JPH04147399A - Crime prevention sensor - Google Patents

Abstract

PURPOSE:To improve the accuracy of a crime prevention sensor for use in an automatic taller machine or the like by counting a frequency of a user of the teller machine coming and going in a detecting area of the crime prevention sensor within an interval of no time to one time based on the output from a human body detecting sensor and by amplifying the degree of sensing of a malicious action if on time is long and a counted value is great. CONSTITUTION:The output from a pyroelectric sensor 1 that is a human body detecting sensor is supplied to the off delay timer 3 and counter 4 of a feature quantity extracting means 2. The timer 3 delays an input signal by a constant period of time, and its output is supplied to an ontime measuring section 5. The ontime measuring section 5 measures a period of time while the output from the timer 3 continues, and by subtracting a delay period of time from the period of time while the output from the timer 3 continues, a period of time while a person stays in the detecting area is measured. The output from the timer 3 and the output from the counter 4 are supplied to a fuzzy inference section 6. The fuzzy inference section 6 carried out fuzzy in accordance with a plurality of fuzzy rules, and as a result of the fuzzy inference, a signal representing a degree of malicious behavior of the person coming and going in the detecting area is exerted to an alarm output means 7 and a malicious display means 8.

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Field of Application] The present invention relates to a security sensor that is disposed inside an automatic teller machine, a safe, etc., and is used to prevent theft and the like.

[Conventional technology]

Conventional crime prevention sensors for detecting intruders include ultrasonic waves and microwaves, as disclosed in, for example, Japanese Patent Application Laid-open No. 1-297791 or 2-8997.

Security sensors such as hot-wire type and security sensors combining these have been proposed.

[Sequence of problems to be solved by the invention] However, if such a conventional security sensor is installed near an automatic teller machine, etc., anyone who enters the sensor's detection zone will be considered a suspicious person. I end up.

Therefore, as a sensor for preventing crime, there were many false alarms, and it was almost meaningless. In recent years, automatic teller machines have been increasingly installed not only in banks and post offices, but also in general businesses, and people often pass by the machines even after their operating hours have ended. Therefore, there is a need for a security sensor that can determine whether a person is suspicious.

The present invention has been made in view of the problems of conventional crime prevention sensors, and a technical object thereof is to provide a crime prevention sensor that can determine with high accuracy whether or not a person is a suspicious person.

[Means for Solving the Problems] The present invention divides the detection area into a large number of small areas, includes a human body detection sensor that detects the arrival of a person in each small area, and continuously provides the output of the human body detection sensor. an on-time measuring means for detecting the time during which the on-time is measured; a counter for counting the number of outputs within the output range of the on-time measuring means; The present invention is characterized by comprising a fuzzy inference unit that performs fuzzy inference to obtain a non-fuzzy definite value so as to reduce the degree of suspicion if the count value of the counter is small.

[Operation] According to the present invention having such characteristics, the number of times the sensor moves back and forth between the detection areas within the range of the ON time and the ON time is counted based on the output from the human body detection sensor. Fuzzy inference is performed such that if the on time is long and the count value of the counter is large, the degree of suspicion is increased, and if the on time is short and the count value of the counter is small, the degree of suspicion is decreased.

〔Example〕

FIG. 1 is a block diagram showing the overall configuration of a security sensor according to an embodiment of the present invention. In this embodiment, a pyroelectric sensor 1 is used as a human body detection sensor, and its output is given to an off-delay timer 3 and a counter 4 of the feature extracting means 2. Off-delay timer 3 outputs a human signal for a certain period of time T.
It is delayed by d, and its output is given to the on-time measuring section 5. The on-time measuring section 5 measures the time T' during which the output of the off-delay timer 3 continues, and calculates the delay time Td.
This is a measuring means that measures the time T during which a person exists in the detection area by subtracting , and its output is given to the fuzzy inference section 6. Further, the count value of the counter 4 is also given to the fuzzy inference section 6. The fuzzy inference unit 6 performs fuzzy inference according to a plurality of fuzzy rules described later, and provides a signal indicating the degree of suspiciousness of the incoming person to the alarm output means 7 and the suspiciousness level display means 8 as a result of the inference. The alarm output means 7 distinguishes between these signals to ensure safety and safety.
It outputs signals such as caution, warning, or danger to the outside. Further, the suspiciousness level display means 8 is for discriminating the signal and displaying the suspiciousness level.

The pyroelectric sensor 1 used in this embodiment is constructed by arranging a shielding plate 12 having a plurality of openings and a Fresnel lens 13 in the vicinity of a pyroelectric sensor body 11, for example, as shown in FIG. In this way, a detection area and a non-detection area passing through the opening can be provided as shown in the figure. By arranging such a pyroelectric sensor 1 near the automatic teller machine 14, for example, as shown in FIG. 3, a large number of detection small areas and non-detection areas are provided, which are indicated by a plurality of hatchings.

Next, the operation of this embodiment will be explained. Figures 4 and 5
The figure is a chart showing the operation of this embodiment. First, when a person comes near the automatic teller machine and moves near the detection area, the pyroelectric sensor 1 produces an output as shown in FIG. 4(a) or FIG. 5(a). . This signal is given to the off-delay timer 3, and from the off-delay timer 3, as shown in FIG. 4(b) or FIG. 5(b),
After the output of the pyroelectric sensor 1 stops, an output that turns off after a certain period of time is obtained. From this time T', the delay time T
The time T obtained by subtracting d is measured as the on time by the on time measuring section 5, and the on time information is given to the fuzzy inference section 6. Further, as shown in FIGS. 4(a) and 5(a), the number of pulses from the pyroelectric sensor 1 is counted by the counter 4. The output is then given to the fuzzy inference section 6.

Now, the fuzzy rule stored in the fuzzy inference section 6 is a rule for alarm output based on the ON time and the count value of the counter 4, and is expressed in the form of a 1fthen rule (if then).

The on time and the number of on times are evaluated as membership functions as shown in FIGS. 6(a) and (b). Here, PS indicates a small positive value, and PL indicates a large positive value. The alarm value also has four states (P
S, PMS, PML, PL). P.S.
indicates a safe state, PMS indicates a caution, PML indicates a warning, and PL indicates a dangerous state.

Now, the fuzzy inference section 6 performs fuzzy inference based on the following inference rules based on these human powers. The fuzzy inference rule is determined such that the shorter the on time X1 is and the fewer the number of on times X2 is, the safer the state is, and vice versa, the state is more dangerous and the alarm value is increased.

(Rule 1) If the on time X1 is PS and the number of on times X2 is PS, let the alarm output Z be PS. This rule 1 is simplified and expressed as follows.

■if X1=PSand X2=PSthen
Z = P S Similarly, other rules are expressed by the following formula.

■if XI-PLand X2=PSthen
Z=PMS ■if X1=PS and X2=PLthe
n Z=PML ■if X1=PL and X2=PLthe
n Z = PL Thus, four inference rules are defined. As shown in the flowchart of FIG. 7, in the fuzzy inference unit 6, when input signals XI and X2 are inputted from the on-time measurement unit 5 and the counter 4, respectively (step 21), the antecedent unit responds to this fuzzy rule. The degree to which it fits the membership function is determined (step 22.23). Then, the one with the lowest degree of fitness is selected and given to the consequent. The consequent part of the fuzzy inference unit 6 obtains a membership function by restricting the membership function of the output Z based on the selected fitness. These membership functions are superimposed by MAX synthesis processing to generate a synthesis output, and fuzzy inference is performed (step 24). Next, in step 25, the center of gravity of this combined output is calculated as a final output by the defassifier. Then, Z is outputted to the outside by the alarm output means 7 as a signal of safety, caution, warning, or danger, and simultaneously displayed by the suspiciousness level display means 8.

In this embodiment, a microcomputer is used to execute the fuzzy) 1# logic processing, but dedicated hardware such as a fuzzy computer or a fuzzy controller as shown in Japanese Patent Application Laid-open No. 63-123177 may also be used. Fuzzy inference may be executed using software and an alarm value may be output.

In this embodiment, the pyroelectric sensor is provided with one feature amount extraction means and outputted to the fuzzy inference section, but a plurality of pyroelectric sensors and feature amount extraction means are provided, and the logic of their output is It is also possible to provide the sum to the fuzzy inference section. Further, in this embodiment, a pyroelectric sensor is used as a human body detection sensor, but it is also possible to use other types of human body detection sensors such as an ultrasonic sensor.

〔Effect of the invention〕

As described above in detail, according to the present invention, when a person comes near an automatic teller machine or the like, the degree of suspiciousness can be estimated and output by fuzzy inference. Therefore, the accuracy of determining whether a person is a suspicious person can be improved. In particular, simply passing by the automatic teller machine after the operating hours of the automatic teller machine has ended will not cause an alarm output, making it possible to provide a highly reliable crime prevention sensor.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the overall configuration of a security sensor according to an embodiment of the present invention, FIG. 2 is a schematic diagram showing the configuration of a pyroelectric sensor 1 used in the security sensor, and FIG. 3 is a block diagram showing the configuration of a pyroelectric sensor 1 used in the security sensor. A diagram showing the detection area of a pyroelectric sensor installed nearby,
Figures 4 and 5 are time charts showing the operation of this embodiment, Figure 6 is a diagram showing membership functions of ON time, number of ON times, and their alarm outputs, and Figure 7 is a flowchart showing fuzzy inference processing. be. 1 -1 - Pyroelectric sensor 2 - Feature extracting means 3 - Off delay timer 4 - Counter 5 -11 - On time measuring unit 6 - Fuzzy inference unit Alarm output means Suspiciousness level display means

Claims (1)

[Claims] (1) The detection area is divided into many small areas, and a human body detection sensor detects the arrival of a person in each small area, and an on-time measurement that detects the time when the output of the human body detection sensor is continuously given. means, a counter for counting the number of outputs within the output range of the on-time measuring means, and if the outputs of the on-time measuring means and the counter are large, an output indicating suspicion is increased, and the on-time and the count value of the counter are increased. A crime prevention sensor comprising: a fuzzy inference unit that performs fuzzy inference to obtain a non-fuzzy definite value so as to reduce the degree of suspicion if the value is smaller.