JP6555617B2 - Human detection system - Google Patents

Description

  The present invention relates to a human detection system.

  2. Description of the Related Art Conventionally, a human detection system that detects an intruder using an image captured by an imaging device has been proposed (see, for example, Patent Document 1).

  The human detection system (human body detection device) described in Patent Literature 1 includes an image input unit that captures an image in a monitoring area, and a determination unit that determines the presence or absence of a human body from the image input by the image input unit. Yes. In this human detection system, a difference image is obtained from two images of the same region imaged with a certain time difference, and an area value for each image appearing in the binarized difference image is used to obtain a human body. The presence or absence of is determined.

JP 2000-348268 A

  By the way, for example, sunlight entering from a window locally illuminates a floor surface or the like, so that a “sunshine spot” having a higher brightness than other areas may be generated in the monitoring area. Such a sunshine spot moves in the monitoring area with the passage of time. In the conventional human detection system that determines the presence / absence of a person from the difference image, such a movement of the sunshine spot is erroneously determined as a movement of the person in the monitoring area, and there is a possibility that erroneous detection occurs.

  The present invention has been made in view of the above-described reasons, and an object thereof is to provide a human detection system in which erroneous detection caused by sunlight or the like hardly occurs.

The human detection system according to the first aspect of the present invention includes a detection unit that detects the presence or absence of a person in the detection region of the analysis image captured by the imaging unit, and the detection unit includes the presence of a person in the detection region. A movement detection mode for detecting the presence or absence of a person and a stay detection mode for detecting the presence or absence of a person in a stay area composed of the identification area stagnant in the detection area among the identification areas surrounding the person in the detection area And when the motion amount accumulated over time with the evaluation value according to the motion of the person in the stay area is equal to or greater than the detection threshold in the stay detection mode, the person is present. When the state where the amount of movement is less than the detection threshold continues for the holding time, it is determined that there is no person, and the amount of movement is reset every time a reset time shorter than the holding time elapses. Characterized in that it is configured to.

The human detection system according to the second aspect of the present invention includes a detection unit that detects the presence or absence of a person in the detection region of the analysis image captured by the imaging unit, and the detection unit includes the presence of a person in the detection region. A movement detection mode for detecting the presence or absence of a person and a stay detection mode for detecting the presence or absence of a person in a stay area composed of the identification area stagnant in the detection area among the identification areas surrounding the person in the detection area And when the motion amount accumulated over time with the evaluation value according to the motion of the person in the stay area is equal to or greater than the detection threshold in the stay detection mode, the person is present. Each time the amount of motion becomes equal to or greater than the detection threshold, the amount of motion is reset, and when the amount of motion is equal to or greater than the detection threshold continuously for a predetermined number of times at a time interval that satisfies a predetermined condition, People exist Characterized in that it is configured to determine not to.

  The present invention has an advantage that erroneous detection caused by sunlight or the like is unlikely to occur.

FIG. 1 is a block diagram of a human detection system according to the first embodiment. FIG. 2 is a perspective view illustrating a usage example of the human detection system according to the first embodiment. 3A to 3G are diagrams showing analysis images of seven frames when a person enters the detection area from the outside of the detection area, stops on the spot, and then leaves the detection area. FIG. 4 is a flowchart illustrating the operation of the human detection system according to the first embodiment. FIG. 5 is a graph illustrating an example of a change in the amount of movement in the stay detection mode for the human detection system and the comparative example according to the first embodiment. FIG. 6A is a graph illustrating an example of a change in the amount of movement in the stay detection mode for the human detection system according to the first embodiment. FIG. 6B is a graph showing an example of a change in motion amount in the stay detection mode for the comparative example. FIG. 7 is a graph showing an example of a change in the amount of movement in the stay detection mode for the human detection system according to the second embodiment. FIG. 8 is a block diagram of a human detection system according to the third embodiment.

  Each of the following embodiments relates to a human detection system, and more particularly to a human detection system that detects a person using an image.

(Embodiment 1)
(1) Outline The human detection system 100 according to the present embodiment includes a detection unit 31 and a creation unit 32, as shown in FIG. The detection unit 31 detects the presence or absence of a person 70 (see FIG. 2) in the detection region 21 (see FIG. 3A) of the analysis image captured by the imaging unit 2. The creation unit 32 creates an identification area 22 (see FIG. 3A) surrounding a person in the detection area 21 in the analysis image.

  The detection unit 31 detects the presence of the person 70 in the stay area 23 (see FIG. 3E) including the movement detection mode for detecting the presence or absence of the person 70 in the detection area 21 and the identification area 22 stagnated in the detection area 21. A stay detection mode for detecting presence or absence.

In the stay detection mode, the detection unit 31 determines that the person 70 exists when the motion amount obtained by accumulating the evaluation values according to the movement of the person 70 in the stay region 23 is equal to or greater than the detection threshold Th2 (second threshold). It is configured. In the stay detection mode, the detection unit 31 is configured to determine that the person 70 does not exist when the state in which the amount of motion is less than the detection threshold Th2 continues for the holding time T1. Moreover, the detection part 31 is comprised so that the amount of movement may be reset whenever reset time T2 shorter than holding time T1 passes in stay detection mode.

The “motion amount” here is a value obtained by accumulating evaluation values along the time axis, that is, an accumulated value of evaluation values. The “evaluation value” here can be derived from the analysis image, and is a value corresponding to the movement of the person 70 in the stay area 23. In this embodiment, as an example, the “motion amount” is a value (cumulative value) obtained by accumulating the amount of change in the luminance value of the pixel in the stay area 23 along the time axis. In this case, the “evaluation value” corresponding to the movement of the person in the stay area 23 is the amount of change in the luminance value of the pixels in the stay area 23.

  According to the human detection system 100 of the present embodiment, the detection unit 31 resets the amount of movement every time a reset time T2 shorter than the holding time T1 elapses in the stay detection mode. For this reason, even if a “sunshine spot” having a higher brightness than the other areas occurs in the stay area 23, the amount of movement in the stay area 23 is reset before the movement amount reaches the detection threshold Th2. Is less likely to occur. Therefore, for example, even if the sunlight (sunlight) entering from the window illuminates the floor surface or the like locally, a sunshine spot is generated, so that erroneous detection due to the movement of the sunshine spot is less likely to occur. As a result, the human detection system 100 according to the present embodiment has an advantage that erroneous detection due to sunlight or the like hardly occurs. The “sunlight” referred to here includes the western sun, the morning sun, reflected light from neighboring buildings, and the like.

(2) Details Hereinafter, the human detection system 100 of the present embodiment will be described in detail with reference to the drawings. However, the configuration described below is merely an example of the present invention, and the present invention is not limited to the following embodiment. Therefore, various modifications other than this embodiment can be made according to the design and the like as long as they do not depart from the technical idea of the present invention.

(2.1) Configuration of Human Detection System The human detection system 100 includes an imaging unit 2, a control unit 3, a transmission unit 4, a storage unit 5, and a timer 7, as shown in FIG. .

  The imaging unit 2 is a camera having an image sensor and an optical member such as a lens. In the present embodiment, as an example, as shown in FIG. 2, the imaging unit 2 is attached to the ceiling 80 of the office and images the inside of the office from above. In the present embodiment, the imaging unit 2 captures a moving image and outputs an image (analysis image) to the control unit 3 at a predetermined frame rate.

  The control unit 3 has a computer such as a microcomputer (microcomputer) as a main component, and performs appropriate processing by executing a program recorded in a memory of the computer by a processor of the computer. The program may be recorded in advance in a memory, may be provided through an electric communication line such as the Internet, or may be provided by being recorded in a recording medium such as a memory card. The control unit 3 includes a detection unit 31, a creation unit 32, and a derivation unit 33.

  The detection unit 31 detects the presence or absence of a person 70 in the detection area 21 of the analysis image captured by the imaging unit 2. The “detection area” referred to here may be, for example, a partial area of an image (analysis image) captured by the imaging unit 2 or may be the entire area of the image captured by the imaging unit 2. In FIG. 2, the detection area 21 is shown to exist in the real space, but this is an expression for facilitating the understanding of the correspondence between the real space and the analysis image. It does not exist in space. That is, the detection area 21 is only virtually shown in FIG. 2, and is actually set in the analysis image.

  The creation unit 32 creates an identification area 22 surrounding the person 70 in the detection area 21 of the analysis image in the analysis image. When creating the identification region 22 by the creation unit 32, the method for extracting the person 70 in the detection region 21 by the creation unit 32 is not particularly limited, and various human detection techniques can be applied. For example, a method of extracting the person 70 in the detection area 21 using the background difference may be used, and the person 70 in the detection area 21 is extracted by deriving the moving speed of the person 70 in the analysis image using the inter-frame difference. The technique to do may be used. In the present embodiment, the detection area 21 is a partial area of the analysis image, and the creation unit 32 creates the identification area 22 that surrounds the person 70 not only for the detection area 21 but also for the entire analysis image. Here, the creation unit 32 creates the rectangular identification region 22 so as to surround the person 70 in the analysis image based on the width and height of the person 70 in the analysis image. That is, the identification area 22 is an area created so as to include the person 70 in the analysis image.

  The deriving unit 33 derives the position change amount of the identification area 22 in the detection area 21. The deriving unit 33 derives the amount of movement in the stay area 23. The “stay area” here is the identification area 22 stagnated in the detection area 21. The method of deriving the position change amount of the identification area 22 and the movement amount in the stay area 23 will be described in the section “(2.2) Basic operation”.

  The transmission unit 4 transmits a control signal corresponding to the detection result of the detection unit 31 (the presence or absence of the person 70) to the load 6 to be controlled. The transmission part 4 may transmit a control signal directly with respect to the load 6, and may transmit a control signal indirectly via a repeater etc. As the communication method of the transmission unit 4, an appropriate communication method such as a wireless communication method using Wi-Fi (registered trademark) or a wired communication method using a signal line can be applied.

  The storage unit 5 stores a stagnation threshold Th1 (first threshold) and a detection threshold Th2 (second threshold) as evaluation indexes. Furthermore, the storage unit 5 also stores values such as the holding time T1 and the reset time T2.

  The timer 7 is used for measuring the holding time T1 (see FIG. 5) and the reset time T2 (see FIG. 5). The timer 7 outputs the count value to the control unit 3. The timer 7 receives the reset signal from the control unit 3 and resets the count value.

  Here, the detection unit 31 detects the presence / absence of the person 70 based on the movement detection mode for detecting the presence / absence of the person 70 in the detection area 21 of the analysis image and the amount of movement in the stay area 23 as the detection mode. And a stay detection mode. In the movement detection mode, the detection unit 31 fixes the position of the identification area 22 in the detection area 21 when the position change amount of the identification area 22 is less than the stagnation threshold Th1, and sets the identification area 22 as the stay area 23. The presence / absence of the person 70 is detected in the stay detection mode for the area 23. That is, the stay area 23 is an identification area 22 that is regarded as stagnant in the detection area 21 because the amount of change in position in the detection area 21 is less than the stagnation threshold Th1.

  As long as the detection unit 31 operates in the movement detection mode, the creation unit 32 updates the position of the identification region 22 as the person 70 moves within the detection region 21 of the analysis image of the human detection system 100. Thus, the position of the person 70 in the detection area 21 is tracked in the identification area 22. That is, in the movement detection mode, the detection unit 31 detects the presence or absence of the person 70 for the entire detection area 21, and the position of the person 70 is determined even when the person 70 moves greatly within the detection area 21. By tracking, the presence or absence of the person 70 is continuously detected.

  On the other hand, in the stay detection mode, the detection unit 31 is in the stay area 23 that is the identification area 22 stagnated in the detection area 21, that is, in the stay area 23 that is the identification area 22 in which the position in the detection area 21 is fixed. The presence / absence of the person 70 is detected. Therefore, in the stay detection mode, the detection unit 31 detects the slight movement of the person 70 even when the person 70 is stopped at one place in the detection area 21 (stagnation). The presence or absence of the person 70 in the area 23 can be detected.

  In the stay detection mode, the detection unit 31 detects the presence or absence of the person 70 in the stay region 23 by comparing the amount of movement in the stay region 23 with the detection threshold Th2. That is, when the amount of motion in the stay area 23 is equal to or greater than the detection threshold Th2, the detection unit 31 determines that the person 70 exists. On the other hand, if the state where the amount of motion in the stay area 23 is less than the detection threshold Th2 continues for the holding time T1, the detection unit 31 determines that the person 70 does not exist. That is, once the detection mode of the detection unit 31 is switched to the stay detection mode, even if the amount of movement is less than the detection threshold Th2, the detection unit 31 does not immediately determine that the person 70 does not exist and retains it. During the time T1, it is determined that the person 70 exists. Since the holding time T1 is timed by the timer 7, the detection unit 31 outputs a reset signal to the timer 7 every time it determines that the person 70 exists in the stay area 23, and the count value of the holding time T1. To reset.

  However, the detection unit 31 does not completely switch between the movement detection mode and the stay detection mode but operates in the movement detection mode in principle. When the stay area 23 appears in the detection area 21 (that is, when the stagnation of the identification area 22 occurs), the detection unit 31 detects the presence or absence of the person 70 only in the stay area 23 in the stay detection mode. And the detection part 31 detects the presence or absence of the person 70 in area | regions other than the stay area | region 23 among the detection areas 21 in a movement detection mode. That is, when the stay area 23 appears in the detection area 21, the detection unit 31 partially sets the detection mode to the stay detection mode only for the stay area 23, and the detection mode is changed from the movement detection mode to the stay detection mode. It is not completely switched to.

  Further, switching from the stay detection mode to the movement detection mode is performed when the detection unit 31 determines that the person 70 does not exist in the stay area 23 in the stay detection mode. That is, when the state where the amount of motion in the stay area 23 is less than the detection threshold Th2 continues for the holding time T1, the detection unit 31 ends the stay detection mode. The length of the holding time T1 is appropriately set depending on the use of the human detection system 100 or the like. In this embodiment, as an example, the holding time T1 is about 3 minutes to 10 minutes. The holding time T1 is preferably variable and can be arbitrarily changed by the user.

  The detection unit 31 cancels the stay area 23 when switching from the stay detection mode to the movement detection mode. The cancellation of the stay area 23 here means that the stay area 23 set in the detection area 21 is deleted. When the stay area 23 is canceled, the detection unit 31 may simply cancel the setting of the stay area 23 instead of returning the stay area 23 to the identification area 22. That is, when the stay area 23 is released, neither the stay area 23 nor the identification area 22 remains at the same position as the stay area 23 in the detection area 21.

  In the state where the stay areas 23 are set at a plurality of locations in the detection area 21, the detection unit 31 may perform switching from the stay detection mode to the movement detection mode for each stay area 23, The stay area 23 may be performed all at once.

  In the movement detection mode, the detection unit 31 detects the presence or absence of the person 70 based on whether or not the identification region 22 created by the creation unit 32 exists in the detection region 21. If the identification area 22 exists in the detection area 21, the detection unit 31 determines that the person 70 exists. If the identification area 22 does not exist in the detection area 21, the detection unit 31 determines that the person 70 does not exist. Thereby, the detection part 31 can distinguish and detect the person 70 in the detection area 21 and the person 70 outside the detection area 21. Note that the detection unit 31 preferably determines that the identification area 22 exists in the detection area 21 when the identification area 22 partially overlaps the detection area 21.

  In addition, the detection unit 31 detects the presence or absence of the person 70 depending on whether or not the stay area 23 exists in the detection area 21. If the stay area 23 exists in the detection area 21, the detection unit 31 determines that the person 70 exists. If the stay area 23 does not exist in the detection area 21, the detection unit 31 determines that the person 70 does not exist. In addition, it is preferable that the detection unit 31 determines that the stay area 23 exists in the detection area 21 when the stay area 23 partially overlaps the detection area 21.

  By the way, when the detection unit 31 determines that the person 70 exists in the detection area 21, the final detection result of the human detection system 100 differs depending on whether the detection mode of the detection unit 31 is the stay detection mode or the movement detection mode. Specifically, in the human detection system 100, when the detection unit 31 determines that the person 70 exists in the movement detection mode, it is determined as “moving state”, and the detection unit 31 determines that the person 70 exists in the stay detection mode. When it is done, it is judged as “stay state”. When the detection results of both “movement state” and “stay state” are obtained in duplicate, the human detection system 100 preferentially adopts the detection result of “stay state”. When the detection unit 31 determines that the person 70 does not exist in either the movement detection mode or the stay detection mode, the human detection system 100 determines that “the absence state”.

  In the present embodiment, the load 6 to be controlled by the human detection system 100 is a lighting device. The load 6 has at least two operation modes of “lighting” and “lighting off”.

  If the detection unit 31 determines that there is a person 70 when the operation mode of the load 6 is “off”, the transmission unit 4 transmits a control signal instructing the “lighting” to the load 6. The load 6 that has received the control signal instructing “lighting” switches the operation mode from “lighting out” to “lighting”. On the other hand, when the operation mode of the load 6 is “lighting”, if the detection unit 31 determines that the person 70 does not exist, the transmission unit 4 transmits a control signal instructing “extinguishing” to the load 6. . The load 6 that has received the control signal instructing “turn off” switches the operation mode from “lit” to “turned off”.

  The transmission unit 4 outputs the detection result in a form that can distinguish whether the person 70 in the detection area 21 is moving (in a moving state) or staying (in a staying state). That is, the transmission unit 4 transmits different control signals to the load 6 when the detection unit 31 determines that the person 70 exists in the stay detection mode and when it determines that the person 70 exists in the movement detection mode. . By doing so, the load 6 can be controlled according to the movement and stay of the person 70.

(2.2) Basic Operation Hereinafter, the basic operation of the human detection system 100 will be described with reference to FIGS. 3A to 3G. FIGS. 3A to 3G show analysis images of seven frames when a person 70 enters the detection area 21 from outside the detection area 21 and stops (stagnates) on the spot, and then leaves the detection area 21. FIG. The analysis images shown in FIGS. 3A to 3G are arranged in time series, but the time interval between frames is appropriate.

  First, in the state of FIG. 3A, the person 70 exists in a position outside the detection area 21 in the analysis image (that is, in the field of view of the imaging unit 2). Therefore, the creation unit 32 creates the identification area 22 in the analysis image so as to surround the person 70 in the analysis image. Here, since the person 70 does not exist in the detection area 21, the identification area 22 is created outside the detection area 21 in the analysis image. At this time, the detection unit 31 operates in the movement detection mode, and determines that the person 70 is not present in the detection area 21, that is, the “absence state”.

  Next, in the state of FIG. 3B, the person 70 has moved into the detection area 21 in the analysis image. Here, the identification area 22 is located at the end of the detection area 21, and a part of the identification area 22 overlaps the detection area 21. At this time, since the detection unit 31 operates in the movement detection mode, it is determined that the person 70 exists, that is, the “movement state”.

  When the determination is made that “the movement state”, the deriving unit 33 calculates the movement distance of the center of the identification area 22 between two consecutive frames as the distance L, and uses the distance L as the position change amount of the identification area 22. To derive. From the state of FIG. 3B to the state of FIG. 3C, the person 70 moves in the detection area 21 in the analysis image. Here, FIGS. 3B and 3C represent two consecutive frames of analysis images. Therefore, in the state of FIG. 3C, the distance L from the center of the identification area 22 in FIG. 3B to the center of the identification area 22 in FIG. 3C is derived as a position change amount.

  If the positional change amount (distance L) of the identification area 22 derived in this way is equal to or greater than the stagnation threshold Th1, the detection unit 31 continues to detect the presence or absence of a person in the movement detection mode. Here, in both cases of FIG. 3B and FIG. 3C, the detection unit 31 operates in the movement detection mode and determines that the person 70 exists (moving state).

  Next, in the state of FIG. 3D, the person 70 exists in the same position as the state of FIG. 3C. Therefore, the position change amount (distance L) of the identification region 22 becomes less than the stagnation threshold Th1, and the detection unit 31 switches the detection mode from the movement detection mode to the stay detection mode. At this time, the detection unit 31 fixes the position of the identification area 22 and changes the identification area 22 to the stay area 23 as shown in FIG. 3E. That is, if there is an identification area 22 whose position change amount is less than the stagnation threshold Th1, the detection unit 31 sets the identification area 22 as the stay area 23 and detects the presence or absence of the person 70 in the stay detection mode only for the stay area 23. As described above, the detection mode is partially set as the stay detection mode.

  In the state of FIG. 3E, since the detection unit 31 operates in the stay detection mode, the presence or absence of the person 70 in the stay area 23 is detected by comparing the amount of movement in the stay area 23 with the detection threshold Th2. . Further, when the detection unit 31 detects the presence or absence of the person 70 in the stay area 23 in the stay detection mode, the presence or absence of another person outside the stay area 23 and in the detection area 21 of the analysis image is determined. Detect in the movement detection mode. Therefore, when a person other than the person 70 in the stay area 23 enters the detection area 21, the detection unit 31 can detect the presence (entrance) of this person in the movement detection mode.

  In the present embodiment, the “motion amount” is a value (cumulative value) obtained by accumulating the amount of change in the luminance value of the pixel in the stay area 23 along the time axis. In this case, the “evaluation value” corresponding to the movement of the person in the stay area 23 is the amount of change in the luminance value of the pixels in the stay area 23.

  That is, as shown in FIG. 3E, when a person 70 in the stay area 23 moves, the luminance value of the pixel in the stay area 23 changes according to the movement of the person. Therefore, in this embodiment, every time the control unit 3 acquires an analysis image from the imaging unit 2, a difference image is generated from the latest analysis image and the analysis image one frame before. The luminance value of each pixel in the difference image corresponds to the amount of change in luminance value. The deriving unit 33 accumulates the luminance value for each of all the pixels in the stay area 23 in the difference image every time a difference image is obtained, thereby obtaining the luminance value for each of all the pixels in the stay area 23. Deriving the cumulative value of change.

  Hereinafter, an example of a method for deriving the amount of motion when focusing on one pixel in the stay area 23 (hereinafter referred to as “target pixel”) will be described. When the luminance value of the pixel of interest changes by “1” with the movement of a person in the stay area 23, the cumulative value of the amount of change in luminance value for the pixel of interest becomes “1”. Further, when the luminance value of the pixel of interest changes by “3” with the movement of the person in the stay area 23, the cumulative value of the amount of change in the luminance value for the pixel of interest becomes “4”. Furthermore, when the luminance value of the pixel of interest changes by “2” with the movement of the person in the stay area 23, the cumulative value of the amount of change in luminance value for the pixel of interest becomes “6”. The cumulative value of the luminance value variation derived in this way is the “motion amount” for the pixel of interest.

  Then, the deriving unit 33 derives a motion amount for each of all the pixels in the stay area 23. In the stay detection mode, the detection unit 31 compares the motion amount with the detection threshold Th2 for each of all the pixels in the stay region 23, and sets a detection flag for a pixel with the motion amount equal to or greater than the detection threshold Th2. When the detection flag is set for pixels having a predetermined number of pixels (for example, 8 pixels) or more, the detection unit 31 determines that a person 70 exists in the stay area 23. In other words, the detection unit 31 determines that the person 70 exists in the stay area 23 when the amount of movement of the pixels in the stay area 23 equal to or greater than the prescribed number of pixels is equal to or greater than the detection threshold Th2.

  And if the detection part 31 judges that the person 70 exists in the stay area | region 23, as mentioned above, the count value of holding | maintenance time T1 will be reset (it makes it zero). Furthermore, if the detection part 31 judges that the person 70 exists in the stay area | region 23, the amount of movement will be reset. That is, each time the detection unit 31 determines that there is a person 70 in the stay region 23, the detection unit 31 resets the accumulated value of the amount of change in luminance value for all the pixels in the stay region 23 (sets it to zero). Note that the specified number of pixels may be one pixel. In this case, the detection unit 31 determines that the person 70 exists in the stay area 23 when the detection flag is set even for one pixel.

  In the state of FIG. 3E, since the person 70 in the stay area 23 is moving and the detection unit 31 is operating in the stay detection mode, the detection unit 31 is in a state where the person 70 exists, that is, the “stay state”. Judge.

  Next, in the state of FIG. 3F, the person 70 has left the stay area 23 and moved outside the stay area 23. Therefore, the creation unit 32 creates a new identification area 22 so as to surround the person 70 in the analysis image. In the state of FIG. 3F, the person 70 is moving in the detection area 21 in the analysis image. At this time, the detection unit 31 operates in the movement detection mode, and determines that the person 70 exists, that is, the “movement state”.

  However, even if the amount of movement in the stay area 23 is less than the detection threshold Th2, the detection unit 31 does not immediately determine that the person 70 does not exist, and determines that the person 70 exists during the holding time T1. Continue. That is, even in the state of FIG. 3F, the stay region 23 shown in FIG. 3E remains without being released. Therefore, in the state of FIG. 3F, the detection unit 31 operates in the stay detection mode, and determines that the person 70 exists, that is, the “stay state”. As a result, in the state of FIG. 3F, both detection results of “moving state” and “stay state” are obtained redundantly, so the detection unit 31 preferentially adopts the detection result of “stay state”. To do.

  Further, from the state of FIG. 3F to the state of FIG. 3G, the person 70 moves in the detection area 21 in the analysis image. Here, FIGS. 3F and 3G show two consecutive frames of analysis images. Therefore, in the state of FIG. 3G, the distance L from the center of the identification area 22 in FIG. 3F to the center of the identification area 22 in FIG. 3G is derived as a position change amount. Further, in the state of FIG. 3G, the stay region 23 remains without being released, and as a result, both detection results of the “movement state” and the “stay state” are obtained in duplicate. 31 preferentially adopts the detection result of “stay state”.

  After the person 70 leaves the detection area 21 after FIG. 3G, the detection result in the movement detection mode becomes “absent state”. Therefore, when the stay area 23 is released after the holding time T1 has elapsed, the detection unit 31. Determines that the person 70 does not exist, that is, the “absent state”. However, even in the movement detection mode, the detection unit 31 is not limited to the configuration in which the identification area 22 disappears in the detection area 21 and immediately determines “absence state”, but determines “absence state” after the delay time has elapsed. May be. The delay time may be the same length as the holding time T1, or may be a different length of time.

(2.3) Specific Operation Hereinafter, a specific operation of the human detection system 100 will be described.

  In the human detection system 100 of the present embodiment, the detection unit 31 has a function of resetting the motion amount derived by the deriving unit 33 every time the reset time T2 elapses in the stay detection mode. Here, the reset time T2 is shorter than the holding time T1. That is, even if the detection unit 31 does not determine that the person 70 exists in the stay area 23, the amount of movement is reset when the reset time T2 elapses.

  In the present embodiment, the detection unit 31 uses the output of the timer 7 to count the reset time T2 inside the control unit 3. Therefore, the detection unit 31 resets the count value of the reset time T2 every time the reset time T2 elapses, but does not output the reset signal to the timer 7. That is, the count value of the holding time T1 is not reset even when the reset time T2 elapses. In short, the reset performed by the detection unit 31 every time the reset time T2 elapses is an internal reset (internal reset) of the control unit 3, and does not include a reset of the timer 7 outside the control unit 3.

  Next, a specific operation of the human detection system 100 will be described with reference to FIG. FIG. 4 is a flowchart illustrating an example of the operation of the detection unit 31.

  In step S1, the detection unit 31 starts operating in the movement detection mode. At this time, the detection unit 31 detects the presence or absence of the person 70 in the detection area 21 in the movement detection mode. As an initial value, the detection unit 31 determines that the person 70 does not exist in the detection area 21, that is, the “absence state” (step S2).

  Thereafter, while the detection unit 31 determines that no person 70 exists in the detection area 21 (NO in step S3), the detection unit 31 maintains the determination result of “absence state”.

  When the detection unit 31 detects that the person 70 exists in the detection area 21, that is, the person 70 has entered the detection area 21 (YES in step S3), the detection unit 31 is in a state where the person 70 exists, That is, it is determined as “moving state” (step S4).

  If it is determined as “moving state”, the detection unit 31 compares the position change amount of the identification region 22 derived by the deriving unit 33 with the stagnation threshold Th1 (step S5). If the position change amount (distance L) of the identification area 22 is equal to or greater than the stagnation threshold Th1 (NO in step S5), the detection unit 31 returns to step S3 and continues to detect the presence or absence of a person in the movement detection mode. To do.

  On the other hand, if the position change amount (distance L) of the identification area 22 is less than the stagnation threshold Th1 (YES in step S5), the detection unit 31 proceeds to step S6 and switches the detection mode from the movement detection mode to the stay detection mode. . When the detection mode of the detection unit 31 is switched to the stay detection mode in step S <b> 6, the detection unit 31 changes the identification area 22 to the stay area 23. At this time, the detection unit 31 resets the amount of motion derived by the deriving unit 33 and resets the count value of the holding time T1. Therefore, at the time when the detection mode of the detection unit 31 is switched to the stay detection mode, the movement amount in the stay region 23 and the count value of the holding time T1 are both zero.

  Thereafter, the detection unit 31 detects the presence or absence of the person 70 in the stay region 23 by comparing the amount of movement in the stay region 23 with the detection threshold Th2 (step S7). When the detection unit 31 detects that the amount of movement is equal to or greater than the detection threshold Th2 and the person 70 is present in the stay area 23 (NO in step S7), the detection unit 31 is in a state where the person 70 exists, that is, “stay state” Is determined (step S8). If it is determined that “stay state”, the process returns to step S7 via steps S9 and S10, and the detection unit 31 continues to detect the presence or absence of a person in the stay detection mode. In step S <b> 9, the detection unit 31 resets the amount of motion derived by the deriving unit 33. In step S10, the detection unit 31 transmits a reset signal to the timer 7 to reset the count value of the holding time T1.

  On the other hand, if the amount of motion is less than the detection threshold Th2 (YES in step S7), the detection unit 31 determines whether or not the reset time T2 has elapsed (S11). If the reset time T2 has elapsed (YES in step S11), the detection unit 31 proceeds to step S13 via step S12. If the reset time T2 has not elapsed (NO in step S11), the detection unit 31 proceeds to step S13 without passing through step S12. In step S <b> 12, the detection unit 31 resets the amount of motion derived by the deriving unit 33.

  In step S13, the detection unit 31 determines whether or not the holding time T1 has elapsed. Since the count value of the holding time T1 is not reset even if the reset time T2 elapses, the elapse of the holding time T1 is synonymous with the state where the movement amount is less than the detection threshold Th2 continues for the holding time T1. . If the holding time T1 has not elapsed (NO in step S13), the process returns to step S7, and the detection unit 31 continues to detect the presence or absence of a person in the stay detection mode. When the holding time T1 has elapsed (YES in step S13), the process proceeds to step S14, and the detection unit 31 switches the detection mode from the stay detection mode to the movement detection mode. For the areas other than the stay area 23 in the detection area 21, since the detection unit 31 originally detects the presence or absence of a person in the movement detection mode, the detection mode is switched from the stay detection mode to the movement detection mode. Is only in the stay area 23.

  When the detection mode of the detection unit 31 is switched to the movement detection mode in step S14, the process returns to step S3, and the detection unit 31 operates in the movement detection mode. When the detection mode of the detection unit 31 is switched to the movement detection mode in step S14, the detection unit 31 cancels the stay area 23.

(2.4) Influence of Sunlight Hereinafter, with reference to FIGS. 5, 6A and 6B, the influence of sunlight in the human detection system 100 of the present embodiment will be described while comparing with the comparative example. In particular, here we assume an evening situation in which a western sun comes in through a window. 5, 6A, and 6B are graphs showing an example of a change in the amount of motion in the stay detection mode with the horizontal axis as the time axis. In addition, all of FIG. 5, FIG. 6A, and FIG. 6B represent the amount of motion when paying attention to one pixel in the stay area 23.

  The comparative example assumed here is different from the human detection system 100 of the present embodiment in that there is no function of resetting the motion amount derived by the deriving unit 33 every time the reset time T2 elapses in the stay detection mode. To do. That is, in the comparative example, if there is no step S11 and step S12 in the flowchart shown in FIG. 4 and the amount of motion is less than the detection threshold Th2 (YES in step S7), the detection unit 31 does not go through steps S11 and S12. Proceed to step S13. Other configurations and functions are the same in the comparative example and the human detection system 100 of the present embodiment.

  First, the influence of sunlight (here, the western sun) in the comparative example will be described. For example, there is a case where a “sunshine spot” having a higher luminance than the other areas is generated in the detection area 21 by sunlight shining from a window locally illuminating a floor surface or the like. Such a sunshine spot moves in the detection area 21 as time passes. Therefore, when a sunshine spot is generated in the stay area 23, the movement of the sunshine spot is erroneously determined as a movement of the person 70 in the stay area 23, and there is a possibility that erroneous detection occurs. Usually, since the sunshine spot moves about several centimeters per minute, in the detection mode in which the detection unit 31 operates with relatively low sensitivity in order to detect the movement of the person 70 as in the movement detection mode, erroneous detection is not performed. Hard to occur. On the other hand, in the detection mode in which the detection unit 31 operates with relatively high sensitivity in order to detect the fine movement of the person 70 in the stay area 23 as in the stay detection mode, erroneous detection is likely to occur due to the movement of the sunlight spot. .

  In other words, in the comparative example, when the detection unit 31 is operating in the stay detection mode and a sunshine spot is generated in the stay area 23, even if no person 70 exists in the stay area 23, FIG. As shown by a broken line in FIG. Specifically, as the sunshine spot moves, the amount of motion, which is the cumulative value of the change in the luminance value of the pixel in the stay area 23, gradually increases with time. When the amount of movement in the stay area 23 is equal to or greater than the detection threshold Th2, the detection unit 31 determines that the person 70 exists in the stay area 23. If it is determined that the person 70 exists in the stay area 23, the movement amount is reset.

  Here, when the state in which the amount of motion in the stay region 23 is less than the detection threshold Th2 continues for the holding time T1, the detection unit 31 ends the stay detection mode. In other words, if the movement amount in the stay region 23 does not reach the detection threshold Th2 within the holding time T1, the detection unit 31 ends the stay detection mode. Therefore, erroneous detection occurs when the time until the amount of movement reaches the detection threshold Th2 from the initial value (zero) due to the movement of the sunshine spot is shorter than the holding time T1.

  Therefore, in the comparative example, as shown in FIG. 6B, it is caused by the movement of the sunshine spot before the holding time T <b> 1 elapses from the time when the person 70 leaves the stay area 23 (“X1” in the figure). Then, when the amount of movement reaches the detection threshold Th2, it is determined as “stay state”. Usually, the moving speed of the sunshine spot increases as the sunset approaches. As a result, the detection unit 31 continues to determine the “stay state” until the sunshine spot is out of the stay area 23 or the sunshine spot disappears at sunset.

  On the other hand, according to the human detection system 100 of the present embodiment, the detection unit 31 moves in the stay detection mode every time the reset time T2 shorter than the holding time T1 elapses. Reset the amount. Therefore, even if the amount of movement in the stay region 23 gradually increases with the passage of time due to the movement of the sunshine spot while the detection unit 31 is operating in the stay detection mode, a solid line in FIG. As shown, the amount of motion is reset in small increments. Specifically, the amount of motion in the stay region 23 is set by setting the reset time T2 to be shorter than the time until the amount of motion reaches the detection threshold Th2 from the initial value (zero) due to the movement of the sunshine spot. , Reset before reaching the detection threshold Th2. That is, in the human detection system 100 of the present embodiment, unlike the comparative example (shown by a broken line in FIG. 5), the amount of motion in the stay area 23 is reset before reaching the detection threshold Th2, and thus erroneous detection is unlikely to occur. Become.

  Therefore, in the human detection system 100 according to the present embodiment, as shown in FIG. 6A, from the time when the person 70 leaves the stay area 23 (time “X1” in the figure) until the holding time T1 elapses. The amount of movement is suppressed to less than the detection threshold Th2. As a result, when the holding time T1 elapses from the time when the person 70 leaves the stay area 23 (time "X1" in the figure), the detection unit 31 ends the stay detection mode and releases the stay area 23. Thereby, the detection unit 31 can determine the “absence state”. In FIG. 6A, for comparison with the comparative example (see FIG. 6B), the movement amount in the stay region 23 changes even after the lapse of the holding time T <b> 1, but actually, the holding time T <b> 1 has elapsed. Then, when the stay area 23 is released, the derivation of the movement amount is also ended. For this reason, in the example of FIG. 6A, the moving speed of the solar radiation spot is increased just before sunset, and no erroneous detection occurs even if the amount of movement reaches the detection threshold Th2 within the reset time T2.

  Here, the influence of the western sun has been described. However, according to the human detection system 100 of the present embodiment, not only the western sun but also misdetection due to, for example, reflected light from the morning sun and nearby buildings is less likely to occur. Become.

  Further, the length of the reset time T2 is appropriately set depending on the use of the human detection system 100 or the relationship with the holding time T1. If the holding time T1 is about 3 to 10 minutes, as an example, the reset time T2 is preferably about 5 to 20 seconds. The reset time T2 is preferably variable and can be arbitrarily changed by the user.

(3) Effect As described above, according to the human detection system 100 of the present embodiment, the detection unit 31 calculates the amount of movement every time the reset time T2 shorter than the holding time T1 elapses in the stay detection mode. Reset. For this reason, even if a “sunshine spot” having a higher brightness than the other areas occurs in the stay area 23, the amount of movement in the stay area 23 is reset before the movement amount reaches the detection threshold Th2. Is less likely to occur. Therefore, for example, even if a sunlight spot is generated by sunlight shining from a window locally illuminating the floor surface or the like, erroneous detection due to the movement of the sunlight spot is less likely to occur. As a result, the human detection system 100 according to the present embodiment has an advantage that erroneous detection due to sunlight or the like hardly occurs.

  In addition, as in the present embodiment, the evaluation value is preferably the amount of change in the luminance value of the pixels in the stay area 23. According to this configuration, for example, it is possible to obtain the amount of motion by a relatively simple process that only accumulates the luminance value for each pixel in the stay area 23 in the difference image. Therefore, it is possible to reduce the processing load for the process for obtaining the motion amount.

(4) Modifications Modifications of Embodiment 1 are listed below.

  In the first embodiment, the imaging unit 2, the transmission unit 4, the storage unit 5, and the timer 7 are components of the human detection system 100, but the imaging unit 2, the transmission unit 4, the storage unit 5, and the timer 7 are human detections. It may not be included in the components of the system 100. Furthermore, in Embodiment 1, the load 6 is not a component of the human detection system 100, but the load 6 may be included in the component of the human detection system 100.

  In addition, a plurality of components of the human detection system 100 may be housed in a single casing to constitute a single device, or may be provided separately in a plurality of devices, and the plurality of devices may be organically related. A single function may be exhibited.

  Further, in the first embodiment, the detection unit 31, the creation unit 32, and the derivation unit 33 are provided in the control unit 3, and the transmission unit 4, the storage unit 5, and the timer 7 are externally attached to the control unit 3. However, it is not limited to this configuration. For example, the transmission unit 4, the storage unit 5, and the timer 7 may be provided in the control unit 3. Furthermore, the deriving unit 33 is not an essential component of the human detection system 100, and the deriving of the movement amount in the stay area 23 and the positional change amount of the identification area 22 may be performed by the detection unit 31, for example.

  Further, the load 6 to be controlled by the human detection system 100 is not limited to the lighting device, and may be an electric device having at least two operation modes such as an air conditioner and a cooking device.

  The deriving unit 33 does not necessarily need to use the center of the identification area 22 as a reference point when deriving the position change amount of the identification area 22 in the detection area 21. That is, for example, the derivation unit 33 may be configured to calculate the distance between the representative points of the identification area 22 before and after moving, and derive this distance as the position change amount of the identification area 22. The representative points here are, for example, the center of gravity and the apex of the identification region 22.

  When the detection unit 31 determines that a person is present in the detection area 21, the final detection result of the human detection system 100 differs depending on whether the detection mode of the detection unit 31 is the stay detection mode or the movement detection mode. Is not a required configuration. That is, in the human detection system 100, when the detection unit 31 determines that there is a person in the movement detection mode (movement state), and when the detection unit 31 determines that there is a person in the stay detection mode (stay state). It does not have to be distinguished. Alternatively, even when the “movement state” and the “stay state” are distinguished by the detection unit 31, the transmission unit 4 applies the same control signal to the load 6 in the “movement state” and the “stay state”. May be sent to.

  In addition, the timer 7 may be provided separately for measuring the holding time T1 and for measuring the reset time T2. Further, the timer 7 may be built in the control unit 3 (microcomputer).

  In the human detection system 100, the detection unit 31 may detect the presence or absence of a person using an output image obtained by applying filter image processing to an analysis image at least in the stay detection mode. In this configuration, the filter image processing preferably uses differential filter image processing such as a Sobel filter. Thereby, the person detection system 100 can detect the presence or absence of a person with higher accuracy. Furthermore, the detection unit 31 may use an output image to which filter image processing is applied when detecting the presence or absence of a person in the movement detection mode. Filter image processing is not limited to Sobel filter image processing. The filter image processing may be, for example, a primary differential filter (eg, edge detection, line detection, etc.), a secondary differential filter (eg, Laplacian filter, etc.), image enhancement (eg, low frequency enhancement, high frequency enhancement, etc.), etc. .

  Further, the “movement amount” used for determining whether or not a person is present in the stay detection mode is not limited to a value obtained by accumulating the amount of change in the luminance value of the pixel in the stay region 23 along the time axis. For example, as the “motion amount”, a cumulative value of the magnitude of variation from the representative value (average value or median value) of the change amount of the luminance value of the pixel in the stay region 23 may be derived as the motion amount. In this case, the “evaluation value” corresponding to the movement of the person in the stay area 23 is the magnitude of variation from the representative value (average value or median value) of the amount of change in the luminance value of the pixels in the stay area 23. .

  The detection unit 31 is configured to terminate the stay detection mode and cancel the stay area 23 when a predetermined condition is satisfied even when the holding time T1 does not elapse in the stay detection mode. Also good. For example, when the amount of movement in the stay area 23 is equal to or greater than the release threshold Th3 (third threshold) greater than the detection threshold Th2, the detection unit 31 ends the stay detection mode and cancels the stay area 23. May be. Thus, if the amount of movement is equal to or greater than the release threshold Th3, the detection mode is switched from the stay detection mode to the movement detection mode, so that the detection is performed when the person 70 moves from the stay area 23 to the outside of the stay area 23. The unit 31 can detect the presence of the person 70 without losing sight.

  In addition, in the comparison between the two values such as the motion amount and the detection threshold Th2, “less than” means that one of the two values is lower than the other. However, the present invention is not limited thereto, and “less than” here may be synonymous with “below” including both the case where the two values are equal and the case where one of the two values is lower than the other. That is, whether or not to include the case where the two values are equal can be arbitrarily changed depending on the setting of the threshold value or the like, so there is no technical difference between “less than” and “less than”. Similarly, “greater than” may be synonymous with “greater than”.

(Embodiment 2)
The human detection system 100 according to the present embodiment is different from the human detection system 100 according to the first embodiment in the configuration of the detection unit 31 that makes it difficult for erroneous detection caused by sunlight or the like to occur. Hereinafter, the same configurations as those of the first embodiment are denoted by common reference numerals, and the description thereof is omitted as appropriate.

  In the present embodiment, in order to make it difficult for erroneous detection due to sunlight or the like to occur, the detection unit 31 has a time interval that satisfies a predetermined condition and a predetermined determination number of times continuously so that the amount of motion becomes equal to or greater than the detection threshold Th2. It has a function of determining that the person 70 does not exist. That is, in the present embodiment, the detection unit 31 replaces the function of resetting the motion amount every time the reset time T2 described in the first embodiment elapses, and the time interval and the number of times that the motion amount is equal to or greater than the detection threshold Th2. Based on the above, it has a function of determining that the person 70 does not exist. The “time interval that satisfies a predetermined condition” here is a time interval that satisfies a condition that is set in order to distinguish erroneous detection caused by sunlight or the like. In the present embodiment, as an example, “a time interval that satisfies a predetermined condition” is a time interval that is substantially equidistant while allowing variation within a predetermined error range.

  Hereinafter, the operation of the detection unit 31 of the present embodiment will be described with reference to FIG. FIG. 7 is a graph showing an example of a change in the amount of movement in the stay detection mode with the horizontal axis as the time axis. FIG. 7 shows the amount of movement when attention is paid to one pixel in the stay area 23.

  When the detection unit 31 is operating in the stay detection mode, the amount of movement in the stay region 23 gradually increases with time as shown in FIG. When the amount of movement in the stay area 23 is equal to or greater than the detection threshold Th2, the detection unit 31 determines that the person 70 exists in the stay area 23. As in the first embodiment, the detection unit 31 resets the movement amount every time the movement amount becomes equal to or greater than the detection threshold Th2.

  Here, for example, with respect to a sunshine spot generated by, for example, sunlight shining through a window locally illuminating a floor surface or the like, usually, a rapid change is hardly generated in the shape and moving speed on the analysis image. For this reason, when the amount of motion increases due to sunlight or the like, a rapid change is unlikely to occur in the rate of increase (the amount of increase per unit time). Therefore, as shown in FIG. 7, the amount of motion increases linearly with the passage of time, and is reset every time it reaches the detection threshold Th2. As a result, the state in which the amount of motion is equal to or greater than the detection threshold Th2 is repeated a plurality of times at time intervals that satisfy a predetermined condition. In other words, a state in which the amount of motion is equal to or greater than the detection threshold Th2 occurs regularly and repeatedly according to specific conditions.

  In the example of FIG. 7, the state where the amount of motion is equal to or greater than the detection threshold Th2 occurs four times at time intervals of “T11”, “T12”, and “T13”. Here, the “time interval satisfying the predetermined condition” is a time interval that is substantially equal with allowance for variation within a predetermined error range. As an example, assume that the predetermined error range is 20%. In this case, if the plurality of time intervals T11, T12, and T13 are 20 seconds, 24 seconds, and 16 seconds, respectively, the average value (Tav) of the plurality of time intervals T11, T12, and T13 is 20 seconds. The error range is 4 seconds (20% of 20 seconds). Then, since the plurality of time intervals T11, T12, and T13 are all within the error range (20%) with 20 seconds as a reference, they are “time intervals that satisfy a predetermined condition”. Here, if the number of determinations is four, in the example of FIG. 7, the amount of motion is equal to or greater than the detection threshold Th <b> 2 at a time interval that satisfies the predetermined condition and continuously for the predetermined number of determinations. The unit 31 determines that the person 70 does not exist.

  Thus, after the amount of movement becomes equal to or greater than the detection threshold Th2 for the fourth time, the detection unit 31 does not determine the “stay state” and switches the detection mode from the stay detection mode to the movement detection mode. That is, the detection unit 31 can end the stay detection mode and cancel the stay area 23 even if the holding time T1 has not elapsed.

  Note that the predetermined error range is appropriately set depending on the use of the human detection system 100 or the like. As an example, the predetermined error range is preferably about 15% to 25%. The number of determinations is preferably about 10 as an example. The predetermined error range and the number of determinations are preferably variable and can be arbitrarily changed by the user.

  The human detection system 100 of the present embodiment described above includes the detection unit 31 and the creation unit 32 as in the first embodiment. The detection unit 31 detects the presence or absence of a person 70 in the detection area 21 of the analysis image captured by the imaging unit 2. The creation unit 32 creates an identification area 22 surrounding the person in the detection area 21 in the analysis image.

  The detection unit 31 detects the presence / absence of a person 70 in the stay area 23 composed of the movement detection mode for detecting the presence / absence of the person 70 in the detection area 21 and the identification area 22 stagnated in the detection area 21. And a detection mode.

In the stay detection mode, the detection unit 31 is configured to determine that the person 70 exists when the amount of movement obtained by accumulating evaluation values corresponding to the movement of the person 70 in the stay area 23 is equal to or greater than the detection threshold Th2. The detection unit 31 is configured to reset the movement amount every time the movement amount becomes equal to or greater than the detection threshold Th2. In addition, the detection unit 31 is configured to determine that the person 70 does not exist when the amount of motion becomes equal to or greater than the detection threshold Th2 at a predetermined time interval and continuously for a predetermined number of determinations.

  That is, when the state in which the amount of motion is equal to or greater than the detection threshold Th2 occurs regularly and repeatedly according to specific conditions, the detection unit 31 determines that the person 70 does not exist. Therefore, even if a “sunshine spot” having a higher luminance than the other areas occurs in the stay area 23, if the state in which the amount of motion is equal to or greater than the detection threshold Th2 is regularly repeated according to a specific condition, a false detection is performed. Can be avoided. Therefore, for example, even if a sunlight spot is generated by sunlight shining from a window locally illuminating the floor surface or the like, erroneous detection due to the movement of the sunlight spot is less likely to occur. As a result, the human detection system 100 according to the present embodiment has an advantage that erroneous detection due to sunlight or the like hardly occurs.

  In short, in the human detection system 100 according to the first embodiment, the detection unit 31 reduces sensitivity to changes in the amount of motion caused by sunlight or the like in the stay detection mode by resetting the amount of motion every time the reset time T2 elapses. I am letting. In the human detection system 100 according to the second embodiment, the event that occurs regularly and repeatedly according to a specific condition is determined separately from the human 70, so that the detection unit 31 detects the amount of motion caused by sunlight or the like in the stay detection mode. It reduces the sensitivity to changes. Therefore, the human detection system 100 according to the first embodiment and the second embodiment is common in that the detection unit 31 reduces the sensitivity to a change in the amount of motion caused by sunlight or the like in the stay detection mode.

  By the way, the “time interval satisfying the predetermined condition” is not limited to a time interval that is substantially equal while allowing variation within a predetermined error range. For example, the “predetermined condition” is that a plurality of time intervals are all within a predetermined range (for example, 15 to 25 seconds), or that the time interval is decreased (or increased) at a certain rate. May be.

  As a modified example of the second embodiment, when it is determined that the amount of motion is equal to or greater than the detection threshold Th2 at a time interval satisfying a predetermined condition and continuously for a predetermined number of determinations, and the person 70 is not present, The presence or absence of a person may be detected based on the output of the infrared sensor. That is, the detection unit 31 does not switch the detection mode from the stay detection mode to the movement detection mode, but shifts to detection of the presence or absence of a person in the detection area 21 based on the output of the infrared sensor. The detection of the presence or absence of a person in the detection area 21 based on the output of the infrared sensor will be described in detail in Embodiment 3.

  Note that the human detection system 100 of the second embodiment can be applied in combination with the configuration (including the modification) of the first embodiment as appropriate.

(Embodiment 3)
As shown in FIG. 8, the human detection system 100 </ b> A of the present embodiment is different from the human detection system 100 of the first embodiment in that it further includes an infrared sensor 8, a clock unit 9, and a switching unit 34. Hereinafter, the same configurations as those of the first embodiment are denoted by common reference numerals, and the description thereof is omitted as appropriate.

  The infrared sensor 8 is a passive sensor that detects the presence of a person by detecting a change in the amount of received infrared light emitted from the person, such as a pyroelectric infrared sensor. This type of differential infrared sensor 8 detects a person who is moving (stagnation) without detecting it because the amount of received infrared light does not change. The infrared light receiving range of the infrared sensor 8 is set in accordance with the detection area 21 of the analysis image captured by the imaging unit 2.

  The clock unit 9 measures the current time. The clock unit 9 may be built in the control unit 3.

  The switching unit 34 detects the presence or absence of a person in the detection region 21 using the analysis image and the output of the infrared sensor 8 according to the time zone to which the current time measured by the clock unit 9 belongs. The state of detecting the presence / absence of a person in the area 21 is switched. As an example, the switching unit 34 outputs the output of the infrared sensor 8 for a time zone in which the western sun falls (for example, 17:00 to 18:00) and a time zone in which the morning sun falls (for example, 7:00 to 8:00). Based on this, a state for detecting the presence or absence of a person in the detection area 21 is selected. For other time zones, the switching unit 34 selects a state in which the presence / absence of a person in the detection area 21 is detected using the analysis image.

  That is, in the present embodiment, in the time zone in which there is an influence of the western sun and the morning sun, the detection unit 31 uses the analysis image that is easily affected by the western sun and the morning sun to determine whether or not there is a person in the detection area 21. Instead of detecting, the presence or absence of a person is detected based on the output of the infrared sensor 8. It is preferable that the user can arbitrarily set the time zone used in the switching unit 34, that is, the time zone in which the analysis image is used, and the time zone in which the output of the infrared sensor 8 is used. In addition, since the sunset and sunrise times vary depending on the season, the switching unit 34 may automatically change the time zone setting depending on the season by using the calendar function of the clock unit 9.

  The human detection system 100A of the present embodiment described above further includes the clock unit 9 that measures the current time and the switching unit 34. The switching unit 34 detects the presence / absence of the presence of the person 70 in the detection area 21 using the analysis image according to the time zone to which the current time measured by the clock unit 9 belongs, and the output of the infrared sensor 8. The state of detecting the presence or absence of the person 70 in the detection area 21 is switched. According to this configuration, the detection unit 31 can detect the presence / absence of a person based on the output of the infrared sensor 8 during a time zone that is affected by sunlight or the like, and thus more erroneous detection due to sunlight or the like occurs. It becomes difficult.

  By the way, as a modification of the third embodiment, the detection unit 31 detects the presence / absence of a person in the detection area 21 using an analysis image and the presence of a person in the detection area 21 based on the output of the infrared sensor 8. The presence / absence detection may be performed in parallel. In this case, for example, even if it is determined that the person 70 is present in the detection using the analysis image, the detection unit 31 is present based on the output from the infrared sensor 8 for a predetermined time (for example, 30 minutes). It is preferable to determine “absent state” when the determination of not to continue is continued. Thereby, although the person 70 is absent, it is possible to suppress an erroneous determination that the person 70 exists in the detection unit 31 for a long time due to the influence of sunlight or the like.

  In the third embodiment, the infrared sensor 8 is a component of the human detection system 100A, but the infrared sensor 8 may not be included in the component of the human detection system 100A.

  Note that the human detection system 100A of the third embodiment can be applied in combination with the configuration of the first or second embodiment (including modifications) as appropriate.

(Embodiment 4)
The human detection system 100A of the present embodiment is different from that of the third embodiment in that the switching unit 34 changes the sensitivity of the detection unit 31 in the stay detection mode according to the time zone to which the current time measured by the clock unit 9 belongs. This is different from the human detection system 100A. In the present embodiment, the infrared sensor 8 is not an essential configuration and can be omitted. Hereinafter, the same configurations as those of the third embodiment are denoted by common reference numerals, and description thereof is omitted as appropriate.

  As an example, for the time zone in which the western sun shines and the time zone in which the morning sun shines, the detection unit 31 is a mode (hereinafter, referred to as a mode in which the sensitivity to a change in the amount of movement caused by sunlight or the like is reduced in the stay detection mode. "Low sensitivity mode"). For other time zones, the detection unit 31 operates in a mode (hereinafter referred to as “low sensitivity mode”) that is highly sensitive to changes in the amount of motion caused by sunlight or the like in the stay detection mode.

  For example, as described in the first embodiment, the low sensitivity mode is a mode in which the sensitivity with respect to a change in the amount of motion caused by sunlight or the like is reduced by resetting the amount of motion every time the reset time T2 elapses. . The high sensitivity mode in this case is a stay detection mode in which the function of resetting the motion amount is invalidated every time the reset time T2 elapses. Alternatively, as described in the second embodiment, the low-sensitivity mode is a sensitivity to a change in the amount of motion caused by sunlight or the like by determining an event that occurs regularly and repeatedly according to a specific condition from the person 70. Is a mode in which The high-sensitivity mode in this case is a stay detection mode in which the function of distinguishing the event that occurs regularly and repeatedly according to a specific condition from the person 70 is disabled.

  In the human detection system 100A of the present embodiment, a clock unit 9 that counts the current time, and a switching unit that changes the sensitivity of the detection unit 31 in the stay detection mode according to the time zone to which the current time measured by the clock unit 9 belongs. 34. According to this configuration, the detection unit 31 can reduce the sensitivity in a time zone that is affected by sunlight or the like, so that erroneous detection due to sunlight or the like is less likely to occur.

  By the way, the configuration described in the third embodiment and the configuration described in the fourth embodiment can be applied in appropriate combination. In this case, the switching unit 34 detects the presence / absence of the presence of a person in the detection area 21 using the analysis image according to the time zone, and the presence of the person in the detection area 21 based on the output of the infrared sensor 8. Switches between the presence / absence detection state. Further, in the former state, the switching unit 34 changes the sensitivity of the detection unit 31 in the stay detection mode according to the time zone.

  Note that the human detection system 100A of the fourth embodiment can be applied in combination with the configurations (including modifications) of the first to third embodiments as appropriate.

  By the way, the control unit 3 described in the first to fourth embodiments can be realized by causing a computer to execute an appropriate program. That is, this program is a program for causing a computer to function as the control unit 3, and is provided by being recorded on a recording medium or provided through an electric communication line. In other words, the configuration of the control unit 3 can be read as a person detection method for detecting the person 70 using a computer.

2 Imaging unit 8 Infrared sensor 9 Clock unit 31 Detection unit 32 Creation unit 34 Switching unit 100, 100A Human detection system T1 Holding time T2 Reset time

Claims (7)

A detection unit that detects the presence or absence of a person in the detection region of the analysis image captured by the imaging unit;
The detector is
A movement detection mode for detecting the presence or absence of a person in the detection area;
A stay detection mode for detecting the presence or absence of a person in a stay area consisting of the identification area stagnant in the detection area, among the identification areas surrounding the person in the detection area ,
The detector is
In the stay detection mode,
When the amount of movement accumulated over time with the evaluation value according to the movement of the person in the stay area is greater than or equal to the detection threshold, it is determined that there is a person,
When the state where the amount of movement is less than the detection threshold continues for a holding time, it is determined that there is no person,
The human detection system is configured to reset the amount of movement every time a reset time shorter than the holding time elapses. A detection unit that detects the presence or absence of a person in the detection region of the analysis image captured by the imaging unit;
The detector is
A movement detection mode for detecting the presence or absence of a person in the detection area;
A stay detection mode for detecting the presence or absence of a person in a stay area consisting of the identification area stagnant in the detection area, among the identification areas surrounding the person in the detection area ,
The detector is
In the stay detection mode,
When the amount of movement accumulated over time with the evaluation value according to the movement of the person in the stay area is greater than or equal to the detection threshold, it is determined that there is a person,
Each time the amount of movement is equal to or greater than the detection threshold, the amount of movement is reset,
A human detection system configured to determine that no person is present when the amount of motion is equal to or greater than the detection threshold for a predetermined number of times at a time interval that satisfies a predetermined condition. The human detection system according to claim 1, wherein the evaluation value is a change amount of a luminance value of a pixel in the stay area. A clock section that counts the current time,
A state in which the presence or absence of a person in the detection area is detected using the analysis image according to a time zone to which the current time measured by the clock unit belongs, and a person in the detection area based on an output of an infrared sensor The human detection system according to claim 1, further comprising: a switching unit that switches between a state in which presence / absence of the presence / absence is detected. A clock section that counts the current time,
The switching part which changes the sensitivity of the said detection part in the said stay detection mode according to the time zone to which the present time time-measured by the said clock part belongs is further provided. The human detection system according to item.   A transmission unit that transmits a control signal according to a detection result of the detection unit;
  The human detection system according to claim 1, wherein:   A load controlled in response to the control signal;
  The human detection system according to claim 6.

Images