JP3701656B2 - Moving pattern recognition device - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a movement pattern recognition device that detects, for example, the moving direction of a human body at an entrance and the like and the number of people passing by using an infrared array sensor.
[0002]
[Prior art]
In recent years, in security, air conditioning control, and lighting control, there is an increasing demand for measurement of the presence and temperature distribution of a person in the room in order to detect the presence or activity of a person in the room. Furthermore, it is required to grasp and manage the number of visitors to and exits from rooms and buildings by performing occupancy management in companies and the like.
[0003]
Conventional occupant management includes a method of registering at the reception of a building or room, a method of counting by a manual counter at the entrance of the room, a method of counting each person by entering and leaving the room mechanically, and the like.
[0004]
Further, a method of detecting the number of people and the moving direction by installing pressure sensors on mats such as entrances and stairs as disclosed in JP-A-3-196286, 5-324955, etc., and JP-A-3-186998, 4-95794, 5- There is a method of counting the moving direction and the number of people using a plurality of infrared sensors such as 81503. In other words, the human body can be detected by detecting the weight of the human body, the direction of movement and the number of passing people can be detected, the human body can be detected by detecting the infrared rays emitted from the human body, and the signal can be air-conditioned. Devices that detect infrared radiation sources using infrared sensors have come to be used to control environmental control equipment such as machines and lighting equipment, security systems, etc., and to measure the number of people entering and leaving. It was.
[0005]
In addition, in order to solve the above problems, the moving direction of the passing human body and the number of passing people are detected by using one infrared sensor provided with a plurality of detection units, and the number of people in the room is managed. A passing-through number detection sensor and a system thereof have been filed by the present applicant.
[0006]
[Problems to be solved by the invention]
However, when the number of people is counted using a person as described above, there is a disadvantage that the cost is high, and it is very troublesome and takes time and effort.
[0007]
In addition, the mechanical detection method requires large-scale equipment, is very expensive, and has a problem that there is no space for installation.
[0008]
In addition, in the method of detecting a human body using a plurality of infrared sensors, since a plurality of sensors are used, the cost is high, and since it is not known whether all the entrances / exits are detected, a lot of false detections are likely to occur. There was a problem that it was impossible to measure the number of people. In particular, in the detection of a sensor in the case where a person moves continuously or crosses slightly apart, there are many false detections and misjudgments, and there is a problem that accurate human body detection cannot be performed. . Furthermore, in each of these methods, a signal processing method, an algorithm, and the like suitable for each method should be necessary, but none of them are disclosed.
[0009]
Therefore, the present invention has been tried in view of the above-described problems. For example, it is possible to recognize a movement pattern that can detect the direction of a moving body or calculate the number of movements for each direction. An object of the present invention is to provide a moving pattern recognition device that performs the above-described operation.
[0010]
[Means for Solving the Problems]
  The present invention uses a condensing lens that condenses infrared rays at a predetermined angle and an infrared array sensor that detects infrared rays collected by the condensing lenses using a plurality of detection units, and detects the condensing infrared rays. Movement signal detection means for detecting a movement signal indicating the position of the heating element that changes with time when the heating element moves in the region, and normalized data of the movement signal detected by the movement signal detection means The movement signal normalizing means to be generated, and data relating to various movement patterns when the heating element moves in the detection area, which is created in advance in consideration of the normalized data by the movement signal normalizing means, are stored. The movement pattern storage means, and the movement pattern storage means select data related to the movement pattern corresponding to the normalized data generated by the movement signal normalization means. And movement pattern selection means for, based on the data selected by the movement pattern selection means, recognizing the movement pattern recognizing means of the movement pattern of the heating element,Reference temperature storage means for storing reference data relating to a predetermined reference temperature,
The movement signal normalization means includes
Using the reference data stored in the reference temperature storage means, obtain the ratio or difference of each of the signals obtained from the plurality of detection units of the infrared array sensor,
By performing binarization processing by comparing the magnitude of the ratio or difference between the detection unit and the detection unit adjacent thereto, normalization data of the movement signal is generated,It is a movement pattern recognition device.
[0011]
The movement signal detecting means of the present invention has a plurality of sets of the condenser lens and the infrared array sensor, and each of the plurality of sets covers a predetermined region where the heating element moves. Also good.
[0012]
The present invention further includes dust data storage means for storing a plurality of dust data, which is data that cannot be generated by the movement signal normalization means, and the movement signal normalization means is provided by the movement signal detection means. When the normalized data of the detected movement signal is generated and the dust data stored in the dust data storage means is included in the normalized data, the dust data is removed from the normalized data. Also good.
[0013]
  In addition,The predetermined reference temperature is a temperature detected by the movement signal detecting means when the heating element is not present in the detection area, and the reference data is obtained by sampling for a predetermined time. The average value and standard deviation of the reference temperature calculated using the reference temperature may be used.
[0014]
Further, according to the present invention, the data relating to the movement pattern stored in the movement pattern storage means has sub-information relating to the movement direction of the movement pattern and / or the number of passages in the direction, and the movement pattern recognition means The movement direction related to the movement pattern of the heating element and / or the number of passages in the direction may be recognized based on the data related to the movement pattern selected by the movement pattern selection means.
[0015]
In the present invention, the movement signal detecting means uses a sensor for detecting the presence of the moving body, and indicates the position of the moving body that changes with time in a detection area where the moving body that can be detected by the sensor moves. A movement signal is detected. The movement pattern storage means stores data relating to various movement patterns when the moving body moves in the detection area. Data relating to the various movement patterns is created in advance. The movement pattern selection means selects data relating to the movement pattern corresponding to the movement signal detected by the movement signal detection means from the movement pattern storage means.
[0016]
In the present invention, the movement signal detecting means uses a condensing lens for condensing infrared rays at a predetermined angle and an infrared array sensor for detecting infrared rays condensed by the condensing lenses by a plurality of detection units. A movement signal indicating a position of the heating element that changes with time when the heating element moves in a detection region that collects infrared rays is detected. The movement signal normalization means generates normalization data of the movement signal detected by the movement signal detection means. The movement pattern storage means stores data relating to various movement patterns when the heating element moves in the detection area, which is created in advance in consideration of the normalized data by the movement signal normalization means. . The movement pattern selection means selects data related to the movement pattern corresponding to the normalized data generated by the movement signal normalization means from the movement pattern storage means. The movement pattern recognition means recognizes the movement pattern of the heating element based on the data selected by the movement pattern selection means.
[0021]
DETAILED DESCRIPTION OF THE INVENTION
Embodiments of the present invention will be described below with reference to the drawings.
[0022]
FIG. 1 is a perspective view showing an embodiment of a movement signal detection means used in the movement pattern recognition apparatus of the present invention.
[0023]
In general, an infrared sensor or the like is used as a sensor used in an apparatus that detects an infrared generation source. Two types of infrared sensors are known: a quantum sensor that captures infrared rays as photons, and a thermal sensor that absorbs infrared rays as electromagnetic waves and uses changes in the physical properties of the element as a result of thermal action that increases the temperature of the element. ing. Since the former usually requires cooling with liquid nitrogen or the like, the latter thermal type sensor is generally used. Among the thermal sensors, the pyroelectric infrared sensor is suitable for infrared source detection because it has higher sensitivity than others.
[0024]
However, since the pyroelectric infrared sensor basically detects a change in the infrared ray, when trying to detect a stationary infrared ray generation source, the infrared ray is intermittently applied to the sensor light receiving unit by any method. It is necessary to devise so that it may enter. Normally, infrared rays are intermittently incident (chopped) on the sensor light receiving unit by rotating or vibrating a chopper such as a disk with a slit or a flat plate.
[0025]
Therefore, the infrared array sensor of the movement signal detecting means of this embodiment is composed of a plurality of line-shaped pyroelectric elements 11. An infrared light shielding plate 12 is attached so as to cover the front surface of the pyroelectric element 11. An infrared transmission lens 13 which is a silicon infrared transmission lens is attached so as to collect infrared rays 17 on the pyroelectric element 11. If a wide-angle lens is used for the infrared transmission lens 13, it is possible to detect heating elements that move in a batch without rotating the infrared array sensor. The chopper 14 is provided to intermittently block the infrared rays 17 incident on the infrared transmission lens 13. The chopper 14 is mechanically connected to the brushless motor 16 via the chopper shaft 15 in a rotatable state. If the chopping frequency is set to 10 Hz or more, human body detection can be performed at higher speed, more accurately, and with higher accuracy.
[0026]
In addition, since this infrared array sensor has a plurality of line-shaped pyroelectric elements 11, the output value for each time of each pyroelectric element 11 is taken into consideration based on the plurality of pyroelectric elements 11. In this case, it is possible to detect in which direction the moving heating element has moved on the plurality of pyroelectric elements 11.
[0027]
FIG. 2 is a configuration diagram of another embodiment according to the movement pattern recognition apparatus of the present invention. A sensor head 21 which is a movement signal detecting means includes the sensor shown in FIG. Since the sensor head 21 has the plurality of line-shaped pyroelectric elements 11 shown in FIG. 1, the output values of the pyroelectric elements 11 for each time with reference to the plurality of pyroelectric elements 11. Can be detected in which direction the heating element has moved on the plurality of pyroelectric elements 11. The output of the sensor head 21 is input to the signal processing circuit 22 through a cable of several meters. The signal processing circuit 22 includes an amplifier circuit (Amp. In FIG. 2) and an A / D converter (A / D in FIG. 2). The output of the sensor head 21 input to the signal processing circuit 22 is amplified by the amplifier circuit and then input to the A / D converter and converted into a digital signal (AD value). The digital signal is arithmetically processed by a CPU, a memory, and the like included in the signal processing circuit 22, and the movement pattern of the heating element that moves based on the infrared line sensor of the sensor head 21 is recognized. The recognition result is output from the RS-232C and transmitted to the computer or the display device 23.
[0028]
The arithmetic processing program (the flowchart will be described later) is stored in a semiconductor memory such as a ROM or a memory card, or a disk memory such as a hard disk, a magneto-optical disk, or a flexible disk. In addition, data created in advance that is referred to in the arithmetic processing is also stored in a semiconductor memory such as a ROM or a memory card, or a disk memory such as a hard disk, a magneto-optical disk, or a flexible disk. The memory included in the signal processing circuit 22 is a main memory for loading the program, and is the memory described above.
[0029]
By the way, the CPU of the signal processing circuit 22 that has started the program corresponds to the movement signal normalization means, movement pattern selection means, and movement pattern recognition means of the movement pattern recognition apparatus of the present invention. The memory of the signal processing circuit 22 corresponds to the movement pattern storage means, dust data storage means, and reference temperature storage means of the present invention.
[0030]
FIG. 4 is a diagram illustrating a state in which the sensor head 21 according to the present embodiment is installed above the entrance of the room. The sensor head 21 is installed at the center upper part of the entrance / exit 44 of a room with a width of about 100 cm and a height of about 220 cm so as to collect infrared rays from the center lower part of the entrance / exit 44 at a predetermined angle. The infrared rays collected by the sensor head 21 are input to the signal processing circuit 22, and information such as entrance / exit obtained by processing there is displayed on the computer or the display device 43. If the incident angle of the sensor at this time is 6 °, the detection area 46 becomes as shown in the figure, and even if the human body 45 passes through the end, it can be sufficiently detected.
[0031]
Next, operation | movement of this apparatus in this installation state is demonstrated.
[0032]
FIG. 5 is a flowchart regarding the processing of the CPU of FIG. 2 in which the program is running. That is, the floor surface temperature, which is always detected, is sampled for a predetermined time, and the average value (Xvar) and standard deviation (S) of the floor surface temperature are calculated. These average values and standard deviations are stored in the memory of the signal processing circuit 22. As the storage operation, if the human body 45 does not exist in the detection area 46 during the predetermined sampling time, the calculated new average value and standard deviation are updated.
[0033]
These average values and standard deviations are used according to (Equation 1), and for each chopping, a plurality of pyroelectric elements are obtained from AD values obtained by inputting from each of the plurality of pyroelectric elements of the infrared array sensor. The rate of change (HENKA) of each of the signals obtained from is calculated.
[0034]
[Expression 1]
HENKA (i, j) = (AD value−Xvar) / S
When this change rate is three times or more of the standard deviation, the change rates between adjacent pyroelectric elements in the infrared array sensor are compared, and binarization processing is performed by setting the larger one to 1.
[0035]
FIG. 6 is a diagram showing how the normalized data of the movement signals detected by the three pyroelectric elements 11 is generated as a specific example. The detection area of the infrared array sensor installed 2.5 m above consists of detection areas 1, 2, and 3. A state A shown in (a) is a moving state when a human body enters the detection area 3. The states B1 and B2 shown in (b) are the movement states when proceeding toward the detection area 2. States C1 and C2 shown in (c) are movement states when the detection area 1 is reached. The binarized data at this time is as shown in the graph on the right side of each figure. That is, if the change rate of 3 times or more of the standard deviation σ (S) is set to 1 which is larger than the adjacent pyroelectric element, A: (001), B1: (011), B2: (010), C1: (110), C2: (100).
FIG. 7 is a diagram illustrating an example of another operation obtained in the same manner as FIG. For example, when a person passes, as shown in (a), it changes as (000) (001) (011) (110) (000). Moreover, when a person is making a U-turn in the middle, it is considered that it changes as shown in (b) or (c).
[0036]
Based on the binarized data obtained in this way, data relating to various movement patterns when the human body moves within the detection area is created in advance and stored in the memory of the signal processing circuit 22. .
[0037]
The actual movement pattern detected and processed by the infrared array sensor is compared with various movement patterns stored in this memory, and data relating to the corresponding movement pattern is selected.
[0038]
The movement pattern of the human body is recognized based on the data related to the selected movement pattern. That is, if the information regarding the direction of movement is included in the data regarding the movement pattern, it is recognized in which direction the human body has moved. In the case of the present embodiment, the movement pattern recognition device recognizes whether a person has entered through the entrance 44 or has exited.
[0039]
By this method, it is possible to detect the human body with high reliability and to determine the moving direction, and the moving direction detection rate which has been about 90% in the past can be increased to 99% or more. .
[0040]
FIG. 8 is a flowchart showing the operation of another embodiment according to the movement pattern recognition apparatus of the present invention. The configuration and operation of the movement pattern recognition apparatus of the present embodiment are the same as those of the embodiment described with reference to FIGS. 1 to 7, and the difference is the type of data information related to various movement patterns. This is a process of recognizing information after the selection is made and the result obtained. FIG. 9 is a diagram illustrating specific binarized data regarding a movement pattern when a person enters and exits an entrance. A case where two human bodies cross each other or a case where the human body passes continuously is shown. In the present embodiment, such a movement pattern is considered, and information indicating in which direction the person has passed through the detection area is used together with the data related to the movement pattern, together with the signal processing circuit 22 of FIG. In advance. Thereby, the movement pattern recognition apparatus of a present Example can count how many people entered or went out from the entrance / exit. It is also possible to calculate the number of visitors by time of day.
[0041]
With this method, it is possible to detect a human body with high reliability and with high accuracy, determine the moving direction, and measure the number of people passing by. I was able to.
[0042]
FIG. 10 is a flowchart showing the operation of another embodiment according to the movement pattern recognition apparatus of the present invention. The configuration and operation of the movement pattern recognition apparatus of the present embodiment are the same as those of the embodiment described with reference to FIGS. 1 to 9, and the handling of dust patterns described below is different. First, the memory of the signal processing circuit 22 in FIG. 2 stores in advance a plurality of dust data, which is data that cannot be obtained by generating the movement signal normalization data in FIG.
[0043]
Therefore, how this dust pattern is used will be described. The floor surface temperature, which is the background, is sampled for a predetermined time, and the average value and standard deviation of the floor surface temperature are calculated. These average values and standard deviations are stored in the memory of the signal processing circuit 22. As a storing operation, if a human body does not exist in the detection area during a predetermined sampling time, the calculated new average value and standard deviation are updated.
[0044]
These average values and standard deviations are used according to (Equation 1), and for each chopping, a plurality of pyroelectric elements are obtained from AD values obtained by inputting from each of the plurality of pyroelectric elements of the infrared array sensor. The rate of change (HENKA) of each of the signals obtained from is calculated.
[0045]
When this change rate is three times or more of the standard deviation, the change rates between adjacent pyroelectric elements in the infrared array sensor are compared, and binarization processing is performed by setting the larger one to 1.
[0046]
If there is data corresponding to the dust pattern in the binarized data, the data is deleted as dust. The dust pattern is data when only 1 appears suddenly, for example, (000) (010) (000). In such a case, the dust pattern is automatically canceled by deleting dust.
[0047]
FIG. 11 is a block diagram relating to another embodiment of the movement pattern recognition apparatus of the present invention. Each of the sensor heads (111a, 111b,...) As a plurality of movement signal detection means includes the sensor of FIG. Each output of the sensor head is input to a respective signal processing circuit (112a, 112b,...). Each of the signal processing circuits includes an amplifier circuit (Amp. In FIG. 11), an A / D converter (A / D in FIG. 11), and a motor driver. The output of the sensor head input to each of the signal processing circuits is amplified by the amplifier circuit, and then input to the A / D converter to be converted into a digital signal (AD value). The digital signal is input to the control device 113 having a CPU, a memory, etc., and is processed by the CPU, the memory, etc. The pattern is recognized. This result is output from the RS-232C or PIO and sent to the computer 114, or from the PIO and sent to the display device 115.
[0048]
The arithmetic processing program (the flowchart will be described later) is stored in a semiconductor memory such as a ROM or a memory card, or a disk memory such as a hard disk, a magneto-optical disk, or a flexible disk. Data referred to in the calculation processing is also stored in a semiconductor memory such as a ROM or a memory card, or a disk memory such as a hard disk, a magneto-optical disk or a flexible disk. The memory included in the control device 113 is a main memory for loading the program, and is the memory described above.
[0049]
By the way, the CPU of the control device 113 that has started the program corresponds to the movement signal normalization means, movement pattern selection means, and movement pattern recognition means of the movement pattern recognition apparatus of the present invention. The memory of the control device 113 corresponds to the movement pattern storage means, dust data storage means, and reference temperature storage means of the present invention.
[0050]
FIG. 12 is a diagram showing a state in which the sensor head of the present embodiment is installed above the entrance of the room. Four sensor heads (111a, 111b, 111c and 111d) are always placed on the floor surface at a certain interval perpendicular to the passing direction of the human body 126 on the upper part of the wide entrance / exit 125 having a width of about 200 cm and a height of about 250 cm. And the arrangement direction of the plurality of pyroelectric elements constituting each infrared array sensor of the sensor head is set so as to be parallel to the passing direction of the human body. If the incident angle of the sensor head at this time is 6 °, the detection region 127a or 127b is as shown in the figure, so that even if the human body 126 passes through the end, it can be sufficiently detected.
[0051]
Next, the process of a present Example is demonstrated. First, the floor surface temperature, which is always detected, is sampled for a predetermined time, and the average value and standard deviation of the floor surface temperature are calculated. These average values and standard deviations are stored in the memory of the control device 113 in FIG. As a storing operation, if a human body 126 does not exist in each detection area during a predetermined time of sampling, the new average value and standard deviation are updated.
[0052]
These average values and standard deviations are used by (Equation 1), and AD values obtained by inputting from each of the plurality of pyroelectric elements constituting the infrared array sensor of each sensor head for each chopping. From this, the rate of change (HENKA) of each of the signals obtained from the plurality of pyroelectric elements is calculated.
[0053]
When this change rate is three times or more of the standard deviation, the change rates between adjacent pyroelectric elements in the infrared array sensor are compared, and binarization processing is performed by setting the larger one to 1.
[0054]
Based on the binarized data obtained as described with reference to FIGS. 6 and 7, data relating to various movement patterns when the human body moves within the detection region is created in advance, and FIG. Are stored in the memory of the control device 113.
[0055]
The actual movement pattern detected and processed by the infrared array sensor is compared with various movement patterns stored in the memory, and data relating to the corresponding movement pattern is selected.
[0056]
The movement pattern of the human body is recognized based on the data related to the selected movement pattern. That is, if information on the direction of movement and the number of passes in that direction is included in the data on the movement pattern, it is recognized how many times, that is, how many people have moved. In the case of the present embodiment, the movement pattern recognition device recognizes how many people have entered from the entrance 125 or how many have exited.
[0057]
As described above, according to the present embodiment, it is possible to easily detect a human body with a non-constant temperature distribution entering a room or building at a low cost, and measure the number of passing people with high accuracy and high reliability. The number of people in the room can be displayed.
[0058]
FIG. 13 is a flowchart showing the operation of another embodiment according to the movement pattern recognition apparatus of the present invention. The configuration of the movement pattern recognition apparatus of this embodiment is the same as that of the embodiment described with reference to FIG. 11, but the processing is performed in consideration of a plurality of movement signals from a plurality of installed sensor heads. Is different.
[0059]
Therefore, processing in the present embodiment will be described.
[0060]
First, the floor surface temperature, which is always detected, is sampled for a predetermined time, and the average value (Xvar) and standard deviation (S) of the floor surface temperature are calculated. These average values and standard deviations are stored in the memory of the control device 113 in FIG. As a storing operation, if there is no person in each detection area during a predetermined time of sampling, the calculated new average value and standard deviation are updated.
[0061]
These average values and standard deviations are used by (Equation 1), and AD values obtained by inputting from each of the plurality of pyroelectric elements constituting the infrared array sensor of each sensor head for each chopping. From this, the rate of change (HENKA) of each of the signals obtained from the plurality of pyroelectric elements is calculated.
[0062]
When this change rate is three times or more of the standard deviation, the change rates between adjacent pyroelectric elements in the infrared array sensor are compared, and binarization processing is performed by setting the larger one to 1.
[0063]
If there is data corresponding to the dust pattern in the binarized data, the data is deleted as dust. Here, the memory of the control device 113 stores in advance a plurality of dust data, which is data that cannot be obtained by the binarization process.
[0064]
Based on the data binarized in this way, data relating to various movement patterns when the human body moves within the detection area is created in advance and stored in the memory of the control device 113. The actual movement pattern detected and processed by the infrared array sensor of each sensor head is compared with various movement patterns stored in this memory, and data relating to the corresponding movement pattern is selected. .
[0065]
The number of persons passing through the human body is detected by comprehensively determining data relating to the movement pattern selected by each sensor head. For example, referring to the example of FIG. 12, when output appears on the sensor heads 111 a and 111 b at the same time, or when output appears on the sensor head 111 a or 111 b, it is determined that only one person has passed. Can do. Further, when outputs appear simultaneously on the sensor heads 111a and 111d, it can be determined that two persons have passed.
[0066]
As described above, according to the present embodiment, it is possible to easily detect a person entering a room or a building even at a wide entrance, and to measure the number of passing people with high accuracy and high reliability. The number of people in the room can be displayed.
[0067]
FIG. 14 is a block diagram relating to another embodiment of the movement pattern recognition apparatus of the present invention. The configuration of the movement pattern recognition apparatus of this embodiment is the same as that of the embodiment described with reference to FIG. As shown in FIG. 14, three sensor heads (111 a, 111 b, and 111 c) are placed on the upper part of the wide entrance / exit 142 having a width of about 160 cm and a height of about 250 cm with a certain interval perpendicular to the passing direction of the human body. The sensor heads (111a and 111c) at both ends are arranged so that the arrangement direction of the infrared array sensors is parallel to the passing direction of the human body, and the center sensor head 111b is vertical. Install to be. By installing in this way, the detection area (143a, 143b and 143c) of each infrared array sensor becomes as shown in the figure. If the incident angles of the sensor heads (111a and 111c) at both ends are 6 °, the detection area (143a and 143c) is as shown in the figure. Therefore, even if the human body passes through the end, it should be detected sufficiently. Can do.
[0068]
Next, the process of a present Example is demonstrated.
[0069]
First, the floor surface temperature, which is always detected, is sampled for a predetermined time, and the average value and standard deviation of the floor surface temperature are calculated. These average values and standard deviations are stored in the memory of the control device 113 in FIG. As a storing operation, if a human body does not exist in each detection area during a predetermined sampling time, the calculated new average value and standard deviation are updated.
[0070]
These average values and standard deviations are used by (Equation 1), and AD values obtained by inputting from each of the plurality of pyroelectric elements constituting the infrared array sensor of each sensor head for each chopping. From this, the rate of change (HENKA) of each of the signals obtained from the plurality of pyroelectric elements is calculated.
[0071]
When this change rate is three times or more of the standard deviation, the change rates between adjacent pyroelectric elements in the infrared array sensor are compared, and binarization processing is performed by setting the larger one to 1.
[0072]
Based on the binarized data obtained as described with reference to FIGS. 6 and 7, data relating to various movement patterns when the human body moves within the detection region is created in advance, and the control device 113 is stored in the memory.
[0073]
The actual movement patterns detected and processed by the respective infrared array sensors of the sensor heads (111a, 111b and 111c) are compared with various movement patterns stored in this memory, and related to the corresponding movement patterns. Data is selected.
[0074]
The number of people passing through the human body is detected by comprehensively judging from the data related to the movement pattern selected by each sensor head. That is, the number of passing people is determined by the detection area 143b of the sensor head 111b including the detection areas (143a and 143c) of the sensor heads (111a and 111c) at both ends. Then, the moving direction and the number of passing people are detected by the sensor heads (111a and 111c) at both ends. For example, it is assumed that the number of pyroelectric elements constituting each infrared array sensor of three sensor heads (111a, 111b, and 111c) is eight. When this sensor head 111b detects a human body with three pyroelectric elements or less, even if the sensor heads (111a and 111c) at both ends detect the human body at the same time, the number of people passing through is counted as one. Further, when the sensor head 111b detects a human body with 4 to 6 pyroelectric elements, it counts as two persons.
[0075]
This method is highly reliable and can detect a human body with high accuracy and determine the number of people passing by. The passing person detection rate, which was about 80% in the past, can be increased to 95% or more. .
[0076]
In the above embodiment, the sensor head has an infrared array sensor composed of a plurality of pyroelectric elements. However, the sensor head is not limited to this, and may be a sensor such as an ultrasonic wave or a radar. . In short, the movement signal detecting means of the present invention only needs to be capable of detecting the directionality of a moving body that moves within a predetermined area.
[0077]
In addition, the sensor head of the above embodiment includes the chopper 14, the chopper shaft 15, and the brushless motor 16. However, the movement signal detection unit of the present invention only needs to detect the moving human body, so the chopper 14, The chopper shaft 15 and the brushless motor 16 are not necessarily required.
[0078]
In the above embodiment, the configuration of the signal processing circuit 22 in FIG. 2 is a configuration including a motor driver. However, the configuration is not limited to this, and as shown in FIG. The relay box 22b may have a different configuration, and the relay box 22b may be incorporated in the sensor head 21.
[0079]
Moreover, in the said Example, although floor surface temperature was used for the reference | standard of calculation of a rate of change, it is not necessarily restricted to this, You may use wall surface temperature.
[0080]
Moreover, in the said Example, although the normalization data of the movement pattern was produced | generated using three pyroelectric elements in FIG.6, FIG.7 and FIG.9, it is not necessarily this, and there are n infrared array sensors. In this case, 2 to n pyroelectric elements may be used to generate movement pattern normalization data.
[0081]
【The invention's effect】
As is clear from the above description, the movement pattern recognition apparatus of the present invention can recognize the movement pattern of the moving body. Alternatively, the moving pattern of the moving heating element can be recognized.
[0082]
Further, according to the present invention, if various information is added to the data related to the movement pattern stored in the movement pattern storage means, an apparatus meeting various needs can be provided. For example, adding the direction of movement and the number of passes in that direction not only enables detection of people entering or leaving a room or building, but also enables accurate entry by measuring the number of people passing through. It has an advantage of providing a device capable of detecting the number of persons and the number of people in the room.
[0083]
Further, in the present invention, when the infrared array sensor of the movement signal detecting means is installed so as to collect infrared rays from the floor surface, it is possible to always feed back the temperature fluctuation of the floor surface. Even if fluctuates, the average value can follow the fluctuation. Since the movement signal detecting means detects the heating element that moves using the average value and the standard deviation, the movement signal detection means has an effect of being able to detect even if the heating element is a low-temperature object having a low temperature.
[0084]
In addition, according to the present invention, there is an effect that the moving body or the moving heating element can be accurately detected even when the moving body or the moving heating body passes continuously or crosses.
[0085]
Further, in the present invention, even when low temperature or high temperature objects are mixed, the dust pattern generated when the movement signal detection means does not clearly detect the moving body or the moving heating element is deleted. This has the effect of enabling accurate detection.
[0086]
In addition, according to the present invention, it is possible to obtain a moving pattern recognition device that can be downsized in terms of system, is low in cost, accurate and reliable, and can easily detect the number of people in the room immediately. it can.
[0087]
Further, in the present invention, no erroneous detection as in the case of using other conventional sensors is observed, and accurate and highly reliable human body detection and passing person detection are possible.
[0088]
In addition, in this embodiment, by using an algorithm that is very simple in terms of program, it is possible to easily realize accurate human body detection with few false detections even at the time of continuous passage or crossing at low cost. it can.
[0089]
In addition, according to the present invention, highly reliable and highly reliable human body detection and occupant number measurement can be performed easily, which can greatly contribute to a comfortable intelligent building system.
[Brief description of the drawings]
FIG. 1 is a perspective view showing an embodiment of a movement signal detection means used in a movement pattern recognition apparatus of the present invention.
FIG. 2 is a block diagram of another embodiment of the movement pattern recognition apparatus of the present invention.
3 is a configuration diagram in which a part of the configuration of the embodiment of FIG. 2 is changed.
FIG. 4 is a diagram showing a state in which the sensor head 21 of this embodiment is installed above the entrance of the room.
5 is a flowchart relating to the processing of the CPU of FIG. 2 in which a program is running.
FIG. 6 is a diagram showing a state of generation of normalized data of movement signals detected by three pyroelectric elements 11;
FIG. 7 is a diagram showing an example of another operation obtained in the same manner as in FIG.
FIG. 8 is a flowchart showing the operation of another embodiment of the movement pattern recognition apparatus of the present invention.
FIG. 9 is a diagram showing specific binarization data regarding a movement pattern when a person enters and exits an entrance / exit.
FIG. 10 is a flowchart showing the operation of another embodiment of the movement pattern recognition apparatus of the present invention.
FIG. 11 is a block diagram relating to another embodiment of the movement pattern recognition apparatus of the present invention.
FIG. 12 is a diagram showing a state in which the sensor head of this embodiment is installed above the entrance of the room.
FIG. 13 is a flowchart showing the operation of another embodiment of the movement pattern recognition apparatus of the present invention.
FIG. 14 is a block diagram relating to another embodiment of the movement pattern recognition apparatus of the present invention.
[Explanation of symbols]
11 Pyroelectric elements
12 Infrared shading plate
13 Infrared transmitting lens
14 Chopper
15 Chopper shaft
16 Brushless motor
17 Infrared
21 Sensor head
22, 22a Signal processing circuit
22b Relay BOX
23 Computer or display device
44 doorway
45 human body
46 Detection area

Claims (5)

Using a condensing lens that condenses infrared rays at a predetermined angle and an infrared array sensor that detects infrared rays collected by the condensing lenses using a plurality of detection units, heat is generated in the detection region that collects the infrared rays. A movement signal detecting means for detecting a movement signal indicating the position of the heating element that changes with time when the body moves;
Movement signal normalization means for generating normalized data of movement signals detected by the movement signal detection means;
Movement pattern storage means for storing data relating to various movement patterns when the heating element moves in the detection area, which is created in advance in consideration of the normalized data by the movement signal normalization means,
Movement pattern selection means for selecting data relating to the movement pattern corresponding to the normalized data generated by the movement signal normalization means from the movement pattern storage means;
Based on the data selected by the movement pattern selection means, movement pattern recognition means for recognizing the movement pattern of the heating element ,
Reference temperature storage means for storing reference data relating to a predetermined reference temperature,
The movement signal normalization means includes
Using the reference data stored in the reference temperature storage means, obtain the ratio or difference of each of the signals obtained from the plurality of detection units of the infrared array sensor,
A movement characterized by generating normalized data of the movement signal by performing binarization processing by comparing the magnitude of the ratio or difference between the detection section and a detection section adjacent to the detection section. Pattern recognition device.   The movement signal detecting means has a plurality of sets of the condenser lens and the infrared array sensor, and each of the plurality of sets covers a predetermined region where the heating element moves. The movement pattern recognition apparatus according to claim 1. Comprising dust data storage means for storing a plurality of dust data, which cannot be generated by the movement signal normalization means,
The movement signal normalization means includes
Normalization data of the movement signal detected by the movement signal detection means is generated;
3. The movement pattern recognition apparatus according to claim 2, wherein when the normalized data includes dust data stored in the dust data storage means, the dust data is removed from the normalized data. . The predetermined reference temperature is a temperature detected by the movement signal detection means when the heating element is not present in the detection region,
Wherein the reference data, the movement pattern recognizing apparatus according to claim 1, characterized in that the mean and standard deviation of said reference temperature calculated by using the reference temperature obtained by sampling a predetermined time. The data relating to the movement pattern stored in the movement pattern storage means has sub-information relating to the movement direction of the movement pattern and / or the number of passages in the direction,
The movement pattern recognition means recognizes the movement direction and / or the number of passages in the direction of the movement pattern of the heating element based on the data related to the movement pattern selected by the movement pattern selection means. The movement pattern recognition apparatus according to any one of claims 1 to 4 .

Images