JPH06215222A - Shape discriminating device for metal chip - Google Patents

Abstract

PURPOSE:To make it possible to identify thinner shapes such as surface ruggedness and holes in addition to the judgement of quality, thickness and size based upon the convensional peak detection of a waveform outputted from a detection circuit in the case of identifying a metal chip such as a coin by a coil sensor. CONSTITUTION:A feature extraction value (Va1-Vb1)X(Tb1-Ta1)<1/2> or the like obtained by differentiating an output waveform 19 outputted from a detection circuit 14 by a differential circuit 72 and correcting the number of peaks or a level difference between a peak and a trough in a differential waveform 23 by a metal chip passing speed is compared with a corresponding feature value stored in a memory 12 by a microcomputer 11, or the peak part of the waveform 19 is extracted and expanded through a peak expanding circuit 73 and the numbers of peaks and troughs, a level difference between a peak and a trough, the time width of a trough, etc., are extracted and compared with feature data stored in the memory 12.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus for identifying the shape of a metal piece using a coil sensor, such as a coin sorting apparatus used in a vending machine. In the drawings below, the same reference numerals indicate the same or corresponding parts.

[0002]

2. Description of the Related Art A conventional technique for a coin sorting device applied as a metal piece identifying device will be described. FIG. 1 is a sectional view showing an example of a principle structure of a sensor portion of a conventional coin sorting device. That is, as shown in FIG. 1, a ferrite core 2 having a diameter smaller than that of a coin 1 and a coil 3 inserted therein is attached to the wall surface of the metal piece passage 4 and used as a sensor. This sensor produces an alternating magnetic field. This coin sorting device extracts characteristics of the coin 1, for example, material, size and thickness. For this reason, one sensor is arranged for each of these features, and each sensor is connected to a separate detection circuit. These sensors are arranged on one side or both sides of the metal piece, and those having different connection methods (in-phase and anti-phase) and detection circuit methods are used.

FIG. 2 is an explanatory diagram of the circuit (waveform detection circuit) and output waveform of the sensor of FIG. In FIG. 2, 15 is the coil sensor described in FIG. 1, 16 is a circuit composed of a capacitor and an active element, and 17 is an envelope detection circuit. That is, the coil sensor 15 includes a capacitor and an active element 16
The oscillator circuit 13 is configured in combination with the above. At this time, the alternating magnetic field generated by the coil sensor 15 changes as the coin 1 passes through the coin passage 4 and between the sensors. The alternating magnetic field is converted into a voltage by the detection circuit 14, but is attenuated when a coin passes, and becomes an oscillating waveform 18. An output waveform 19 having the characteristics of a coin is taken out by detecting and smoothing the amplitude by the envelope wave circuit 17.

FIG. 3 is an explanatory diagram of a coin identifying method. In the memory 12 of FIG. 3, peak value data 63 of the waveform of coins to be selected, which are respectively obtained from different sets of the detection circuit 14 described in FIG. 2, are stored in advance. The three sets of detection circuits 14 have waveforms 19a, 19b, 19 for the passage of a coin, respectively.
Output c. The microcomputer 11 outputs each output waveform 19a, 19
The voltage values of b and 19c are sampled, and the peak values V1, V2 and V3 of the voltage are detected. Then, it compares the peak value data 63 of each coin in the memory 12 one by one. Voltage peak values V1, V2, V3 sampled at this time
However, if the corresponding stored peak value data a, b, c for the coin A is within the same range or within the error range, the discriminated coin 1 is discriminated as the coin A.

[0005]

However, the peak value of the voltage detected by the detection circuit of FIG. 2 represents the material, size, thickness, etc. of the coin, and represents the characteristics of the detailed shape of the surface of the coin 1. Was not. As a result, coins with similar patterns, edges, and similar material, size, and thickness were sometimes mistakenly selected.

Therefore, the present invention is to provide a metal piece shape identifying device capable of selecting metal pieces which are similar in material, size and thickness, but have different patterns and edge shapes, which are the above problems. It is an issue.

[0007]

In order to solve the above-mentioned problems, the shape identifying apparatus of claim 1 provides a high-frequency alternating magnetic field to a passage (such as a coin passage 4) through which at least a metal piece (such as 1A) passes. A coil sensor (15 or the like) to be generated and means (detection circuit 14 or the like) for detecting and smoothing the output of the coil sensor and outputting a smoothed signal (output waveform 19 or the like) representing the envelope of the output waveform. In the identification device for identifying the property of the metal piece from the peak value of the smoothed signal accompanying the passage of the metal piece, a means (differential circuit 72 etc.) for differentiating the smoothed signal is further provided, Features such as irregularities on the surface of are identified.

The shape identification device of claim 2 is the same as that of claim 1.
In the shape identifying apparatus described in (1), in order to absorb the difference in the passing speed of the metal piece, the calculated differential signal is normalized by the passing time or the square root of the passing time. A shape identifying apparatus according to claim 3 is the shape identifying apparatus according to claim 1 or 2, wherein the metal piece is a coin (1 or the like) and peak points of peaks and valleys of the differential signal waveform are detected. Feature data related to the lever metal piece is extracted.

Further, the shape identifying device according to claim 4 is a coil sensor (15 or the like) for generating a high frequency alternating magnetic field in a passage (the coin passage 4 or the like) through which at least a metal piece (1A or the like) passes, and the coil sensor. Means for detecting and smoothing the output and outputting a smoothed signal (such as the output waveform 19) representing the envelope of the output waveform, the peak of the smoothed signal accompanying the passage of the metal piece. And the peak point of the valley is detected to identify the shape of the metal piece.

The shape identifying device of claim 5 is the same as that of claim 4.
In the shape identifying device described in (1), the metal piece is a coin (1 or the like).

[0011]

[Action]

(1) Differentiation method: When the output waveform 19 obtained using the coil sensor 15 and the detection circuit 14 of FIG. 2 described in the prior art is differentiated, the amount of change in the voltage level per unit time can be detected. The features due to the detailed shape of the metal piece and the unevenness of the surface are greatly shown.

FIG. 4 is a comparison of the output waveform 19 and its differential waveform corresponding to the slow speed of passage of the metal piece. Here, 19-1 and 19-2 correspond to the waveform 19 output from the detection circuit 14 of FIG. 2 when the coin passing speed is slow and fast, respectively, and 23-1, 23-2 are waveform 1 respectively.
9-1 and 19-2 are differential waveforms. That is, the output waveform 19
The levels of -1, 19-2 do not change depending on the coin passing speed (thus the passing times 20-1, 20-2), but the level of the differential waveform differs depending on the speed at which the metal piece passes between the sensors, and the speed is slow. In this case, the level becomes low like the differential waveform 23-1, and when the passing speed is fast, the differential waveform 23
The level becomes higher like -2. In the present invention, the differential waveforms 23-1 and 23-2 are calculated with the passage times 20-1 and 20-2, and are constant regardless of the passage speed like the normalized differential waveforms 25-1 and 25-2. Normalize to obtain the waveform of. This normalized corrugation depends only on the detailed shape of the metal piece and the surface features. The type of metal piece is identified by utilizing this feature.

(2) In the case of the beak shape method: The enlarged waveform of the peak portion of the output waveform 19 obtained from the coil sensor 15 and the detection circuit 14 of FIG. 2 described in the prior art represents information on the surface of the metal piece. However, if the metal surface is deeply carved, the output signal at that portion decreases, and if the metal surface is exposed, the output waveform at that portion increases. This output signal represents surface information averaged over the area of the sensor rather than a fine pattern on the metal surface. An example of this is shown in FIG. In FIG. 5, 29 indicates the output waveform 19 as the threshold value Vth.
It is an enlarged waveform of the peak part cut out in. in this way,
The unevenness of the peak waveform 29 changes in accordance with the shape of the metal piece and the unevenness of the surface. The type of metal piece is identified by utilizing the characteristics of this waveform.

[0014]

FIG. 6 is a block circuit diagram showing a configuration as an embodiment of the present invention. In the present invention, in addition to the detection circuit 14 described in FIG. 2, a differentiation circuit 72 and a peak expansion circuit 73 for differentiating the output 19 of the detection circuit 14 are provided, and the detection circuit 14, the differentiation circuit 72 and the peak expansion circuit 73 are provided.
The microcomputer 11 inputs the output of 1 to perform the determination process of the metal piece 1A. Note that 11A is a determination processing means as a main function part of the microcomputer 11 related to the present invention. Note that 1A is a metal piece such as a coin to be identified.

FIGS. 7, 8 and 9 show the detection circuit 1 of FIG. 6, respectively.
4, an embodiment of specific circuits of the differentiating circuit 72 and the peak expanding circuit 73 will be shown. The detection circuit 14 in FIG. 7 has the same configuration as that described in FIG. That is, it is composed of an oscillating circuit 13 including a coil sensor 15 and a envelope detecting circuit 17 which are assembled in reverse phase. The feature calculation of the metal piece 1A when the differential processing is performed and when the waveform peak shape processing is performed will be described below.

(1) In the case of the differentiation method: As described above, the differentiation circuit 72 differentiates the output waveform 19 obtained by passing the metal piece 1A and outputs it to the microcomputer 11. The microcomputer 11 extracts the feature of the shape of the metal piece 1A from the differential waveform, compares it with the corresponding feature data of the metal piece 1A stored in the memory 12 in advance, and determines whether the data is the same or within the error range. Identify piece 1A.

Next, the differential waveform 23 executed by the microcomputer 11
The method of extracting features from is described. When the metal piece 1A is a coin, the output waveform 19 of the detection circuit 14 is differentiated to obtain a waveform 23 in FIG. 10, which has a characteristic attributed to the end of the metal piece 1A and has a valley B on the ten side. Two peaks of C and two peaks of D and F sandwiching the valley E on one side are generated. As described above, this mountain changes in size depending on the passing speed of the metal piece 1A, but can be extracted as a feature if corrected. Here, as the correction method, a method of calculating the passage time 20 of the metal piece 1A or the square root of the passage time is used.

The differential feature values DF1 to DF4 represented by the following equations (1) to (4) are examples of feature values that can be used for feature extraction of coin shapes.

[0019]

[Formula 1] Differential feature value DF1 = (V _a1 −V _b1 ) × (T _b1 −T _a1 ) ... (1) Differential feature value DF2 = (V _c1 −V _b1 ) × (T _c1 −T _b1 ). … (2) Differential feature value DF3 = (V _a1 −V _b1 ) × (T _b1 −T _a1 ) ^1/2 (3) Differential feature value DF4 = (V _c1 −V _b1 ) × (T _c1 − T _b1 ) ^1/2 (4) where V _a1 : Peak value of peak A (maximum value) V _b1 : Peak value of valley B (minimum value) V _c1 : Peak value of peak C (maximum value) T _a1 : Detection time of V _a1 T _b1 : Detection time of V _b1 T _c1 : Detection time of V _c1 Correction by the square root as in equations (3) and (4)
This is proposed in the present invention because the most stable signal is obtained experimentally.

By using these calculations, the characteristic of the shape of the metal piece 1A is corrected and extracted, and the metal piece 1A is selected. If the shape of the metal piece 1 is highly uneven and the differential waveform has a great feature, it is possible to select the metal piece 1A using V _a1 , V _b1 , and V _c1 as they are. In addition, as shown in FIG. 6, the number of peaks in the differential waveform 23, the level difference between peaks and valleys V _a1 −V _{b1, and the} like may be used as feature values.

(2) In case of the peak shape method: The peak expanding circuit 73 amplifies the peak of the output waveform 19 obtained by passing the metal piece 1A and outputs the peak expanding waveform 29 to the microcomputer 11. The microcomputer 11 extracts the characteristic of the shape of the metal piece 1A from the shape of the peak waveform, compares it with the characteristic data of the metal piece 1A stored in the memory 12 in advance, and determines whether the data is the same or within the error range. Identify 1A.
Next, a method of extracting a feature from the peak waveform by the microcomputer 11 will be described.

FIG. 11 shows a peak expansion waveform 29 in which the peak of the output waveform 19 of the detection circuit 14 when the metal piece is a coin is cut at the threshold voltage Vth34 and inverted and amplified. This peak waveform 29 represents the shape of the metal piece 1A. If the metal piece 1A is flat, the extreme value B in FIG.
C and D also become flat without recessing, and if the metal piece 1A has unevenness, the extreme values B, C, and D also become uneven according to the shape. Generally, in the case of a circular metal piece, for example, a coin, a pattern is often arranged in a concentric uniform pattern when viewed macroscopically. Therefore, the peak waveform 29 also has a concave shape in the center and a peripheral shape. Become. If there is a hole in the center of the metal piece, the dent in the center becomes larger and is peculiar to coins, and the influence of the outer shape can be seen. Further, the width T _d2 −T _b2 between the valley B and the valley D of the waveform in FIG. 11 represents the interval between the irregularities of the metal piece. However, since the time changes depending on the passing speed of the metal piece, it can be extracted by correcting the effective time 31 (= T _e2 −T _a2 ) with the size of the metal piece. The size of the metal piece can be detected by the conventional peak level method. The correction formula is as shown in the following formula (5).

[0023]

[Equation 2] Interval of unevenness of metal piece = (T _d2 −T _b2 ) × (size of metal piece) / (effective time 31) (5) As described above, various waveforms are obtained depending on the shape of the metal piece. Therefore, the difference in the surface shape of the metal piece can be detected by extracting the information such as the number of peaks, the level difference between peaks and valleys, and the temporal width of valleys.

[0024]

According to the present invention, the output of the coil sensor is detected and smoothed, a smoothed signal indicating the envelope of the output waveform is obtained, and a metal piece such as a coin is identified from the peak value thereof. A signal obtained by differentiating the smoothed signal or peaks and valleys of the smoothed signal is detected to extract finer features of the metal piece, and if necessary, the moving speed of the metal piece is included in the extracted data. Since the correction of is performed, it is possible to identify the detailed shape of the metal piece that could not be detected only by conventional peak value detection, the selection of metal pieces becomes more accurate, and the selection of metal pieces of various shapes can be performed. it can.

[Brief description of drawings]

FIG. 1 is a cross-sectional view showing an example of the principle structure of a coil sensor unit.

FIG. 2 is a diagram showing an output waveform of a coil sensor circuit.

FIG. 3 is an explanatory diagram of a coin identifying method.

FIG. 4 is a diagram showing a differential waveform of a coil sensor output corresponding to a difference in passing speed of metal pieces.

FIG. 5 is a diagram showing a peak expansion waveform of a coil sensor output.

FIG. 6 is a functional configuration diagram as an embodiment of the present invention.

FIG. 7 is a concrete circuit diagram of the detection circuit.

FIG. 8 is a concrete circuit diagram of the differentiating circuit.

FIG. 9 is a concrete circuit diagram of the peak expansion circuit.

FIG. 10 is a detailed diagram of the differential waveform as well.

FIG. 11 is a detailed diagram of a peak expansion waveform.

[Explanation of symbols]

 1 Coin 1A Metal Piece 2 Ferrite Core 3 Coil 4 Coin Passage 11 Microcomputer 11A Judgment Processing Means 12 Memory 13 Oscillation Circuit 14 Detection Circuit 15 Coil Sensor 17 Envelope Detection Circuit 19 (19a-19c, 19-1, 19-2) Output Waveform 20 (20-1, 20-2) Coin transit time 23 Differential waveform 29 Peak expansion waveform 31 Effective time 63 Peak value data 72 Differentiation circuit 73 Peak expansion circuit

Claims (5)

[Claims] 1. A coil sensor for generating a high-frequency alternating magnetic field in a passage through which at least a metal piece passes, and means for detecting and smoothing an output of the coil sensor and outputting a smoothed signal representing an envelope of the output waveform. In the identification device for identifying the property of the metal piece from the peak value of the smoothed signal accompanying the passage of the metal piece, further comprising means for differentiating the smoothed signal, and the like on the surface of the metal piece. A shape identification device for a metal piece, characterized in that the features of the metal piece are identified. 2. The shape identifying apparatus according to claim 1, wherein in order to absorb a difference in passing speed of the metal piece, the differential signal calculated as above is normalized by a passing time or a square root of the passing time. Characteristic shape identification device for metal pieces. 3. The shape identifying apparatus according to claim 1 or 2, wherein the metal piece is a coin, peak points of peaks and valleys of the differential signal waveform are detected, and characteristic data relating to the metal piece is detected. A shape identifying device for a metal piece, characterized by being extracted. 4. A coil sensor for generating a high frequency alternating magnetic field in a passage through which at least a metal piece passes, and means for detecting and smoothing an output of the coil sensor and outputting a smoothed signal representing an envelope of the output waveform. A shape identifying device for a metal piece, comprising: a peak of a peak and a valley of the smoothed signal when the metal piece passes; and the shape of the metal piece is identified. 5. The shape identifying device for a metal piece according to claim 4, wherein the metal piece is a coin.