EA044353B1 - METHOD FOR DETECTING OBJECT MOVEMENT - Google Patents

Description

Field of technology

The invention relates to security methods with protection against false alarms and relates to a method for detecting the movement of objects using a passive infrared sensor. The invention can be used for security alarm systems.

Terms used in the description

A false positive is an erroneous alarm.

Sensitivity zone - active zone, determines the adjustable detection range.

Signal intensity is the rate of change of the signal.

A function is a rule that associates each element from the first set - the domain of definition - with an element from another set - the domain of values.

Derivative is a concept in differential calculus that characterizes the rate of change of a function.

Threshold value is a parameter value limit that cannot be exceeded during normal operation.

Reference value is a certain value of a quantitative characteristic that can serve as a standard for measurements.

The Pearson correlation coefficient (denoted by r) is an indicator of correlation (linear dependence) between two variables X and Y, which takes values from -1 to +1 inclusive.

Discretization is the transformation of functions of continuous variables into functions of discrete variables, from which the original continuous functions can be restored with a given accuracy.

State of the art

The problem of protection against false alarms in security systems constantly faces security system manufacturers. The street is always in motion: the wind sways the leaves of trees and bushes heated by the sun, animals and birds walk around the area, insects land on the sensor body. The diversity of flora and fauna creates a flurry of obstacles, preventing an outdoor motion sensor from accurately recognizing the specific type of object that may be a threat, in most cases a person. The same goes for indoor sensors when there may be pets in buildings.

There is a known method and device for reducing the number of false alarms due to white light in a motion detection system (EP1544823B1 dated July 18, 2007). The method is implemented through a motion detection system including a first sensor sensitive to infrared light in at least one sensitivity zone and generating a first output signal that represents the detected level of infrared light. The second sensor is sensitive to visible light and generates a second output signal that represents the detected level of visible light. The second sensor is located next to the first sensor. The processor is programmed to generate an alarm based on the first and second output signals. An alarm is generated when the first and second conditions are met. The first condition is satisfied when the first output signal indicates that movement has occurred in at least one sensitivity zone. The second condition is satisfied when the second output signal does not correlate with the first output signal.

This analogue has disadvantages for solving the problem, since it reveals a method for reducing false alarms specifically for white light, while using different types of sensors that are sensitive to different parts of the radiation spectrum. This method does not allow you to pre-select a specific type of object to which the alarm will be triggered, and, as a result, does not provide error-free confirmation of the detection of movement of a specific type of obstacle.

A motion detection method is known that can be implemented using an infrared motion detection sensor, including an infrared sensor configured to receive infrared radiation from a monitored area within a detection field of view and to generate a waveform output signal that indicates time-dependent changes in the received infrared radiation. in response to the movement of an object in the detection field of view. In another embodiment, the infrared motion detection sensor may also include a processor configured to receive a waveform output generated by the infrared sensor and perform signal analysis based on the waveform output to determine whether a motion event has occurred, which may include finding matches between a waveform output and one or more waveform reference signals (US20200111335A1 04/09/2020).

The basis of this method is a comparison with amplitude components, which does not ensure the reliability and accuracy of signal processing.

There is a known method for testing body motion detection using a passive infrared sensor, including generating motion data based on a first signal received in a main detection channel, the first signal indicating possible body movement; generating verification data based on the second signal or absence of the second signal received on the secondary detection channel; verifying that the first signal represents body movement based on comparison of the movement data with the verification data; wherein the primary detection channel is configured to receive a first predetermined range of signals, including at least some infrared signals, and the secondary detection channel is configured to receive a second predetermined range of signals, excluding at least some infrared signals through the infrared blocker ( US2021080482A1 dated 03/18/2021).

This method was based on improving the design of the sensor, which made it possible to cut off false alarms caused by interference and heating of the sensor, but did not provide the ability to accurately and reliably determine the type of moving object.

Objective of the invention

The basis of the invention is the task of developing a method for detecting the movement of an object, which will allow you to quickly and reliably separate types of moving objects in order to accurately select the required object for which an alarm will be triggered.

The technical result that is achieved in this case is to increase the accuracy of determining the type of object to reduce the number of false alarms of sensors in security systems due to several-stage signal processing.

Disclosure of the Invention

The problem is solved by the fact that a method for detecting the movement of an object is proposed, which includes detecting movement within the sensitivity zones of at least two pyrosensors and generating a signal from these pyrosensors, then measuring the signal intensity by determining the rate of change of the signal during a measurement period of maximum 150 ms from each pyrosensor . , in the case where the degree of similarity of signal shapes is lower than the reference value, the signals are distributed into at least 8 spectral components in the form of sinusoidal functions and compared with each other and with the reference values by establishing the percentage of similarity of each spectral component of each signal with the reference values by determining the total share the areas of two spectral distributions, where one distribution is a predetermined reference, and the other is obtained from the signal; a certain percentage of similarity is compared with a threshold value;

determining the percentage of similarity of each spectral component of the signals to each other by determining the total share of the area of two spectral distributions, where one distribution belongs to the signal from one pyrosensor, and the other is obtained from the signal of another pyrosensor; a certain percentage of similarity is compared with threshold values, in case the percentage of similarity is at least 25%, an object motion detection signal is generated.

According to one of the preferred embodiments of the method, the measurement period is in the range of 50-150 ms from each pyrosensor.

According to another preferred embodiment of the method, the signal rate of change function is constructed with a number of measurements of 64.

It should be understood that the above general description and the following detailed description are illustrative and explanatory only and do not limit the claimed invention.

Brief description of drawings

The provided drawings, which are included in and constitute a part of this specification, illustrate embodiments of the invention and, together with the description, serve to explain the principles of the invention. The drawings show:

fig. 1 is a simplified block diagram showing the operation of the method when using two pyrosensors;

fig. 2 is a simplified block diagram showing the operation of the method when using three pyrosensors;

fig. 3 is a diagram showing the input signal f(x,t) (in the case of using two pyrosensors);

fig. 4 is a diagram showing the derivative of the signal corresponding to the growth rate fd'(x,t) (in the case of using two pyrosensors);

fig. 5 is a diagram showing the input signal f(y,t) (in the case of using two pyrosensors);

fig. 6 is a diagram showing the derivative of the signal corresponding to the growth rate fd'(y,t) (in the case of using two pyrosensors);

fig. 7 - schematically shows the distribution of pyrosensor signals fd(z,t) into 8 spectral components in the form of a sinusoidal function;

- 2 044353 fig. 8 - schematically shows the distribution of pyrosensor signals fd(y,t) into 8 spectral components in the form of a sinusoidal function;

fig. 9 - schematically shows the distribution of pyrosensor signals fd(x,t) into 8 spectral components in the form of a sinusoidal function;

fig. 10 - schematically shows a comparison of the spectral components from three pyrosensors fd(x,t), fd(y,t), fd(z,t).

Detailed description

An illustrative embodiment of the invention is described in detail below using the accompanying drawings. The implementations set forth in the following description of the embodiment do not cover all implementations of the invention, but only serve to further explain its essence.

To implement the claimed method, a system for detecting the movement of an object can be used, containing a sensor and a receiver, between which data is transmitted. The sensor may contain two or more pyrosensors that continuously generate signals. These pyrosensors are placed one above the other, thus forming sensitivity zones.

The sensor contains a signal analysis block and a data block. The signal analysis block contains three consecutive subblocks: a delta analysis subblock, a correlation analysis subblock and a spectral signal analysis subblock.

The data block contains a block of reference values and a block of threshold values. These blocks are equipped with a data array with the values of most types of objects that can be potential obstacles to the operation of security systems. The indicated values were obtained as a result of long-term experimental studies. These blocks allow you to select the type of motion detection object. In most cases, this is, of course, a person.

The method for detecting the movement of an object is implemented as follows.

For example, two pyrosensors (x, y) were used (see Fig. 1). These pyrosensors within the sensitivity zones detect the presence of any movement, generating signals f(x,t), f(y,t), respectively. In this example, the goal is to establish that a given movement belongs to a person, but the method provides for the possibility of establishing any predetermined movement.

Next, using the delta analysis subblock, the signal intensity is measured by determining the rate of change of the signal during the measurement period of 50-150 ms (6-20 Hz) from each pyrosensor, namely, the speed function D(x,t), D(y,t) is built signal changes when the number of measurements is 64 for the specified measurement period. Thus, the signals f(x,t), f(y,t) are sampled with a period of 50-150 ms (6-20 Hz) ^ fd(x,t), fd(y,t) and the signal derivative corresponding to growth rates fd'(x,t), fd'(y,t) - respectively, which is a delta function D(x,t), D(y,t).

Next, the measurement data of the function derivative curve is compared with the threshold values from the threshold values block. When the threshold values are exceeded by 3-5%, it is established that the signal fd(x,t), fd(y,t) is probably from a person, therefore, this signal goes into processing by the correlation analysis subblock.

Using the correlation analysis subblock, the shapes of the generated curves from pyrosensors are compared using the Pearson correlation coefficient R(x), R(y), thus determining the similarity of the signals.

The Pearson correlation coefficient is calculated using the formula

where are the sample means,

Ж™ sample variances, € [-1.1]·

In the case where the degree of similarity of the waveforms is below the reference value taken from the reference value block, it is determined that the signals fd(x,t), fd(y,t) and R(x), R(y) are likely to belong to a person, and Then they go to processing by the signal spectral analysis subblock.

It is this last stage of signal verification that provides accurate and unmistakable confirmation that the signal belongs/does not belong to a person.

Signals fd(x,t), fd(y,t) are distributed into 8 spectral components in the form of sinusoidal functions and compared with each other and with reference values from the reference values block. There may be more spectral components, depending on the type of object and environmental conditions.

Comparison of spectral components with each other and with reference values can be carried out in different ways, but the following are mandatory.

The percentage of similarity of each spectral component of each signal with the reference ones from the block of reference values is established by determining the total area fraction of two spectral distributions, where one distribution is a predetermined reference, and the other is obtained from the signal.

The determined similarity percentage is compared with a threshold value from the threshold value block.

Calculation of the threshold value S is the function S' = W(a,b,c,d), where W is a function depending on the corresponding arguments, a = D(x,t), b = R(x), c = D(y ,t), d = R(y).

Thus, the threshold value depends on the growth rate of the signals and on the correlation coefficient.

The percentage of similarity of each spectral component of the signals to each other is also determined by determining the total share of the area of two spectral distributions, where one distribution belongs to the signal from one pyrosensor, and the other is obtained from the signal of another pyrosensor.

The determined similarity percentage is compared with the threshold values from the threshold value block.

Calculation of the threshold value S is the function S' = W(a,b,c,d), where W is a function depending on the corresponding arguments, a = D(x,t), b = R(x), c = D(y ,t), d = R(y).

Thus, the threshold value depends on the growth rate of the signals and on the correlation coefficient.

If the percentage of similarity from these stages is 25%, a motion detection signal belonging to a person is generated, after which the signal is transmitted to the receiver.

For example, three pyrosensors (x, y, z) were used (see Fig. 2). These pyrosensors within the sensitivity zones detect the presence of any movement, generating signals f(x,t), f(y,t), f(z,t), respectively. In this example, the goal is to establish that a given movement belongs to a person, but the method provides for the possibility of establishing any predetermined movement.

Next, using the delta analysis subblock, the signal intensity is measured by determining the rate of change of the signal during the measurement period of 50-150 ms (6-20 Hz) from each pyrosensor, namely, the function D(x,t), D(y,t) is constructed, D(z,t) rate of signal change with the number of measurements 64 for the specified measurement period. Thus, the signals f(x,t), f(y,t), f(z,t) are sampled with a period of 50-150 ms (6-20 Hz) ^ fd(x,t), fd(y,t ), fd(z,t) and the derivative of the signal is calculated corresponding to the growth rate fd'(x,t), fd'(y,t), fd'(z,t) - respectively, which is the delta function D(x, t), D(y,t), D(z,t).

Next, the measurement data of the function derivative curve is compared with the threshold values from the threshold values block. When the threshold values are exceeded by 3-5%, it is established that the signal fd(x,t), fd(y,t), fd(z,t) is probably from a person, therefore, this signal goes into the processing of the correlation analysis subblock.

Using the correlation analysis subunit, the shapes of the generated curves from pyrosensors are compared using the Pearson correlation coefficient R(x), R(y), R(z), thus determining the similarity of the signals.

The Pearson correlation coefficient is calculated using the formula

ΣΤ1 - g) (yi - y) covfa y)

Where

sample means w” ¹ i Y ^t > si, sample variances, € [ ^— 1]

In the case where the degree of similarity of the signal shapes is lower than the reference value taken from the block of reference values, it is established that the signals fd(x,t), fd(y,t), fd(z,t) and R(x), R( y), R(z) probably belong to a person, and are then processed by the spectral signal analysis subunit.

It is this last stage of signal verification that provides accurate and unmistakable confirmation that the signal belongs/does not belong to a person.

Signals fd(x,t), fd(y,t), fd(z,t) are distributed into 8 spectral components in the form of sinusoidal

Claims (3)

functions and compare them with each other and with reference values from the block of reference values. There may be more spectral components, depending on the type of object and environmental conditions. Comparison of spectral components with each other and with reference values can be carried out in different ways, but the following are mandatory. The percentage of similarity of each spectral component of each signal with the reference ones from the block of reference values is established by determining the total area fraction of two spectral distributions, where one distribution is a predetermined reference, and the other is obtained from the signal. The determined similarity percentage is compared with a threshold value from the threshold value block. Calculation of the threshold value S is the function S' = W(a,b,c,d,e,g), where W is a function depending on the corresponding arguments, a = D(x,t), b = R(x), c = D(y,t), d = R(y), e = D(z,t), g = R(z). Thus, the threshold value depends on the growth rate of the signals and on the correlation coefficient. The percentage of similarity of each spectral component of the signals to each other is also determined by determining the total share of the area of two spectral distributions, where one distribution belongs to the signal from one pyrosensor, and the other is obtained from the signal of another pyrosensor. The determined similarity percentage is compared with the threshold values from the threshold value block. Calculation of the threshold value S is the function S' = W(a,b,c,d,e,g), where W is a function depending on the corresponding arguments, a = D(x,t), b = R(x), c = D(y,t), d = R(y), e = D(z,t), g = R(z). Thus, the threshold value depends on the growth rate of the signals and on the correlation coefficient. If the percentage of similarity from these stages is 25%, a motion detection signal belonging to a person is generated, after which the signal is transmitted to the receiver. CLAIM 1. A method for detecting the movement of an object, which includes detecting movement within the sensitivity zones of at least two pyrosensors and generating a signal from these pyrosensors, then measuring the signal intensity by determining the rate of change of the signal during a measurement period of maximum 150 ms from each pyrosensor, in particular constructing a function the rate of signal change with a number of measurements of at least 32 for the specified measurement period, compare the measurement data of the derivative curve of the function with threshold values, if the threshold values are exceeded by at least 3%, compare the shapes of the curves from the pyrosensors using the Pearson correlation coefficient, in the case when the degree of similarity of the signal shapes is below the reference value, the signals are distributed into at least 8 spectral components in the form of sinusoidal functions and compared with each other and with the reference values by establishing the percentage of similarity of each spectral component of each signal with the reference values by determining the total area fraction of the two spectral distributions, where one distribution is a predetermined reference, and the other is obtained from the signal; a certain percentage of similarity is compared with a threshold value; determining the percentage of similarity of each spectral component of the signals to each other by determining the total share of the area of two spectral distributions, where one distribution belongs to the signal from one pyrosensor, and the other is obtained from the signal of another pyrosensor; a certain percentage of similarity is compared with threshold values, in case the percentage of similarity is at least 25%, an object motion detection signal is generated. 2. The method for detecting the movement of an object according to claim 1, characterized in that the measurement period is within 50-150 ms from each pyrosensor. 3. The method for detecting the movement of an object according to claim 1, characterized in that the function of the rate of change of the signal is constructed with a number of measurements of 64. -