged at a short parallel interval from the beam sensor arranged at a long parallel interval in a traveling direction in parallel with the movement path plane. a) The detection time when the beam sensor arranged at the long parallel interval detected the front side of the object in the traveling direction and the beam sensor arranged at the long parallel interval were arranged in the traveling direction in parallel. Second moving speed calculation for calculating the second moving speed of the object from the time difference from the detection time when the further one beam sensor detects the front side of the object in the traveling direction. B) is calculated from the time difference between detection times when the at least three beam sensors including the beam sensors arranged at long parallel intervals detected the front side in the traveling direction of the object and the moving speed. When a deviation exceeding a predetermined ratio is detected by comparing the positional relationship of the three points on the front contour of the object with the positional relationship of the three points calculated by the position calculating means The system according to claim 5, further comprising verification means for executing At least four beam sensors including the beam sensors arranged in the traveling direction parallel to the movement path plane from the at least four beam sensors and spaced apart from each other by the long parallel distance are perpendicular to the movement path plane. Further, one beam sensor is arranged at a short parallel interval from the beam sensor arranged at a long parallel interval in a traveling direction in parallel with the movement path plane. a) The detection time when the beam sensor arranged at the long parallel interval detected the front side of the object in the traveling direction and the beam sensor arranged at the long parallel interval were arranged in the traveling direction in parallel. Second moving speed calculation for calculating the second moving speed of the object from the time difference from the detection time when the further one beam sensor detects the front side of the object in the traveling direction. B) is calculated from the time difference between detection times when the at least four beam sensors including the beam sensors arranged at long parallel intervals detected the front side in the traveling direction of the object and the moving speed. The positional relationship between the four points on the front contour of the object and the positional relationship of the four points calculated by the position calculating means, and a notification operation is performed when a deviation exceeding a predetermined ratio is detected The system according to claim 6, further comprising verification means for executing The delay time, which is the time difference between the center passage time at which the contour passes the center point of the cross section of the beam and the detection time, is determined according to the inclination of the tangent of the contour trajectory formed by the front side in the traveling direction. From the relationship of changing, it is estimated from the inclination of the contour passing through the beam and the moving speed determined by the height of the beam from the moving path surface and the trajectory obtained by the feature extracting means, and the detection time is determined as the center. 4. The system according to claim 2, further comprising detection time correction means for correcting the time to the passage time. 12. The apparatus according to claim 11, further comprising a time correction repeating unit that calculates the detection time by repeating the time correction based on the contour recalculated by the feature extraction unit using the time corrected detection time. The described system. The system according to claim 12, wherein the detection time is calculated by repeating the time correction a plurality of times.

Images