JP6139919B2 - Signal control device - Google Patents

Description

  The present invention relates to a signal control device that controls the light color of a traffic light that indicates whether or not a pedestrian can cross a road, and more particularly to a signal control device that secures an optimal blue time for a pedestrian to cross.

  Conventionally, traffic lights for pedestrians provided on pedestrian crossings and the like on roads are switched between a green signal and a red signal every predetermined cycle, and when switching from a green signal to a red signal, the green signal is temporarily It was controlled to flash. Here, if the time of the green light for pedestrians is short, there is a risk of hindering the crossing of the traffic weak, and conversely, even if the time of the green light is too long, there is a risk of excessively hindering the passage of the vehicle.

Therefore, various signal control devices have already been proposed in order to suppress unnecessary stopping of the vehicle while paying sufficient attention to the crossing of pedestrians.
For example, Patent Literature 1 sequentially detects the density and number of pedestrians in a detection area set near a pedestrian crossing, detects the maximum number of pedestrians and the number of people after a green light is lit, and A technique for requesting the start of blinking of a green light when the successively detected ratio falls below a predetermined threshold is disclosed.

  Further, in Patent Document 2, the number of pedestrians in a signal waiting area such as the vicinity of a pedestrian crossing is detected, and the time width of a green signal for pedestrians is stepped or stepped according to the number of pedestrians in the signal waiting area. A technique for performing control to increase or decrease continuously is disclosed. Furthermore, when there is a pedestrian on the pedestrian crossing at the end of the green light, special control for extending the lighting time of the green light is also performed.

Japanese Patent No. 4368259 Japanese Patent No. 3736463

  However, in the technique described in Patent Document 1 described above, when the number of pedestrians is small, there is no difference from the control that simply waits for the pedestrians to cross the pedestrian crossing. Moreover, when there are no pedestrians, the maximum value (0 people) continues, and the signal for pedestrians is unnecessarily maintained at a green light, which may hinder the passage of the vehicle. In addition, when pedestrians enter the end of the green light one after another (3 → 2 → 3 → 2 etc.), there is a problem that the lighting time of the green light is prolonged and the redundancy is increased. It was.

  Moreover, in the technique described in Patent Document 2, the control is based on the number of pedestrians in the signal waiting area, and is not control based on the actual number of pedestrians actually crossing the pedestrian crossing. Therefore, control according to the walking speed of the pedestrian cannot be performed, and it is necessary to perform special control for a traffic weak person whose walking speed is slow. Furthermore, there is a problem that this special control may be performed unnecessarily not only by a traffic weak person but also by the end of the green light, and the control becomes complicated.

  The present invention has been made paying attention to the above-mentioned problems of the prior art, and by tracking the actual crossing situation of the pedestrian and controlling the light color of the traffic light for the pedestrian, It is an object of the present invention to provide a signal control device that can secure an optimum green time for crossing a person and can also realize smooth traffic of a vehicle.

The gist of the present invention for achieving the object described above resides in the inventions of the following items.
[1] In the signal control device (10) for controlling the color of the traffic light (11) instructing whether or not the pedestrian can cross the road,
Detection means (43) capable of detecting a specific individual or group as a tracking target among pedestrians that start crossing after the traffic light (11) switches from a red signal to a green signal;
Tracking means (44) capable of tracking a situation in which the tracking target moves from a reference position on the crossing start side to a designated position on the crossing end side on the road;
Determination means (45) capable of determining whether or not the tracking target has reached the designated position,
In the detection means (43), the traffic light (11) is switched from a red signal to a blue signal based on images sequentially input from the photographing unit (20) capable of photographing at least a range from the reference position to the designated position. Detecting pedestrians or groups of pedestrians that have passed through the reference position during a predetermined time from the time as a tracking target,
A signal control device (10), characterized in that, based on a result determined by the determination means (45) that the tracking target has reached the designated position, a signal color of the traffic light (11) is switched.

  [2] When the pedestrian who starts crossing after the traffic signal (11) switches from the red signal to the blue signal is not detected by the detection means (43), the traffic signal (11) switches from the red signal to the green signal. The signal control device (10) according to [1], characterized in that the signal color of the traffic light (11) is switched based on a lapse of a predetermined time from the time when the signal is transmitted.

  [3] The signal control device (10) according to [2], wherein the predetermined time can be changed in accordance with predetermined parameters according to various parameters.

[4] The tracking means (44) is based on the time until the tracking target detected by the detection means (43) passes the specified position based on the images sequentially input from the imaging unit (20). In addition, when the individual or group of pedestrians to be tracked is separated or combined, tracking is performed while recognizing as a new tracking target that inherits the attribute of the tracking target [1], [ The signal control device (10) according to [2] or [3] .

[5] The determination unit (45) is based on the images sequentially input from the photographing unit (20), and the tracking unit (44) detects the tracking target detected by the detection unit (43). The determination result is output when all the tracking targets including the new tracking target reach the designated position when there is a recognized new tracking target, [4] Signal control device (10).

[6] Separately includes a control means (50) for controlling switching of the lamp color of the traffic light (11),
The signal control device (10 ) according to [5] , wherein the control means (50) executes control for blinking a green light for pedestrians based on a determination result from the determination means (45). ).

[7] The signal control according to [6] , wherein the control means (50) executes control to switch to a red signal for pedestrians based on the determination result from the determination means (45). Device (10).

Next, the operation based on the above solution will be described.
According to the signal control device (10) described in [1], the detection means (43) detects a specific individual or group of pedestrians starting to cross as a tracking target. When crossing a pedestrian crossing, pedestrians may act alone (defined as “individual”), or may act as a group of multiple people (defined as “group”). Accordingly, the detection means (43) selects a specific individual or group as an object to be tracked.
The detection means (43) is a traffic light (11) based on images sequentially input from a photographing unit (20) capable of photographing at least a range from a reference position on the crossing start side to a designated position on the crossing end side on the road. ) Detects a pedestrian or a group of pedestrians who have passed the reference position during a predetermined period from when the red signal is switched to the green signal.
In this detection, the difference between the current frame that is sequentially input and the previous frame is binarized (threshold processing) in image processing, and is further labeled from the background in the image. Pedestrians (individuals or groups) can be detected as tracking targets. Details of such image processing will be described later.
In selecting a tracking target, the pedestrian who has passed the reference position between the time when the traffic light (11) is switched to a green light and a predetermined time has elapsed, is the pedestrian who starts to cross first. Will be targeted. In general, it is considered that a weak traffic person starts to cross immediately after changing to a green light. Therefore, the possibility of including a weak traffic person in the tracking target can be increased by such a time selection.

  Next, the tracking unit (44) tracks the situation in which the tracked object moves from the reference position on the crossing start side to the specified position on the crossing end side. Here, the tracking means (44) grasps the movement of the individual or group to be tracked in real time. Then, the determination means (45) determines whether or not the tracking target has reached the designated position.

  If it is determined by the determination means (45) that the tracking target has reached the specified position, the light color of the traffic light (11) is switched based on the determination result. The switching of the lamp color of the traffic light (11) here does not have to be performed by the determination means (45) or the signal control device (10) itself, but may be performed by another control device. The trigger for switching the lamp color includes not only the output of a red signal request signal for switching from a blue signal to a red signal, but also an output of a blue blink request signal for starting blinking of the blue signal.

  As described above, since all the tracking targets have been crossed, the lamp color is changed. Therefore, by including vulnerable persons such as elderly people, disabled persons, and children in the tracking target, it is safer for the traffic weak. Signal control that can be passed with peace of mind can be realized. In addition, even if a pedestrian crossing in the last half of the green light appears, by not including such a rushed pedestrian in the tracking target, the switching of the light color is not affected, and the switching can be performed. There will be no significant delay.

  According to the signal control device (10) described in [2], when the pedestrian that starts crossing after the traffic light (11) is switched from the red signal to the blue signal is not detected by the detection means (43). Is an opportunity to switch the color of the traffic light (11) based on the passage of a predetermined time from when the traffic light (11) switches from a red signal to a blue signal.

  The predetermined time here may be set to a minimum necessary short time so as not to obstruct the passage of the vehicle as much as possible. Thereby, when there is no pedestrian, the lighting time of the green light for pedestrians becomes the shortest, and the vehicle can be distributed efficiently.

  However, as described in [3] above, the predetermined time may be changed according to various parameters for the predetermined time. Here, the specified value is a time set as a reference that is initially set, and is set to a minimum necessary short time as described above. Depending on various parameters such as the day of the week, time, and season, for example, walking On weekends when the number of people is expected to be large, it can be changed flexibly such as setting a predetermined time longer than the specified value.

Further, according to the signal control device (10) described in [4] , the tracking unit (44) is controlled by the detection unit (43) based on images sequentially input from the photographing unit (20). If the pedestrians or groups that are the tracking target are separated or combined before the detected tracking target passes through the specified position, the tracking target attribute is taken as a new tracking target. Track while recognizing.
Thus, according to various movements of the tracking target, the tracking target detected first can be accurately tracked without omission.

Further, according to the signal control device (10) described in [5] , the determination unit (45) is based on the images sequentially input from the photographing unit (20) by the detection unit (43). If there is a new tracking target recognized by the tracking means (44) in addition to the detected tracking target, the determination is made when all the tracking targets including the new tracking target have reached the designated position. Output the result.
Thereby, it is possible to perform control for switching the lamp color of the traffic light (11) based on the determination result by outputting the determination result when all of the first detected tracking objects have successfully passed. Become.

Here, the control for switching the color of the traffic light (11) does not have to be performed by the signal control device (10) itself as described above, but the control means (50) as a part of the configuration of the signal control device (10). ).
According to the signal control device (10) described in [6] , the control means (50) performs control for blinking a pedestrian green signal based on the determination result from the determination means (45). To do.
Alternatively, as described in [7] above, in addition to the blinking of the blue signal, control for switching the blue signal to a red signal may also be executed.

  According to the signal control device of the present invention, it is possible to ensure the optimal blue time for pedestrian crossing by tracking the actual crossing situation of the pedestrian and controlling the light color of the traffic light for pedestrians. In addition, smooth traffic of the vehicle can be realized.

It is a block diagram which shows the structural example of the signal control apparatus which concerns on embodiment of this invention. It is explanatory drawing which shows the coefficient table used by control of the signal control apparatus which concerns on embodiment of this invention. It is a series of screen transition diagrams showing an example of image processing by the signal control device according to the embodiment of the present invention. It is an image figure which shows an example of the image processing by the signal control apparatus which concerns on embodiment of this invention. It is a flowchart which shows the processing content of the signal control apparatus which concerns on embodiment of this invention. It is a flowchart which shows the modification of the processing content of the signal control apparatus which concerns on embodiment of this invention. It is a flowchart which shows the area | region detection process of the signal control apparatus which concerns on embodiment of this invention. It is a flowchart which shows the tracking determination process of the signal control apparatus which concerns on embodiment of this invention.

DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiments representing the present invention will be described below with reference to the drawings.
The signal control device 10 according to the present embodiment controls the color of the traffic light 11 that instructs whether or not a pedestrian can cross the road. Hereinafter, the case where it applies to the traffic signal 11 only for the pedestrian in a pedestrian crossing is demonstrated.

  FIG. 1 is a block diagram illustrating a configuration example of a signal control device 10 according to the present embodiment. The signal control device 10 includes an imaging unit 20, a setting unit 30, a processing unit 40, and a signal control device 50. And. Here, the signal control device 50 is a “control unit” that comprehensively controls switching of the lamp color of the traffic light 11, and originally constitutes another system independently of the signal control device 10, It can also be understood as a part of the signal control device 10. Since this signal control device 50 is a known system, a detailed description thereof will be omitted.

  The imaging unit 20 is mainly configured by, for example, an infrared video camera or the like in addition to a general video camera equipped with a CCD image sensor, a CMOS sensor, or the like. The imaging unit 20 can cover the pedestrian crossing as a shooting area from the reference position on the crossing start side to the designated position on the crossing end side of the pedestrian crossing in an appropriate place such as a post where the traffic light 11 is installed. Are arranged as follows.

  Images (videos) photographed by the photographing unit 20 are sequentially input and stored in the processing unit 40, and are subjected to image processing by a human detection unit 43 and the like described later. Note that the video camera constituting the photographing unit 20 is generally one that shoots 30 frames per second, for example, but may be one having a variable frame rate that can arbitrarily change the number of frames per second. good. In addition, it is preferable to control the video camera to start at a predetermined timing, for example, when the traffic light 11 is switched to a green signal, a point just before that, or the like.

  In the present signal control device 10, when all the pedestrians detected as tracking targets have crossed the pedestrian crossing by the image processing of the image photographed by the photographing unit 20, an opportunity to switch the light color of the traffic light 11, that is, Via the signal control device 50, the traffic light 11 is switched to blue flashing for notice to change from a green signal to a red signal. However, if a pedestrian that starts crossing after the traffic light 11 is switched from the red signal to the green signal is not detected, the traffic light 11 is based on the passage of a predetermined time from when the traffic light 11 is switched from the red signal to the green signal. It is an opportunity to switch the lamp color from green to red again.

  The setting unit 30 can change the above-described predetermined time from predetermined values according to various parameters. For example, the setting unit 30 is connected to the processing unit 40 in the form of an external input unit. Here, as various parameters, “day of the week”, “time”, and “season” shown in the coefficient table of FIG. 2 are defined as coefficient items to be multiplied by the specified value in the present embodiment. Note that the setting unit 30 may appropriately change a detection region, a reference position, a designated position, and the like in image processing to be described later.

  The type of each coefficient item or a specific coefficient for each coefficient item can be appropriately changed by a predetermined input operation by the setting unit 30. For example, if “day of the week” is Sunday, the coefficient 1.1 is multiplied by the specified value, if “time” is in the midnight range, the coefficient 0.5 is further multiplied, and if “season” is the year-end and New Year period, the coefficient is multiplied. 2.0 is further multiplied to calculate a specific predetermined time. The prescribed value is preferably set to a minimum value (required in short time) as much as possible so as not to hinder the passage of the vehicle when there are no pedestrians.

  The parameters 31 set by the setting unit 30 are stored in the processing unit 40, and the date, day of the week, and season are measured in real time by a clock 41 in the processing unit 40. Based on these data, the processing unit 40 is measured. The predetermined time is calculated each time by the coefficient calculation unit 42. And when there is no pedestrian crossing a pedestrian crossing, when the calculated predetermined time passes, it is controlled via the signal control apparatus 50 that the traffic light 11 is switched from a green light to blinking blue.

  The processing unit 40 is temporarily necessary to execute a CPU (central processing unit) that performs a central function of various controls, a ROM that stores a program executed by the CPU and various fixed data, and a program. It consists of a RAM for storing data and a microcomputer whose main part is an interface. The processing unit 40 includes a human detection unit 43, a human tracking unit 44, and a determination unit 45 as representative functions. These are functions for detecting and tracking pedestrians by image processing techniques.

The human detection unit 43 is a “detection unit” that can detect a specific individual or group of pedestrians starting to cross a pedestrian crossing after the traffic light 11 switches from a red signal to a green signal as a tracking target.
The person tracking unit 44 is a “tracking unit” capable of tracking a situation in which the tracking target moves from a reference position on the crossing start side to a designated position on the crossing end side on the pedestrian crossing (road).
Here, as shown in the screen transition diagram of FIG. 3, the reference position on the crossing start side is a line arbitrarily defined on the crossing start side on the pedestrian crossing in the image inputted from the photographing unit 20 and performing image processing. Corresponds. The designated position on the crossing end side also corresponds to a line arbitrarily defined on the crossing end side on the crosswalk in the image.

The determination unit 45 is “determination means” capable of determining whether or not the tracking target has reached the designated position.
Based on the result of the determination unit 45 determining that the tracking target has reached the designated position, the light color of the traffic light 11 is switched. That is, when the determination unit 45 determines that the tracking target has reached the designated position, the determination unit 45 outputs a blue blinking request signal for starting the blinking of the blue light of the traffic light 11 to the signal control device 50. Is set to be. Here, instead of the blue blinking request signal, a red signal request signal for switching from the blue signal to the red signal may be output.

  Specifically, based on the images sequentially input from the photographing unit 20, the human detection unit 43 has passed through the reference position after a predetermined time has elapsed since the traffic light 11 switched from a red signal to a blue signal. Individuals or groups of pedestrians are detected as tracking targets. Specific image processing here will be described later.

  In addition, the person tracking unit 44 is based on the images sequentially input from the photographing unit 20 until the tracking target detected by the person detecting unit 43 passes through the designated position. When an individual or group of pedestrians are separated or combined, the tracking is performed while recognizing the attribute of the tracking target as a new tracking target. Specific image processing here will also be described later.

  Based on the images sequentially input from the imaging unit 20, the determination unit 45 includes a new tracking target recognized by the human tracking unit 44 in addition to the tracking target detected by the human detection unit 43. In some cases, the determination result is set to be output when all the tracking targets including the new tracking target reach the designated position. Specific image processing here will also be described later.

  Further, in FIG. 1, the traffic light 11 is a red signal when the vehicle traffic signal (not shown) is a green signal under the control of the signal control device 50, and the vehicle traffic signal becomes a yellow signal when a predetermined time elapses. At the same time, the traffic light 11 switches from a red signal to a blue signal. Then, as described above, when all the tracking targets reach the designated position, the light blinks in blue and then switches to a red signal, and at the same time, the vehicle traffic signal switches to a green signal.

Next, the operation of the signal control apparatus 10 according to the present embodiment will be described focusing on image processing.
FIG. 5 is a flowchart showing the processing contents of the processing unit 40. This routine is started when the traffic light 11 is switched from a red signal to a blue signal (step S100). Note that switching to the green light is determined from information obtained from the signal control device 50. First, in step S101, the blue blinking request signal is turned off, and in the next step S102, the blue blinking flag is turned off.

  Subsequently, when the process proceeds to step S103, an image (video) input from the photographing unit 20 is captured and set as a target of the current image processing. In step S104, the difference between the image captured in step S103 and the image before the designated frame is obtained. Note that the image that takes the difference is not the close frame, for example, the previous frame, but the image before the specified frame, because if the walking speed of the pedestrian is slow, the moving area is difficult to appear in the difference between the close frames. is there.

  Next, in step S105, the difference image generated in step S104 is binarized into, for example, luminance 0 and luminance 2 (8 bits) with a predetermined threshold. By such binarization (threshold processing), the brightness of each pixel in the image can be alternatively converted into, for example, black and white with a constant reference value. Here, in the binarized image, a labeling process is performed in order to detect a pedestrian (an individual or a group) from the background as a tracking target.

  The labeling process is a process of adding the same label (number) as an attribute to each pixel connected to each other in the binarized image, and classifying a plurality of areas as a group and extracting a specific area Can do. Therefore, in step S106, the binary image generated in step S105 is labeled to obtain the coordinates and area of the area corresponding to the pedestrian (or individual or group). As a result, pedestrians (individuals or groups) can be detected from the background as tracking targets in the image, and the tracking attributes are added with group attributes A, B, C... Is done. There are various types of specific labeling methods, but since they are already known, detailed description thereof is omitted.

  Next, in step S107, an area detection process is executed, and in step S108, a tracking determination process is executed. Details of these processes will be described later. Then, it progresses to step S109 and it is determined whether there exists the tracking target (for example, division A) crossing the pedestrian crossing. Here, if there is no tracking target (Yes in step S109), the process proceeds to step S110 to determine whether or not a predetermined time has elapsed. On the other hand, if there is a tracking target (No in step S109), the process proceeds to step S113 to determine whether or not the tracking target has been crossed.

  If it is determined in step S110 that the predetermined time has elapsed (Yes in step S110), the blue blinking flag is turned on in step S111, and the blue blinking request signal is turned on in the subsequent step S112. Thereafter, the process proceeds to step S116. On the other hand, if the predetermined time has not elapsed (No in step S110), the process jumps to step S116 as it is. Here, the predetermined time is the above-described predetermined time when there is no pedestrian crossing the pedestrian crossing.

  If it is determined in step S113 that the crossing of the tracking target (for example, section A) has been completed (Yes in step S113), the blue blinking flag is turned on in step S114, and the blue blinking request signal is transmitted in the subsequent step S115. Turn it on. On the other hand, if the traversing of the tracking target (for example, section A) is not completed and still exists between the reference position and the designated position (No in step S113), the process jumps to step S116 as it is.

  In step S116, it is determined whether or not the traffic light 11 is a green light. If the traffic light 11 is still a green signal (Yes in step S116), the process returns to the above-described step S103, and the subsequent processing is repeated. On the other hand, if the traffic light 11 is not a blue signal (No in step S116), the blue blinking request signal is turned ON in step S117, the processing of this routine is once ended (step S118), and the standby state of step S100 is entered again.

  According to this routine, after the traffic light 11 is switched to a green light, for example, a pedestrian at the beginning of a transition detected in x seconds is tracked as a tracking target (section A). Then, when all of the tracking objects (section A) have been crossed, a blue blinking request signal is transmitted. When no tracking target (section A) is generated, a blue blinking request signal is transmitted after a predetermined time has elapsed. As a result, it is possible to secure the optimal green time for crossing pedestrians and to realize smooth traffic of the vehicle.

  FIG. 6 is a flowchart illustrating a modification of the processing content of the processing unit 40. This routine is also started when the traffic light 11 is switched from a red signal to a blue signal, similarly to the processing of FIG. 5 described above (step S200). First, in step S201, the blue blink request signal is turned OFF, and in the next step S202, the red signal request signal is turned OFF. In step S203, the blue blinking flag is turned off, and in the next step S204, the red flag is turned off.

  Next, in step S205, the image (video) input from the photographing unit 20 is captured and set as a target of the current image processing. In step S206, the difference between the image captured in step S205 and the image before the designated frame is obtained. Here, the image for which the difference is taken, for example, the image before the designated frame without being a close frame such as the previous one is the same as the case of the step S104, as in the case of the step S104. This is because the movement area is difficult to appear in the difference.

  In step S207, the difference image generated in step S206 is binarized to, for example, luminance 0 and luminance 2 (8 bits) with a predetermined threshold. In step S208, the binary image generated in step S207 is labeled. Here, as in the case of step S106, a group attribute of sections A, B, C,... Is added to the tracking target detected from the background in the image every predetermined second, and coordinates and Find the area.

  Next, in step S209, area detection processing is executed, and in step S210, tracking determination processing is executed. Details of these processes will be described later. Then, it progresses to step S211 and it is determined whether the blue blinking flag is OFF. Here, if the blue blinking flag is OFF (Yes in step S211), the process proceeds to step S212, and it is determined whether or not there is a tracking target (section A) crossing the pedestrian crossing. On the other hand, if the blue blink flag is not OFF (No in step S211), the process proceeds to step S219, and it is determined whether or not the tracking target (section B) has been crossed.

  If it is determined in step S212 that there is no tracking target (section A) (Yes in step S212), the process proceeds to step S213 to determine whether a predetermined time has elapsed. On the other hand, if there is a tracking target (section A) (No in step S212), the process proceeds to step S216 to determine whether or not the crossing of the tracking target (section A) has ended.

  If it is determined in step S213 that the predetermined time has elapsed (Yes in step S213), the blue blinking flag is turned on in step S214, and the blue blinking request signal is turned on in the next step S215. Thereafter, the process proceeds to step S223. On the other hand, if the predetermined time has not elapsed (No in step S213), the process jumps to step S223 as it is. Here, the predetermined time is the above-described predetermined time when there is no pedestrian crossing the pedestrian crossing.

  If it is determined in step S216 that the tracking target (for example, section A) has been crossed (Yes in step S216), the blue blinking flag is turned on in step S217, and the blue blinking request signal is transmitted in the next step S218. Set to ON. On the other hand, if the traversing of the tracking target (for example, section A) is not completed and still exists between the reference position and the designated position (No in step S216), the process jumps to step S223 as it is.

  Further, when the process proceeds from step S211 to step S219, it is determined whether or not the traversing of the tracking target (section B) is completed. If the traversing of the tracking target (section B) has been completed (Yes in step S219), the red flag is turned on in step S220, and the red signal request signal is turned on in the next step S221. In step S222, the blue blinking flag is turned ON, and the process proceeds to step S223. On the other hand, if the traversing of the tracking target (section B) has not ended (No in step S219), the process jumps to step S222 to turn on the blue blinking flag, and the process proceeds to step S223.

  Then, when it progresses to step S223, it will be determined whether the traffic signal 11 is a blue signal or a blue blinking signal. Here, if the traffic light 11 is a blue signal or a blue blinking signal (Yes in step S223), the process returns to step S205 described above, and the subsequent processing is repeated. On the other hand, if the traffic light 11 is not a blue signal or a blue blinking signal (No in step S223), the blue blinking request signal is turned on in step S224, and the red signal request signal is turned on in the next step S225. Is temporarily terminated (step S226), and the standby state of step S200 is entered again.

  According to this routine, after the traffic light 11 is switched to a green light, for example, a pedestrian at the beginning of a transition detected in x seconds is tracked as a tracking target (section A). Further, the person who starts the next transition detected in y seconds after elapse of x seconds is tracked as a tracking target (section B). Then, when all of the tracking objects (section A) have been crossed, a blue blinking request signal is transmitted. When the tracking target (section A) does not occur, a blue blinking request signal is transmitted after a predetermined time has elapsed. In addition, after the blue blink request signal is transmitted, the red signal request signal is transmitted when all the tracking targets (section B) have passed. When the tracking target (section B) does not occur, a red signal request signal is transmitted after a predetermined time has elapsed. Thus, based on the determination result that the tracking target has finished crossing, not only the blinking of the green light may be started, but also a trigger for switching to the red light.

  FIG. 7 is a flowchart showing a region detection process performed by the human detection unit 43 or the human tracking unit 44 of the processing unit 40. This routine corresponds to step S107 in the processing content described in FIG. 5 and step S209 in the processing content described in FIG. This routine is started as a part of the above-described image labeling process (step S300). First, in step S301, a variable i is initialized with zero.

  Next, in step S302, it is determined whether or not x seconds have elapsed since the start of processing at the time when the traffic light 11 was switched from a red signal to a blue signal. If x seconds have not yet elapsed (Yes in step S302), the current section (variable) is set to (section A) in step S303. On the other hand, if x seconds have already elapsed (No in step S302), it is determined in step S304 whether y seconds have elapsed.

  If y seconds have not yet elapsed (Yes in step S304), the process proceeds to step S305, and the current section (variable) is set to (section B). On the other hand, if y seconds have already elapsed (No in step S304), the process proceeds to step S306, and the current section (variable) is set to (section C).

  In such a labeling process, as will be described later, as shown in the screen transition diagram of FIG. 3, the area to which the group attribute (section) is added every predetermined second from the start of the process is limited to one at a time. There is no. That is, when the areas detected as pedestrians in the actual image are separated from each other, branch numbers (A1, A2...) Are further added in each section (A, B. They are identified as individuals or groups.

  Thereafter, when the process proceeds to step S307, it is determined whether or not the variable i has reached the number of labels. Here, if the variable i has reached the number of labels (No in step S307), the processing of this routine is terminated as it is (step S331). On the other hand, if the variable i has not reached the number of labels (Yes in step S307), the process proceeds to step S308, and the tracking area closest to the label i is searched. Note that the distance for determining whether or not it is “closest” is, for example, the distance between the midpoints of the respective regions.

  Next, in step S309, it is determined whether or not the distance between the tracked area searched in step S308 and the label i is closer than a preset tracking distance (parameter). Here, if the distance is short (Yes in step S309), the process proceeds to step S310, and it is determined whether or not there is an area currently associated with the area being tracked searched in step S308. On the other hand, if the distance is long or there is no area being tracked (No in step S309), the process proceeds to step S323 described later.

  If it is determined in step S310 that there is an area currently associated with the area being tracked searched in step S308 (Yes in step S310), the process proceeds to step S311 where the area of label i is the outer frame of the area detection range. It is determined whether or not it overlaps. On the other hand, if there is no area associated this time (No in step S310), the process proceeds to step S322, and after label i is associated with the area being tracked searched in step S308, 1 is added to variable i in step S330. Then, the process returns to step S307, and the subsequent processing is repeated.

  If it is determined in step S311 that the area of label i overlaps the outer frame of the area detection range (Yes in step S311), the process proceeds to step S312 and the label i is being tracked as being associated as a new detection area. Add to the area. In the subsequent step S313, after the area segment added in step S312 is set as the current segment (set in steps S302 to S306), 1 is added to the variable i in step S330, and the process returns to step S307. Repeat the subsequent processing.

  On the other hand, if it is determined in step S311 that the area of label i does not overlap the outer frame of the area detection range (No in step S311), the process proceeds to step S314, and the tracking area and label searched in step S308 It is determined whether or not the distance to i is closer than a preset combination determination distance (parameter).

  Here, if it is closer than the combination determination distance (parameter) (Yes in Step S314), in Step S315, the region associated with the region being tracked searched in Step S308 and the label i are combined and combined. Update the association to be the largest rectangle. Thereafter, 1 is added to the variable i in step S330, and the process returns to step S307 to repeat the subsequent processing.

  On the other hand, if the distance is longer than the combination determination distance (parameter) (No in step S314), the process proceeds to step S316, and it is determined whether or not the label i is farther than the associated area from the tracked area searched in step S308. . If it is far (Yes in step S316), in step S317, the label i is added as a newly detected area to the area being tracked, and in step S318, the area segment added in step S317 is set to the current segment. Classification (set in steps S302 to S306).

As described above, when the tracked area (tracking target) is separated, the attributes of the same section (for example, A) are inherited as shown in the image diagram of FIG.
Thereafter, 1 is added to the variable i in step S330, and the process returns to step S307 to repeat the subsequent processing.

  If it is determined in step S316 that the label i is closer to the associated area from the tracking area searched in step S308 (No in step S316), the process proceeds to step S319, and the tracking searched in step S308. The area associated with the middle area is added to the area being tracked as being associated as a new detection area. In subsequent step S320, the label i is associated with the area being tracked searched in step S308. In step S321, the new area added in step S319 is the same as the area in step S320. Thereafter, 1 is added to the variable i in step S330, and the process returns to step S307 to repeat the subsequent processing.

  Next, returning to step S309, when the determination result in step S309 is No, the process proceeds to step S323, where it is determined whether or not the area of label i overlaps the area being tracked searched in step S308. To do. Here, if they overlap (Yes in step S323), the process proceeds to step S324, and after associating the label i with the tracked area searched in step S308, 1 is added to the variable i in step S330. The process returns to step S307, and the subsequent processing is repeated. On the other hand, if they do not overlap (No in step S323), the process proceeds to step S325, and it is determined whether or not the area of label i overlaps the outer frame of the area detection range.

  Here, if the area of the label i overlaps with the outer frame of the area detection range (Yes in step S325), in step S326, the label i is added as a newly detected area to the area being tracked. Further, in step S327, the new area added in step S326 is set as the current category (set in steps S302 to S306), and 1 is added to the variable i in step S330, and then the step S307 is performed. And the subsequent processing is repeated.

  On the other hand, if the area of the label i does not overlap the outer frame of the area detection range (No in step S325), in step S328, the label i is added as a newly detected area to the area being tracked. Further, in step S329, the new area added in step S328 is set as the current area (set in steps S302 to S306), 1 is added to the variable i in step S330, and then step S307 is executed. And the subsequent processing is repeated.

  FIG. 8 is a flowchart showing the tracking determination process by the human tracking unit 44 through the determination unit 45 of the processing unit 40. This routine corresponds to step S108 in the processing content described in FIG. 5 and step S210 in the processing content described in FIG. This routine is also started as a part of the above-described image labeling process (step S400). First, the variable i is initialized to 0 in step S401.

  Next, in step S402, it is determined whether or not the variable i is larger than the number of areas being tracked. If there is more (No in step S402), the processing of this routine is terminated as it is (step S415). On the other hand, if it is less (Yes in Step S402), the process proceeds to Step S403, and it is determined whether or not there is an associated area in the area i being tracked.

  Here, if it is a new detection area (Yes in step S403), in step S404, the number of missing areas (variables) of the area i being tracked is initialized to 0, and the process proceeds to step S405. On the other hand, if there is no associated area (No in step S403), in step S407, a tracking area closest to the tracking area i is searched, and then the process proceeds to step S408. Here, the distance for determining whether or not it is “closest” is, for example, the distance between the midpoints of the respective regions, as in the case of step S308.

  In step S405, it is determined whether the area i being tracked exceeds a specified position (parameter). Here, if the specified position is exceeded (Yes in step S405), the process proceeds to step S406, where the section of the area i being tracked is excluded from tracking (may be deleted). Thereafter, 1 is added to the variable i in step S414, and the process returns to step S402 to repeat the subsequent processing. On the other hand, if the tracked area i has not yet exceeded the specified position (No in step S405), the process proceeds to step S414, and 1 is added to the variable i. Then, the process returns to step S402, and the subsequent processing is performed. repeat.

  Next, returning to the step S407, searching for the tracking area closest to the tracking area i and then proceeding to the step S408, the distance to the area searched in the step S407 is set in advance. It is determined whether it is within the distance (parameter). If the distance is within the coupling distance (Yes in step S408), in step S409, the area i being tracked and the area searched in step S407 are unified on the dominant side, and in the following step S410, the track is being tracked. The area i and the area searched in step S407 are combined as one area.

In the step S409, “unification on the dominant side” means that, as shown in the image diagram of FIG. 4B, among the different regions (tracking targets), the earlier classification (A>B>C> It means that it will be unified. Further, by combining them as one area in step S410, as a result, the number of tracking objects that are areas being tracked is reduced by one.
Thereafter, 1 is added to the variable i in step S414, and the process returns to step S402 to repeat the subsequent processing.

  If the distance between the area searched in step S407 and the area i being tracked is not within the combined distance (No in step S408), the number of missing areas (variables) in the area i being tracked is determined in advance in step S411. It is determined whether the value is equal to or less than the reference value (parameter). If it is below the reference value (Yes in step S411), 1 is added to the number of missing areas (variables) of the area i being tracked in the following step S412. Thereafter, 1 is added to the variable i in step S414, and the process returns to step S402 to repeat the subsequent processing.

  On the other hand, if the missing number (variable) of the area i being tracked is not less than or equal to the reference value (No in step S411), the area i being tracked is deleted (recognized as missing) in step S413. Thereafter, 1 is added to the variable i in step S414, and the process returns to step S402 to repeat the subsequent processing.

  FIG. 3 is an image diagram showing a situation in which the tracking target moves in the above-described image processing on the display display of the processing unit 40. As shown in FIGS. 3A to 3H, the tracking target is tracked while being detected every predetermined second from the time when the traffic light 11 is switched to the green light. Each screen is obtained by processing images sequentially input from the photographing unit 20, and the upper side of the pedestrian crossing as the background is the reference position and the lower side is the designated position. These specific positions can be appropriately changed by the setting unit 30 as described above.

  In the example shown in FIG. 3, FIG. 3A is the time when the signal is switched to a green light, and at this time, the tracking target has not yet appeared on the pedestrian crossing. After that, two (A1, A2) of the tracking target (section A) appear by the elapse of x seconds shown in FIGS. (B) to (c), and the following (d) to (e) are shown. Three new tracking targets (section B) (B1, B2, B3) appear before y seconds elapse. It should be noted that the tracking target is detected until the end of the new area (section C) after y seconds shown in FIG.

  Then, as shown in FIG. 3G, when all the first tracking objects (section A) reach the designated position, a blue blinking request signal is output to the signal control device 50 at that time. Further, as shown in FIG. 3 (h), when all the following tracking objects (section B) have reached the designated position, a red signal request signal is output to the signal control device 50. The output of the red signal request signal here corresponds to step S221 in the flowchart of FIG. Thus, a red signal request signal can be output to the signal control device 50 under the control of the determination unit 45.

  According to the signal control apparatus 10 according to the present embodiment (including the modification) described above, the human detection unit 43 detects a specific individual or group among pedestrians that start crossing as a tracking target. When a pedestrian crosses a pedestrian crossing, the pedestrian may act alone or may act as a group of a plurality of people. Depending on these forms, the human detection unit 43 may be a specific individual or group. Are selected as a tracking target.

  Next, the person tracking unit 44 tracks the situation in which the tracking target moves from the reference position on the crossing start side to the designated position on the crossing end side. Here, the person tracking unit 44 grasps the movement of the individual or group to be tracked in real time. Then, the determination unit 45 determines whether or not the tracking target has reached the specified position.

  When the determination unit 45 determines that the tracking target has reached the specified position, the determination unit 45 sets an opportunity to switch the light color of the traffic light 11 based on the determination result. The switching of the lamp color of the traffic light 11 does not need to be performed by the determination unit 45 or the signal control device 10 itself, but may be performed by another signal control device 50. As described above, the trigger for switching the lamp color includes not only the output of the red signal request signal for changing the blue signal to the red signal but also the output of the blue blink request signal for starting the blinking of the blue signal.

  In this way, since the lighting color is changed when all of the tracking targets have been crossed, if the tracking target includes traffic weak persons such as elderly people, disabled persons, and children, the traffic weak persons are safer. In addition, signal control that can be passed with peace of mind can be realized. In addition, even if a pedestrian crossing in the last half of the green light appears, by not including such a rushed pedestrian in the tracking target, the switching of the light color is not affected, and the switching can be performed. There will be no significant delay.

  In addition, when the pedestrian who starts crossing after the traffic light 11 is switched from the red signal to the green signal is not detected by the human detection unit 43, a predetermined time from when the traffic light 11 is switched from the red signal to the green signal is detected. Based on the progress, it is assumed that the lamp color of the traffic light 11 is switched. As described above, the predetermined time here may be set to a minimum necessary short time so as not to obstruct the passage of the vehicle as much as possible. Thereby, when there is no pedestrian, the lighting time of the green light for pedestrians becomes the shortest, and the vehicle can be distributed efficiently.

  For the predetermined time, the specified value determined in advance by the setting unit 30 can be changed according to various parameters. Here, the specified value is a time set as an initial reference, and is set to the minimum necessary time as described above. However, as described with reference to FIG. 2, the day of the week, time, season, etc. Depending on various parameters, it can be changed flexibly, for example, at a weekend when the number of pedestrians is expected to be large, such as setting a predetermined time longer than a prescribed value.

  In this embodiment, in order to realize the control by each means described above, image processing is performed in which the situation of the pedestrian crossing can be confirmed by only one photographing unit 20. That is, the human detection unit 43 has passed a predetermined second from the time when the traffic light 11 is switched from a red signal to a blue signal based on images sequentially input from the photographing unit 20 capable of photographing at least a range from a reference position to a designated position. The number or group of pedestrians who have passed the reference position until the time is detected as a tracking target.

  In selecting the tracking target, the pedestrian who has passed the reference position between the time when the traffic light 11 is switched to the green light and the predetermined time has elapsed, the main point is the pedestrian who starts to cross first. Will do. In general, it is considered that a weak traffic person starts to cross immediately after changing to a green light. Therefore, the possibility of including a weak traffic person in the tracking target can be increased by such a time selection.

  In addition, the person tracking unit 44, based on the images sequentially input from the imaging unit 20, until the tracking target detected by the human detection unit 43 passes the designated position, the pedestrian that is the tracking target. When the individual or group is separated or combined, the tracking is performed while recognizing the tracking target attribute as a new tracking target. Thus, according to various movements of the tracking target, the tracking target detected first can be accurately tracked without omission.

  Further, the determination unit 45, when there is a new tracking target recognized by the human tracking unit 44 in addition to the tracking target detected by the human detection unit 43 based on the images sequentially input from the imaging unit 20. When all the tracking targets including the new tracking target reach the designated position, the determination result is output. As a result, when all of the tracking targets detected first have successfully passed, it is possible to perform control to switch the color of the traffic light 11 based on the determination result by outputting the determination result.

  As described above, the embodiments of the present invention have been described with reference to the drawings. However, the specific configuration is not limited to these embodiments, and the present invention can be modified or added without departing from the scope of the present invention. Included in the invention. The road is not limited to a pedestrian crossing, but also includes various roads and railroad crossings that can be crossed by pedestrians, and the traffic light 11 is not limited to pedestrians but may be used as a vehicle. . Moreover, you may include a wheelchair, a bicycle, etc. in a pedestrian.

In addition, regarding detection of a pedestrian for realizing the control by each means described above, a technique other than image processing, such as detection using infrared light, an ultrasonic sensor, or laser light, may be used.
In addition, the longest time to continue the green light for pedestrians is set in advance, and even if the arrival to the specified position of the tracking target cannot be confirmed for some reason, it will switch to flashing blue light or red signal after the longest time has elapsed May be set.

  INDUSTRIAL APPLICABILITY The present invention can be widely used for a signal control device that controls the color of a traffic light that indicates whether or not a pedestrian can cross a road.

DESCRIPTION OF SYMBOLS 10 ... Signal control apparatus 11 ... Traffic light 20 ... Image pick-up part 30 ... Setting part 40 ... Processing part 41 ... Clock 42 ... Coefficient calculation part 43 ... Person detection part 44 ... Person tracking part 45 ... Determination part 50 ... Signal control apparatus

Claims (7)

In a signal control device that controls the color of a traffic light that indicates whether or not a pedestrian can cross a road,
Detection means capable of detecting a specific individual or group as a tracking target among pedestrians that start crossing after the traffic light switches from a red signal to a green signal;
A tracking means capable of tracking a situation in which the tracking object moves from a reference position on the crossing start side to a designated position on the crossing end side on the road;
Determination means capable of determining whether or not the tracking target has reached the specified position,
The detection means is based on an image sequentially input from a photographing unit capable of photographing at least a range from the reference position to the designated position until a predetermined second elapses from when the traffic light is switched from a red signal to a blue signal. Detecting the individual or group of pedestrians who have passed the reference position as tracking targets,
A signal control device characterized in that the signal control device triggers switching of the color of the traffic light based on the result of the determination by the determination means that the tracking target has been determined to have reached the designated position.   When the pedestrian that starts crossing after the traffic light is switched from the red signal to the green signal is not detected by the detection means, based on the passage of a predetermined time from when the traffic signal is switched from the red signal to the green signal, The signal control apparatus according to claim 1, wherein the signal control apparatus is an opportunity to switch a lamp color of the traffic light.   The signal control apparatus according to claim 2, wherein the predetermined time can change a predetermined specified value according to various parameters. The tracking means is based on the images sequentially input from the imaging unit, and the individual pedestrian or the pedestrian that is the tracking target until the tracking target detected by the detection means passes through the specified position. If the group is separated or binding signal control apparatus according to claim 1, 2 or 3, characterized in that track while recognizing as the new tracking target took over attributes tracked. The determination unit, when there is a new tracking target recognized by the tracking unit in addition to the tracking target detected by the detection unit based on the images sequentially input from the imaging unit, The signal control apparatus according to claim 4 , wherein the determination result is output when all tracking objects including a tracking object reach the designated position. A separate control means for controlling the switching of the light color of the traffic light,
The signal control device according to claim 5 , wherein the control unit performs control for blinking a pedestrian green signal based on a determination result from the determination unit. The signal control apparatus according to claim 6 , wherein the control unit performs control to switch to a red signal for pedestrians based on a determination result from the determination unit.

Images