JP3625775B2 - Anomaly detection device - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an abnormality detection device that detects, for example, an abnormality such as a person falling down in a room such as a bathroom or a toilet.
[0002]
[Prior art]
The number of accidents in which bathers die while taking a bath in the bathroom has been increasing year by year, and the number of deaths during bathing has exceeded the number of deaths due to traffic accidents. Yes.
[0003]
The direct cause of an accident while taking a bath is that the patient falls down due to heart failure or stroke while taking a bath, or a consciousness disorder occurs in the bathtub and drowns. However, although the cause of such a state is being investigated, the current situation is that it is difficult to identify the current lifestyle changes. Therefore, it is not easy to prevent heart failure or stroke during bathing.
[0004]
However, even if it is not possible to prevent the occurrence of symptoms, early detection has a much higher chance of retaining a life.
[0005]
Therefore, by detecting an image in the bathroom, to protect the privacy, a very rough image that barely shows the presence of the bather is captured, the center of gravity of the bather is calculated, and the movement of the center of gravity is detected. For example, Japanese Patent Laid-Open No. 11-101502 discloses a system that detects a bather's movement and, when this movement cannot be detected for a certain period of time, notifies the kitchen of an abnormality.
[0006]
However, in the bathroom, the lens of the image pickup device used when capturing an image may become cloudy due to the growth of water stains, mold, steam, etc., and the bather's movement may not be detected. Then, even if the bather is moving in the bathroom, an alarm is output.
[0007]
Also, in a bathroom with a mist function, turning on the mist function switch causes the bathroom to become cloudy, making it difficult for the bather's movement to be detected. For this reason, even if the bather is moving in the bathroom, there is a high possibility that an alarm is output.
[0008]
[Problems to be solved by the invention]
An object of the present invention is to provide an anomaly detection device that can accurately detect that an anomaly has occurred such as a person falling down in a room such as a bathroom or toilet.
[0009]
[Means for Solving the Problems]
A first abnormality detection apparatus according to the present invention includes an imaging unit that captures an image in a room, an extraction unit that extracts an image feature amount from the image captured by the imaging unit, and a predetermined time interval of the image feature amount extracted by the extraction unit. Detecting means for detecting a temporal change amount of the image, comparing the temporal change amount of the image feature amount detected by the detection means with the threshold value for motion determination, In some cases, it is determined that the monitoring target in the room is in a stationary state, and the motion determination means determines that the monitoring target in the room is in a moving state when it is greater than the motion determination threshold, and the past predetermined number of times by the motion determination means In the abnormality detection device provided with abnormality determination means for determining whether an abnormality has occurred in the monitoring target in the room based on the movement determination result of the movement of the lens of the imaging means based on the image captured by the imaging means Haze determination data generating means for generating a chloride determination data, and based on the haze determination data, threshold control hand to control the movement determination thresholdStepprepare forThe cloudiness determination data generation means extracts a high frequency component from the image captured by the imaging means as the cloudiness determination data.It is characterized by that.
[0010]
A second abnormality detection device according to the present invention is:An imaging means for capturing an image in the room, an extracting means for extracting an image feature quantity from the video imaged by the imaging means, a detecting means for detecting a temporal change amount of the image feature quantity extracted by the extracting means at predetermined time intervals, The temporal change amount of the image feature amount detected by the detection means is compared with the threshold value for motion determination, and the monitoring target in the room is stationary when the temporal change amount of the image feature amount is equal to or less than the threshold value for motion determination. Based on the motion determination unit that determines that the monitoring target in the room is in a moving state when the threshold is larger than the motion determination threshold, and the motion determination result of the past predetermined number of times by the motion determination unit, In an abnormality detection apparatus having an abnormality determination unit that determines whether an abnormality has occurred in a monitoring target, a haze determination unit that generates haze determination data for a lens of the imaging unit based on an image captured by the imaging unit Data generation means and threshold value control means for controlling the threshold value for motion determination based on the cloudiness degree determination data. The cloudiness degree determination data generation means extracts a high frequency component from the image captured by the imaging means. High-frequency component extraction means, and high-frequency component correction means for removing noise components from the high-frequency components extracted by the high-frequency component extraction means to generate haze degree determination dataIt is characterized by having.
[0011]
According to a third abnormality detection apparatus of the present invention, an imaging unit that captures an image in a room, an extraction unit that extracts an image feature amount from the image captured by the imaging unit, and a predetermined time interval of the image feature amount extracted by the extraction unit. Detecting means for detecting a temporal change amount of the image, comparing the temporal change amount of the image feature amount detected by the detection means with the threshold value for motion determination, and if the temporal change amount of the image feature amount is equal to or less than the threshold value for motion determination In some cases, it is determined that the monitoring target in the room is in a stationary state, and the motion determination means determines that the monitoring target in the room is in a moving state when it is greater than the motion determination threshold, and the past predetermined number of times by the motion determination means In the abnormality detection device provided with abnormality determination means for determining whether an abnormality has occurred in the monitoring target in the room based on the movement determination result of the movement of the lens of the imaging means based on the image captured by the imaging means The cloudiness degree determination data generation means for generating the cloudiness degree determination data, and the threshold value control means for controlling the movement determination threshold value based on the cloudiness degree determination data, the cloudiness degree determination data generation means includes: The image picked up by the image pickup means is divided into a plurality of feature amount calculation areas, and the extraction means for extracting the image feature amount from the video picked up by the image pickup means for each feature amount calculation area, the extraction means for each feature amount calculation area Based on the extracted image feature quantity, for each of a plurality of predetermined feature quantity calculation areas, means for calculating a value related to a difference in image feature quantity between the area and the surrounding area, and the plurality of feature quantities Means is provided for generating cloudiness determination data by accumulating values relating to differences in image feature amounts calculated for the respective calculation areas.
[0012]
According to a fourth abnormality detection apparatus of the present invention, an imaging unit that captures an image in a room, an extraction unit that extracts an image feature amount from the image captured by the imaging unit, and a predetermined time interval of the image feature amount extracted by the extraction unit. Detecting means for detecting a temporal change amount of the image, comparing the temporal change amount of the image feature amount detected by the detection means with the threshold value for motion determination, and if the temporal change amount of the image feature amount is equal to or less than the threshold value for motion determination In some cases, it is determined that the monitoring target in the room is in a stationary state, and the motion determination means determines that the monitoring target in the room is in a moving state when it is greater than the motion determination threshold, and the past predetermined number of times by the motion determination means In the abnormality detection device provided with abnormality determination means for determining whether an abnormality has occurred in the monitoring target in the room based on the movement determination result of the movement of the lens of the imaging means based on the image captured by the imaging means The cloudiness degree determination data generation means for generating the cloudiness degree determination data, and the threshold value control means for controlling the movement determination threshold value based on the cloudiness degree determination data, the cloudiness degree determination data generation means includes: The image picked up by the image pickup means is divided into a plurality of feature amount calculation areas, and the extraction means for extracting the image feature amount from the video picked up by the image pickup means for each feature amount calculation area, the extraction means for each feature amount calculation area Based on the extracted image feature quantity, for each of a plurality of predetermined feature quantity calculation areas, means for calculating a value related to a difference in image feature quantity between the area and the surrounding area, and the plurality of feature quantities Means is provided for generating cloudiness determination data by squaring and summing values relating to differences in image feature amounts calculated for the respective calculation areas.
[0013]
According to a fifth abnormality detection apparatus of the present invention, an imaging unit that captures an image in a room, an image captured by the imaging unit is divided into a plurality of feature amount calculation regions, and each feature amount calculation region is captured by the imaging unit. Extraction means for extracting image feature values from video, detection means for detecting temporal changes in image feature values extracted by the extraction means for each feature value calculation area for each predetermined time interval, and feature value calculation areas by the detection means The temporal change amount of the image feature amount detected for each time and the threshold value for motion determination are compared for each feature amount calculation region, and if the temporal change amount of the image feature amount is equal to or less than the threshold value for motion determination, A first motion determination unit that determines that there is no motion in the feature amount calculation area and determines that the feature amount calculation region has a motion when the feature amount calculation region is larger than the motion determination threshold; the number of feature amount calculation regions that has been determined to have a motion In the room when is less than the predetermined number A second motion determining means that determines that the visual target is in a stationary state, and determines that the monitoring target in the room is in a moving state when the number of feature amount calculation areas determined to be in motion is greater than a predetermined number; and Video captured by the imaging unit in the abnormality detection device including the abnormality determination unit that determines whether or not an abnormality has occurred in the monitoring target in the room based on a predetermined number of past movement determination results by the second movement determination unit And a threshold value control unit for controlling a threshold value for motion determination based on the cloudiness degree determination data, based on the cloudiness degree determination data generation unit. The degree determination data generation means is characterized in that it extracts a high frequency component from the image picked up by the image pickup means as haze degree determination data.
[0014]
According to a sixth abnormality detection apparatus of the present invention, an imaging unit that captures an image in a room, the image captured by the imaging unit is divided into a plurality of feature amount calculation areas, and each feature amount calculation area is captured by the imaging unit. Extraction means for extracting image feature values from video, detection means for detecting temporal changes in image feature values extracted by the extraction means for each feature value calculation area for each predetermined time interval, and feature value calculation areas by the detection means The temporal change amount of the image feature amount detected for each time and the threshold value for motion determination are compared for each feature amount calculation region, and if the temporal change amount of the image feature amount is equal to or less than the threshold value for motion determination, A first motion determination unit that determines that there is no motion in the feature amount calculation area and determines that the feature amount calculation region has a motion when the feature amount calculation region is larger than the motion determination threshold; the number of feature amount calculation regions that has been determined to have a motion In the room when is less than the predetermined number A second motion determining means that determines that the visual target is in a stationary state, and determines that the monitoring target in the room is in a moving state when the number of feature amount calculation areas determined to be in motion is greater than a predetermined number; and Video captured by the imaging unit in the abnormality detection device including the abnormality determination unit that determines whether or not an abnormality has occurred in the monitoring target in the room based on a predetermined number of past movement determination results by the second movement determination unit And a threshold value control unit for controlling a threshold value for motion determination based on the cloudiness degree determination data, based on the cloudiness degree determination data generation unit. The degree determination data generation means includes a high frequency component extraction means for extracting a high frequency component from an image captured by the imaging means, and a noise component from the high frequency component extracted by the high frequency component extraction means. And removed by, characterized in that it comprises a high-frequency component correction means for generating a haze determination data.
[0015]
According to a seventh abnormality detection apparatus of the present invention, an imaging unit that captures an image in a room, the image captured by the imaging unit is divided into a plurality of feature amount calculation regions, and each feature amount calculation region is captured by the imaging unit. Extraction means for extracting image feature values from video, detection means for detecting temporal changes in image feature values extracted by the extraction means for each feature value calculation area for each predetermined time interval, and feature value calculation areas by the detection means The temporal change amount of the image feature amount detected for each time and the threshold value for motion determination are compared for each feature amount calculation region, and if the temporal change amount of the image feature amount is equal to or less than the threshold value for motion determination, A first motion determination unit that determines that there is no motion in the feature amount calculation area and determines that the feature amount calculation region has a motion when the feature amount calculation region is larger than the motion determination threshold; the number of feature amount calculation regions that has been determined to have a motion In the room when is less than the predetermined number A second motion determining means that determines that the visual target is in a stationary state, and determines that the monitoring target in the room is in a moving state when the number of feature amount calculation areas determined to be in motion is greater than a predetermined number; and Video captured by the imaging unit in the abnormality detection device including the abnormality determination unit that determines whether or not an abnormality has occurred in the monitoring target in the room based on a predetermined number of past movement determination results by the second movement determination unit And a threshold value control unit for controlling a threshold value for motion determination based on the cloudiness degree determination data, based on the cloudiness degree determination data generation unit. The degree determination data generating means divides the video captured by the imaging means into a plurality of feature amount calculation areas, and extracts extraction means for extracting image feature quantities from the video captured by the imaging means for each feature amount calculation area. For each of a plurality of predetermined feature quantity calculation areas, a value related to the difference between the image feature quantities between the area and the surrounding area is calculated based on the image feature quantity extracted by the extraction unit for each quantity calculation area. And a means for generating haze degree determination data by accumulating values related to the difference between the image feature amounts calculated for each of the plurality of feature amount calculation regions.
[0016]
According to an eighth abnormality detection apparatus of the present invention, an image capturing unit that captures an image in a room, the image captured by the image capturing unit is divided into a plurality of feature amount calculation regions, and each feature amount calculation region is captured by the image capturing unit. Extraction means for extracting image feature values from video, detection means for detecting temporal changes in image feature values extracted by the extraction means for each feature value calculation area for each predetermined time interval, and feature value calculation areas by the detection means The temporal change amount of the image feature amount detected for each time and the threshold value for motion determination are compared for each feature amount calculation region, and if the temporal change amount of the image feature amount is equal to or less than the threshold value for motion determination, A first motion determination unit that determines that there is no motion in the feature amount calculation area and determines that the feature amount calculation region has a motion when the feature amount calculation region is larger than the motion determination threshold; the number of feature amount calculation regions that has been determined to have a motion In the room when is less than the predetermined number A second motion determining means that determines that the visual target is in a stationary state, and determines that the monitoring target in the room is in a moving state when the number of feature amount calculation areas determined to be in motion is greater than a predetermined number; and Video captured by the imaging unit in the abnormality detection device including the abnormality determination unit that determines whether or not an abnormality has occurred in the monitoring target in the room based on a predetermined number of past movement determination results by the second movement determination unit And a threshold value control unit for controlling a threshold value for motion determination based on the cloudiness degree determination data, based on the cloudiness degree determination data generation unit. The degree determination data generating means divides the video captured by the imaging means into a plurality of feature amount calculation areas, and extracts extraction means for extracting image feature quantities from the video captured by the imaging means for each feature amount calculation area. For each of a plurality of predetermined feature quantity calculation areas, a value related to the difference between the image feature quantities between the area and the surrounding area is calculated based on the image feature quantity extracted by the extraction unit for each quantity calculation area. And a means for generating haze degree determination data by squaring and summing the values relating to the difference between the image feature amounts calculated for each of the plurality of feature amount calculation regions. .
As the threshold control means, for example, one that makes the motion determination threshold smaller as the degree of cloudiness represented by the cloudiness determination data is higher. As the threshold control means, for example, when the haze degree determination data generated by the haze degree determination data generation means is larger than the first predetermined value and not more than the second predetermined value which is larger than the first predetermined value. In addition, the one that sets the motion determination threshold value to a value smaller than the normal motion determination threshold value is used. It is preferable to provide means for notifying that the abnormality determination is impossible when the cloudiness determination data generated by the cloudiness determination data generation means is equal to or less than the first predetermined value.
[0017]
DETAILED DESCRIPTION OF THE INVENTION
Embodiments of the present invention will be described below.
[0018]
[1] Description of the first embodiment
A first embodiment of the present invention will be described with reference to FIGS.
[0019]
FIG. 1 is a schematic diagram illustrating an arrangement example of an abnormal state detection device.
[0020]
In FIG. 1, 1 is a bathroom. 2 is a dressing room. 3 is a corridor. 4 is a kitchen.
[0021]
A door 5 is provided on the wall between the bathroom 1 and the dressing room 2 so that the door 5 can be opened and closed. When the door 5 is opened, a person enters the bathroom 1 from the dressing room 2 or exits from the bathroom 1 to the dressing room 2. can do.
[0022]
Reference numeral 8 denotes a camera that captures an image in the bathroom 1. Reference numeral 9 denotes an abnormal state detection device that detects an abnormal state based on the video of the camera 8.
[0023]
Reference numeral 10 denotes a first alarm device arranged in the bathroom 1 in order to issue an alarm to a person (a bather) entering the bathroom 1. Reference numeral 11 denotes a second alarm device, for example, disposed in the kitchen 4 in order to notify a person outside the bathroom 1 of the bather's abnormality. Reference numeral 12 denotes a third alarm device disposed in the kitchen 4, for example, in order to notify a person outside the bathroom 1 that movement determination (abnormality determination) is impossible.
[0024]
FIG. 2 shows the configuration of the camera 8 and the abnormal state detection device 9.
[0025]
The camera 8 includes a lens 8a and a CCD 8b. The CCD 8b photoelectrically converts the optical image in the bathroom through the lens 8a and outputs it as an electrical signal. The output signal of the CCD 8b is sent to the preprocessing circuit 91. The preprocessing circuit includes a CDS circuit, an AGC circuit, and an ADC.
[0026]
The output signal of the CCD 8b is subjected to correlated double sampling processing by the CDS circuit, further gain adjusted to an optimum amplitude by the AGC circuit, converted to a digital signal by the ADC, and input to the YC separation circuit 92.
[0027]
The YC separation circuit 92 generates a luminance signal Y and color difference signals RY and BY from the digital CCD signal. The signals Y, RY, and BY generated by the YC separation circuit 92 are sent to the integration circuit 93 for each area.
[0028]
As shown in FIG. 3, the region-by-region integration circuit 93 divides one screen into a plurality of regions (feature amount calculation regions) D11 to D44 having the same area, and in each feature amount calculation region, signals Y, RY, An integration process or an average process is performed for each BY. In this example, it is assumed that integration processing is performed for each of the signals Y, RY, and BY. The results (hereinafter referred to as image feature amounts) D11 to D44 accumulated for each region are stored in the image feature amount RAM 96.
[0029]
The luminance signal Y generated by the YC separation circuit 92 is also sent to the HPF 94. The HPF 94 passes only the high frequency component of the input luminance signal Y. The high frequency component of the Y signal that has passed through the HPF 94 is sent to the high frequency component integrating circuit 95. The high frequency component integration circuit 95 integrates the high frequency components of the input luminance signal Y during the effective video period, and stores the integration result (hereinafter referred to as high frequency component integration data) in the image feature amount RAM 96.
[0030]
The reason why the HPF 94 and the high frequency component integrating circuit 95 are provided will be described. When dirt is attached to the surface of the lens 8a of the camera 8 and the image picked up by the camera 8 is blurred, the high frequency component extracted by the HPF 94 and the high frequency component integrating circuit 95 decreases. Also, when the bathroom becomes difficult to see due to mist or steam due to the mist function or the like, the high frequency components extracted by the HPF 94 and the high frequency component integrating circuit 95 are lowered.
[0031]
Therefore, in this embodiment, the degree of blurring of the image, in other words, the cloudiness of the lens 8a is detected by the HPF 94 and the high frequency component integrating circuit 95. As will be described later, when the degree of dirt (cloudiness) on the surface of the lens 8a is light and motion detection is possible, the motion detection sensitivity is increased and good motion detection is performed. When the dirt (cloudiness) on the surface of the lens 8a is heavy and motion detection becomes impossible, a warning is given to notify that motion detection is impossible.
[0032]
The timing control unit 99 outputs a CCD drive pulse for driving the CCD 8b. The timing control unit 99 outputs a sampling pulse for the CDS circuit and a sampling clock for the ADC.
[0033]
The timing control unit 99 is configured so that predetermined YC separation, region-by-region integration, high-frequency component integration, and writing to the image feature RAM are normally performed on the CCD digital signal input at a fixed timing by the timing control unit 99. Send control signals to each circuit. Further, the timing control unit 99 outputs a Vsync signal, which is a video vertical synchronization signal, as an interrupt signal to a microcomputer 97 (hereinafter referred to as a microcomputer) that performs motion determination.
[0034]
The microcomputer 97 performs parameter setting such as region position setting on the timing control unit 99. Further, the microcomputer 97 performs gain control of the AGC circuit and timing control through the CPU I / F 98 in order to optimize the brightness of the image, and controls the exposure time of the CCD 8b. Further, the microcomputer 97 receives an interrupt signal from the timing control unit 99 and performs a motion determination in synchronization with the video output.
[0035]
FIG. 4 shows a motion determination processing procedure performed by the microcomputer 97.
[0036]
The microcomputer 97 is in an input standby state until the Vsync signal is input (step 1). When the Vsync signal is input, the microcomputer 97 performs the following processing.
[0037]
When the Vsync signal is input, the microcomputer 97 sets parameters in the timing control circuit 99 through the CPU I / F 98 (step 2). Thereafter, the image feature values in the entire feature value calculation area of the previous field stored in the image feature value RAM 96 are read through the CPU I / F 98 (step 3) and stored in the RAM inside the microcomputer 97 (step 4).
[0038]
After storing the image feature values of the entire feature value calculation area in the RAM inside the microcomputer 97 (step 5), the microcomputer 97 reads the high-frequency component integrated data of the previous field stored in the image feature value RAM 96 through the CPU I / F 98. Then, correction processing is performed to remove noise components from the read high-frequency component integration data (step 6).
[0039]
This correction process will be described. When the subject is dark, the microcomputer 97 controls the gains of the CDS circuit, the AGC circuit, and the ADC to increase the gain. When the gain is increased, the noise component increases, and thus the high frequency component also increases. Therefore, it is necessary to perform correction such as removing noise components from the acquired high-frequency component integrated data in accordance with the gain setting value of the AGC circuit.
[0040]
It is necessary to measure in advance the noise component necessary for performing the correction. The AGC circuit gain setting value is automatically changed by capturing a solid image having no change in luminance on the entire screen and changing the overall brightness of the solid image. At this time, since the high frequency component integration data acquired by the high frequency component integration circuit 95 becomes a high frequency component integration value by noise components corresponding to each gain setting value of the AGC circuit, this data is used as noise component data for the gain setting value. Is held in the microcomputer 97.
[0041]
In step 6, the noise component data for the same gain setting value as the AGC gain setting value in that field is subtracted from the high-frequency component integration data in the previous field acquired from the image feature amount RAM 96, so that only the signal component is obtained. The integrated high-frequency component value is calculated. The calculated high-frequency component integrated value is used as haze degree determination data for the lens surface 8a. Note that the degree of haze increases as the haze degree determination data (high-frequency component integrated value) decreases.
[0042]
When the high-frequency component integrated value of only the signal component is calculated in this way, the microcomputer 97 determines that the high-frequency component integrated value of the previous field (high-frequency component integrated value of only the signal component, hereinafter the same) is the predetermined second. It is determined whether or not it is equal to or less than a threshold value (TH2) (step 7). If the high-frequency component integrated value in the previous field is larger than the second threshold value (TH2), the microcomputer 97 sets the threshold value SUBTH (motion determination threshold value) used for motion determination to a predetermined value SUBTH1 (step 8), Proceed to step 12.
[0043]
When the high-frequency component integrated value of the previous field is equal to or smaller than the second threshold (TH2), the microcomputer 97 has a predetermined first threshold (TH1) equal to or smaller than the high-frequency component integrated value of the previous field smaller than the second threshold (TH2). It is determined whether or not (step 9). When the high-frequency component integrated value in the previous field is equal to or smaller than the second threshold value (TH2) and greater than the first threshold value (TH1), the microcomputer 97 uses the threshold value SUBTH used for motion determination to increase the sensitivity of motion detection. Is set to ½ of the predetermined value SUBTH1 (step 10), and then the process proceeds to step 12.
[0044]
If the integrated value of the high frequency component in the previous field is equal to or less than the first threshold value (TH1), the microcomputer 97 determines that the motion determination is impossible and outputs a third alarm signal (step 11). The third alarm device 12 is driven by the third alarm signal. And it returns to step 1.
[0045]
In step 12, the microcomputer 97 sets a variable CountM representing the number of regions with movement to zero. Then, a motion determination for each feature amount calculation region is performed from the amount of change in the feature amount of each feature amount calculation region. That is, the microcomputer 97 calculates the absolute value of the difference between the image feature value Dxy (t) read this time from the image feature value RAM 96 and the image feature value Dxy (t−1) read last time from the image feature value RAM 96 for each calculation area. It is determined whether or not | Dxy (t) −Dxy (t−1) | is larger than the threshold value SUBTH (step 13).
[0046]
This determination is made for each feature amount calculation region for the change amount of the integrated value of the luminance signal Y, the change amount of the integrated value of the color difference signal RY, and the change amount of the integrated value of the color difference signal BY. If any one of the change amounts is larger than SUBTH, it is determined that there is a motion in the feature amount calculation area.
[0047]
If it is determined that there is a motion in the feature amount calculation area, the microcomputer 97 increments the variable CountM by 1 (step 14). In this way, when the motion determination for all the feature quantity calculation regions is completed (step 15), the microcomputer 97 determines whether or not the number of regions CountM determined to have motion is greater than a predetermined value TH_Count ( Step 16).
[0048]
If the number of areas CountM determined to have movement is less than or equal to the predetermined value TH_Count, the microcomputer 97 determines that there is no movement in the bather (step 17), and the number of areas CountM determined to have movement is If it is greater than the predetermined value TH_Count, the microcomputer 97 determines that the bather has movement (step 18). Then, the process proceeds to Step 19.
[0049]
In step 19, the microcomputer 97 determines whether or not it has been determined that the bather has moved a predetermined M times or more in the movement determination for the past N times (N fields). If the bather has moved more than M times in the past N movement determinations, the microcomputer 97 determines that the bather is in a moving state (step 20), and if not, the microcomputer 97 stops the bather. It determines with it being in a state (step 21).
[0050]
If it is determined that the bather is in a stationary state, it is determined whether or not a timekeeping operation is being performed by an internal timer (not shown) (step 22). When the timekeeping operation is not being performed, the timekeeping operation by the internal timer is started (step 23), and then the process proceeds to step 24.
[0051]
In step 24 and subsequent steps, the microcomputer 97 determines whether or not the time measured by the internal timer has exceeded the first predetermined time (step 24), and the second predetermined time longer than the first predetermined time has elapsed. It is determined whether or not (step 26).
[0052]
When the time measured by the internal timer has passed the first predetermined time, the microcomputer 97 outputs a first alarm signal, drives the first alarm device 10 installed in the bathroom 1, and the person in the bathroom 1. A warning is urged (step 25). Then, the process proceeds to step 26.
[0053]
When the time measured by the internal timer has passed the second predetermined time, the microcomputer 97 determines that an abnormality has occurred in the bather in the bathroom 1, outputs a second alarm signal, The second alarm device 11 installed in the kitchen 4 is driven to notify an external person of an emergency (step 27). And it returns to step 1.
[0054]
The alarm in step 25 is executed to wake up a bather who may be asleep in the bathroom 1. The alarm in step 27 is to determine that an abnormal state such as unconsciousness has occurred in the bathroom and to notify an external person of an emergency situation.
[0055]
If it is determined in step 19 that the bather is in a moving state (step 20), the microcomputer 97 resets the timekeeping operation by the internal timer (if the timekeeping operation is performed) and outputs an alarm signal. After stopping (when the first, second or third alarm signal is output) (step 28), the process returns to step 1.
[0056]
According to the above embodiment, even when the image captured by the camera 8 is slightly blurred, it is possible to accurately detect that the bather is in an abnormal state. In addition, an alarm (third alarm device) is issued when the blur of the image taken by the camera 8 becomes serious due to dirt on the lens 8a of the camera 8, fog in the bathroom, or the like, and the motion cannot be determined normally. , False detection can be prevented.
[0057]
In the above embodiment, the image feature amount is extracted for each of a plurality of feature amount calculation areas set in one screen, but may be extracted only in one region set in one screen. . In this case, it is determined whether the bather is in a moving state or a stationary state based on whether the temporal change in the image feature amount extracted in one region is larger than a predetermined motion determination threshold value. do it.
[0058]
In the above embodiment, as shown in FIG. 5, when the high-frequency component integrated value is greater than TH2, the motion determination threshold SUBTH is set to SUBTH1, and when the high-frequency component integrated value X is greater than TH1 and less than TH2, The motion determination threshold value SUBTH is set to SUBTH1 / 2. However, as shown in FIG. 6, when the high-frequency component integrated value X is greater than TH1 and less than or equal to TH2, the range is from SUBTH1 / 2 to SUBTH1. It is preferable to change the motion determination threshold SUBTH stepwise in accordance with the high-frequency component integrated value.
[0059]
[2] Description of the second embodiment
[0060]
A second embodiment of the present invention will be described with reference to FIGS.
[0061]
Also in the second embodiment, the arrangement example of the abnormal state detection device is the same as that in FIG.
[0062]
FIG. 7 shows the configuration of the camera and the abnormal state detection device. In FIG. 7, the same components as those in FIG.
[0063]
In the abnormal state detection device of FIG. 7, the HPF 94 and the high-frequency component integration circuit 95 that are provided for calculating the cloudiness determination data in FIG. 2 are not provided.
[0064]
In the second embodiment, the cloudiness determination data for determining the cloudiness of the lens surface 8a is calculated based on the integrated luminance values of the feature amount calculation regions D11 to D44. A method for calculating the cloudiness determination data will be described.
[0065]
First, the basic concept will be described. When the haze of the lens surface 8a increases, the difference in image data (brightness integrated value) between adjacent areas decreases. Therefore, the difference between the image data between all adjacent regions is integrated, and the integration result is used as the cloudiness determination data. The smaller the cloudiness determination data (integration result), the higher the cloudiness.
[0066]
Since the amount of processing increases when image data differences between all adjacent regions are integrated, here, the image data differences between some adjacent regions are integrated. Since there are two methods for calculating the cloudiness determination data, each method will be described.
[0067]
(1) First method
In the first method, first, as shown in FIG. 8, each feature amount calculation area D11 to D44 is divided into an outermost area (without hatching) and a central area (area with hatching; D22, D23) surrounded by them. , D32, D33).
[0068]
Next, for each of the regions D22, D23, D32, and D33 other than the outermost region, a value related to a difference in image feature amount between that region and its surrounding region is calculated. In the first method, for each of the regions D22, D23, D32, and D33 other than the outermost region, an absolute value of difference data between the region and the average value of the surrounding region is calculated.
[0069]
For example, when the area is D22, the absolute value ΔD22 of the difference data from the average value of the surrounding area is calculated based on the following equation (1). Dxy is a luminance integrated value of the region Dxy.
[0070]
ΔD22 = | D22 − {(D11 + D12 + D13 + D21 + D23 + D31 + D32 + D33) / 8} (1)
[0071]
That is, ΔD22 is obtained as an absolute value of the difference between the average value of the luminance integrated values of the eight regions around the region D22 and the luminance integrated value of the region D22.
[0072]
The absolute values ΔD22, ΔD23, ΔD32, and ΔD33 of the difference data with respect to the average value of the peripheral area obtained for each of the areas D22, D23, D32, and D33 are integrated, and the integration result is used as cloudiness determination data.
[0073]
In order to make the degree of change in cloudiness steep, the values obtained by squaring ΔD22, ΔD23, ΔD32, and ΔD33 may be integrated, and the result of the integration may be used as cloudiness determination data.
[0074]
(2) Second method
In the second method, first, as shown in FIG. 8, each feature amount calculation area D11 to D44 is divided into an outermost area (without hatching) and a central area (area with hatching; D22, D23) surrounded by them. , D32, D33).
[0075]
Next, for each of the regions D22, D23, D32, and D33 other than the outermost region, a value related to a difference in image feature amount between that region and its surrounding region is calculated. In the second method, for each of the regions D22, D23, D32, and D33 other than the outermost region, the maximum value of the absolute values of the difference between that region and its surrounding region is calculated.
[0076]
For example, when the region is D22, the maximum value ΔD22 of the absolute values of the differences from the surrounding region is calculated based on the following equation (2). Dxy is a luminance integrated value of the region Dxy.
[0077]
ΔD22 = MAX [| D22-D11 |, | D22-D12 |, | D22-D13 |, | D22-D21 |, | D22-D23 |, | D22-D31 |, | D22-D32 |, | D22-D33 |] (2)
[0078]
The maximum values [Delta] D22, [Delta] D23, [Delta] D32, [Delta] D33 among the absolute values of the differences from the surrounding areas obtained for each of the areas D22, D23, D32, D33 are integrated, and the integration result is used as cloudiness determination data. .
[0079]
In order to make the degree of change in cloudiness steep, the values obtained by squaring ΔD22, ΔD23, ΔD32, and ΔD33 may be integrated, and the result of the integration may be used as cloudiness determination data.
[0080]
FIG. 9 shows a motion determination processing procedure performed by the microcomputer 97.
[0081]
9, the same steps as those in FIG. 4 are denoted by the same step numbers, and the description thereof is omitted.
[0082]
The difference between the processing procedure of FIG. 9 and the processing procedure of FIG. 4 is the following two points.
[0083]
(1) Between step 3 and step 4, the peripheral light amount correction of the lens 8a (lens peripheral light amount correction) is performed on the image feature amounts in the entire feature amount calculation area of the previous field acquired from the image feature amount RAM 96 in step 3. The processing step (step 31) for performing () is inserted.
[0084]
The lens peripheral light amount correction is for compensating the light amount input from the peripheral portion of the lens 8a. As shown in FIG. 10, the image feature amounts d11 to 11 for each feature amount calculation region acquired from the image feature amount RAM 96 are used. This is performed by multiplying d44 by a lens peripheral light amount correction coefficient predetermined for each feature amount calculation region.
[0085]
In step 4, the image feature amount after the lens peripheral light amount correction is stored in the RAM inside the microcomputer 97.
[0086]
(2) In step 6, the cloudiness determination data is calculated based on the image feature values in the entire feature value calculation region of the previous field stored in the RAM inside the microcomputer 97 in step 4.
[0087]
That is, in step 6, the cloudiness determination data is calculated by the first method or the second method described above.
[0088]
Also in this embodiment, as shown in steps 7 to 10, when the cloudiness determination data is larger than TH2, the motion determination threshold SUBTH is set to SUBTH1, and the cloudiness determination data is larger than TH1. When TH2 or less, the motion determination threshold SUBTH is set to SUBTH1 / 2, but when the cloudiness determination data is greater than TH1 and less than TH2, the range is from SUBTH1 / 2 to SUBTH1, It is preferable that the motion determination threshold SUBTH is changed stepwise according to the cloudiness determination data.
[0089]
【The invention's effect】
According to the present invention, it is possible to accurately detect the occurrence of an abnormality such as a person falling down in a room such as a bathroom or toilet.
[Brief description of the drawings]
FIG. 1 is a schematic diagram illustrating an arrangement example of an abnormal state detection device.
FIG. 2 is a block diagram illustrating a configuration of a camera and an abnormal state detection apparatus according to a first embodiment.
FIG. 3 is a schematic diagram showing a plurality of feature amount calculation areas set in one screen.
FIG. 4 is a flowchart showing the operation of the microcomputer according to the first embodiment.
FIG. 5 is a graph showing an example of a relationship between a high-frequency component integrated value and a motion determination threshold value SUBTH.
FIG. 6 is a graph showing another example of the relationship between the high-frequency component integrated value and the motion determination threshold value SUBTH.
FIG. 7 is a block diagram illustrating a configuration of a camera and an abnormal state detection apparatus according to a second embodiment.
FIG. 8 is an explanatory diagram for explaining a method of calculating haze degree determination data.
FIG. 9 is a flowchart showing the operation of the microcomputer according to the second embodiment.
FIG. 10 is an explanatory diagram for explaining lens peripheral light amount correction;
[Explanation of symbols]
8 Camera
8a lens
8b CCD
91 Pre-processing circuit
92 YC separation circuit
93 Integration circuit for each area
94 HPF
95 High frequency component integration circuit
96 Image feature amount RAM
97 Microcomputer
98 CPU I / F
99 Timing controller

Claims (8)

An imaging means for capturing an image in the room, an extracting means for extracting an image feature quantity from the video imaged by the imaging means, a detecting means for detecting a temporal change amount of the image feature quantity extracted by the extracting means at predetermined time intervals, The temporal change amount of the image feature amount detected by the detection means is compared with the threshold value for motion determination, and the monitoring target in the room is stationary when the temporal change amount of the image feature amount is equal to or less than the threshold value for motion determination. Based on the motion determination unit that determines that the monitoring target in the room is in a moving state when the threshold is larger than the motion determination threshold, and the motion determination result of the past predetermined number of times by the motion determination unit, In the abnormality detection apparatus provided with abnormality determination means for determining whether an abnormality has occurred in the monitoring target,
Haze degree determination data generating means for generating haze degree determination data for the lens of the imaging means based on the video imaged by the imaging means, and a threshold control means for controlling the motion determination threshold value based on the haze degree determination data With steps ,
The abnormality detection apparatus, wherein the cloudiness determination data generation means extracts a high-frequency component from the image captured by the imaging means as the cloudiness determination data . An imaging means for capturing an image in the room, an extracting means for extracting an image feature quantity from the video imaged by the imaging means, a detecting means for detecting a temporal change amount of the image feature quantity extracted by the extracting means at predetermined time intervals, The temporal change amount of the image feature amount detected by the detection means is compared with the threshold value for motion determination, and the monitoring target in the room is stationary when the temporal change amount of the image feature amount is equal to or less than the threshold value for motion determination. Based on the motion determination unit that determines that the monitoring target in the room is in a moving state when the threshold is larger than the motion determination threshold, and the motion determination result of the past predetermined number of times by the motion determination unit, In the abnormality detection apparatus provided with abnormality determination means for determining whether an abnormality has occurred in the monitoring target,
Haze degree determination data generating means for generating haze degree determination data for the lens of the imaging means based on the video imaged by the imaging means, and threshold control means for controlling the motion determination threshold value based on the haze degree determination data With
The cloudiness determination data generating means removes noise components from the high frequency component extracted by the high frequency component extraction means and the high frequency component extraction means for extracting the high frequency component from the image picked up by the imaging means, and the cloudiness degree determination data An abnormality detection apparatus comprising high-frequency component correction means for generating An imaging means for capturing an image in the room, an extracting means for extracting an image feature quantity from the video imaged by the imaging means, a detecting means for detecting a temporal change amount of the image feature quantity extracted by the extracting means at predetermined time intervals, The temporal change amount of the image feature amount detected by the detection means is compared with the threshold value for motion determination, and the monitoring target in the room is stationary when the temporal change amount of the image feature amount is equal to or less than the threshold value for motion determination. Based on the motion determination unit that determines that the monitoring target in the room is in a moving state when the threshold is larger than the motion determination threshold, and the motion determination result of the past predetermined number of times by the motion determination unit, In the abnormality detection apparatus provided with abnormality determination means for determining whether an abnormality has occurred in the monitoring target,
Haze degree determination data generating means for generating haze degree determination data for the lens of the imaging means based on the video imaged by the imaging means, and threshold control means for controlling the motion determination threshold value based on the haze degree determination data With
The cloudiness determination data generation means
An extraction unit that divides a video captured by the imaging unit into a plurality of feature amount calculation regions and extracts an image feature amount from the video captured by the imaging unit for each feature amount calculation region;
Based on the image feature amount extracted by the extraction means for each feature amount calculation region, for each of a plurality of predetermined feature amount calculation regions, a value related to the difference in image feature amount between that region and its surrounding region Means for calculating, and
Means for generating cloudiness determination data by accumulating values relating to differences in image feature amounts calculated for each of the plurality of feature amount calculation regions;
An abnormality detection device comprising: An imaging means for capturing an image in the room, an extracting means for extracting an image feature quantity from the video imaged by the imaging means, a detecting means for detecting a temporal change amount of the image feature quantity extracted by the extracting means at predetermined time intervals, The temporal change amount of the image feature amount detected by the detection means is compared with the threshold value for motion determination, and the monitoring target in the room is stationary when the temporal change amount of the image feature amount is equal to or less than the threshold value for motion determination. Based on the motion determination unit that determines that the monitoring target in the room is in a moving state when the threshold is larger than the motion determination threshold, and the motion determination result of the past predetermined number of times by the motion determination unit, In the abnormality detection apparatus provided with abnormality determination means for determining whether an abnormality has occurred in the monitoring target,
Haze degree determination data generating means for generating haze degree determination data for the lens of the imaging means based on the video imaged by the imaging means, and threshold control means for controlling the motion determination threshold value based on the haze degree determination data With
The cloudiness determination data generation means
An extraction unit that divides a video captured by the imaging unit into a plurality of feature amount calculation regions and extracts an image feature amount from the video captured by the imaging unit for each feature amount calculation region;
Based on the image feature amount extracted by the extraction means for each feature amount calculation region, for each of a plurality of predetermined feature amount calculation regions, a value related to the difference in image feature amount between that region and its surrounding region Means for calculating, and
Means for generating haze degree determination data by squaring and summing values relating to differences in image feature amounts calculated for the plurality of feature amount calculation regions,
An abnormality detection device comprising: An image capturing unit that captures an image in a room, an image capturing unit that divides the image captured by the image capturing unit into a plurality of feature amount calculation regions, and extracts an image feature amount from the image captured by the image capturing unit for each feature amount calculation region; A detecting means for detecting a temporal change amount of the image feature amount extracted by the extracting means for each feature amount calculation region for each predetermined time interval, and a temporal feature of the image feature amount detected for each feature amount calculation region by the detecting means. The amount of change and the threshold for motion determination are compared for each feature amount calculation area, and if the temporal change amount of the image feature amount is equal to or less than the threshold for motion determination, it is determined that there is no motion in the feature amount calculation area. First motion determination means for determining that there is motion in the feature amount calculation area when the motion determination threshold value is greater than the threshold for motion determination, and monitoring in the room when the number of feature amount calculation regions determined to have motion is equal to or less than a predetermined number Determine that the subject is stationary and move Second motion determination means for determining that the monitoring target in the room is in a moving state when the number of feature amount calculation areas determined to have is greater than a predetermined number, and a predetermined predetermined number of motion determinations by the second motion determination means On the basis of the result, in the abnormality detection device provided with abnormality determination means for determining whether an abnormality has occurred in the monitoring target in the room,
Haze degree determination data generating means for generating haze degree determination data for the lens of the imaging means based on the video imaged by the imaging means, and threshold control means for controlling the motion determination threshold value based on the haze degree determination data With
The abnormality detection apparatus, wherein the cloudiness determination data generation means extracts a high-frequency component from the image captured by the imaging means as the cloudiness determination data . An image capturing unit that captures an image in a room, an image capturing unit that divides the image captured by the image capturing unit into a plurality of feature amount calculation regions, and extracts an image feature amount from the image captured by the image capturing unit for each feature amount calculation region; A detecting means for detecting a temporal change amount of the image feature amount extracted by the extracting means for each feature amount calculation region for each predetermined time interval, and a temporal feature of the image feature amount detected for each feature amount calculation region by the detecting means. The amount of change and the threshold for motion determination are compared for each feature amount calculation area, and if the temporal change amount of the image feature amount is equal to or less than the threshold for motion determination, it is determined that there is no motion in the feature amount calculation area. First motion determination means for determining that there is motion in the feature amount calculation area when the motion determination threshold value is greater than the threshold for motion determination, and monitoring in the room when the number of feature amount calculation regions determined to have motion is equal to or less than a predetermined number Determine that the subject is stationary and move Second motion determination means for determining that the monitoring target in the room is in a moving state when the number of feature amount calculation areas determined to have is greater than a predetermined number, and a predetermined predetermined number of motion determinations by the second motion determination means On the basis of the result, in the abnormality detection device provided with abnormality determination means for determining whether an abnormality has occurred in the monitoring target in the room,
Haze degree determination data generating means for generating haze degree determination data for the lens of the imaging means based on the video imaged by the imaging means, and threshold control means for controlling the motion determination threshold value based on the haze degree determination data With
The cloudiness determination data generating means removes noise components from the high frequency component extracted by the high frequency component extraction means and the high frequency component extraction means for extracting the high frequency component from the image picked up by the imaging means, and the cloudiness degree determination data An abnormality detection apparatus comprising high-frequency component correction means for generating An image capturing unit that captures an image in a room, an image capturing unit that divides the image captured by the image capturing unit into a plurality of feature amount calculation regions, and extracts an image feature amount from the image captured by the image capturing unit for each feature amount calculation region; A detecting means for detecting a temporal change amount of the image feature amount extracted by the extracting means for each feature amount calculation region for each predetermined time interval, and a temporal feature of the image feature amount detected for each feature amount calculation region by the detecting means. The amount of change and the threshold for motion determination are compared for each feature amount calculation area, and if the temporal change amount of the image feature amount is equal to or less than the threshold for motion determination, it is determined that there is no motion in the feature amount calculation area. First motion determination means for determining that there is motion in the feature amount calculation area when the motion determination threshold value is greater than the threshold for motion determination, and monitoring in the room when the number of feature amount calculation regions determined to have motion is equal to or less than a predetermined number Determine that the subject is stationary and move Second motion determination means for determining that the monitoring target in the room is in a moving state when the number of feature amount calculation areas determined to have is greater than a predetermined number, and a predetermined predetermined number of motion determinations by the second motion determination means On the basis of the result, in the abnormality detection device provided with abnormality determination means for determining whether an abnormality has occurred in the monitoring target in the room,
Haze degree determination data generating means for generating haze degree determination data for the lens of the imaging means based on the video imaged by the imaging means, and threshold control means for controlling the motion determination threshold value based on the haze degree determination data With
The cloudiness determination data generation means
An extraction unit that divides a video captured by the imaging unit into a plurality of feature amount calculation regions and extracts an image feature amount from the video captured by the imaging unit for each feature amount calculation region;
Based on the image feature amount extracted by the extraction means for each feature amount calculation region, for each of a plurality of predetermined feature amount calculation regions, a value related to the difference in image feature amount between that region and its surrounding region Means for calculating, and
Means for generating cloudiness determination data by accumulating values relating to differences in image feature amounts calculated for each of the plurality of feature amount calculation regions;
An abnormality detection device comprising: An image capturing unit that captures an image in a room, an image capturing unit that divides the image captured by the image capturing unit into a plurality of feature amount calculation regions, and extracts an image feature amount from the image captured by the image capturing unit for each feature amount calculation region; A detecting means for detecting a temporal change amount of the image feature amount extracted by the extracting means for each feature amount calculation region for each predetermined time interval, and a temporal feature of the image feature amount detected for each feature amount calculation region by the detecting means. The amount of change and the threshold for motion determination are compared for each feature amount calculation area, and if the temporal change amount of the image feature amount is equal to or less than the threshold for motion determination, it is determined that there is no motion in the feature amount calculation area. First motion determination means for determining that there is motion in the feature amount calculation area when the motion determination threshold value is greater than the threshold for motion determination, and monitoring in the room when the number of feature amount calculation regions determined to have motion is equal to or less than a predetermined number Determine that the subject is stationary and move Second motion determination means for determining that the monitoring target in the room is in a moving state when the number of feature amount calculation areas determined to have is greater than a predetermined number, and a predetermined predetermined number of motion determinations by the second motion determination means On the basis of the result, in the abnormality detection device provided with abnormality determination means for determining whether an abnormality has occurred in the monitoring target in the room,
Haze degree determination data generating means for generating haze degree determination data for the lens of the imaging means based on the video imaged by the imaging means, and threshold control means for controlling the motion determination threshold value based on the haze degree determination data With
The cloudiness determination data generation means
An extraction unit that divides a video captured by the imaging unit into a plurality of feature amount calculation regions and extracts an image feature amount from the video captured by the imaging unit for each feature amount calculation region;
Based on the image feature amount extracted by the extraction means for each feature amount calculation region, for each of a plurality of predetermined feature amount calculation regions, a value related to the difference in image feature amount between that region and its surrounding region Means for calculating, and
Means for generating haze degree determination data by squaring and summing values relating to differences in image feature amounts calculated for the plurality of feature amount calculation regions,
An abnormality detection device comprising:

Images