JP4374609B2 - People counting device - Google Patents

Description

  The present invention relates to a people counting device for counting the number of people passing through.

  It has been conventionally performed to count the number of people passing through a passage at a specific place such as a store, a station, or a facility. For example, in the following Patent Document 1, a camera is installed from the upper side of the passage toward directly below, a moving object is extracted from an image captured by the camera, and it is determined whether or not the moving object is a human head. If it is a person's head, it counts as the number of passing people. Also, in Patent Document 2 below, a camera is installed at a bus entrance and exit, face data indicating the characteristics of a person's face is extracted from an image captured by the camera, stored in storage means, and stored face data The number of passengers is counted and managed based on the above. Further, in Patent Document 3 below, a pressure-sensitive sensor is laid on the floor of the doorway, and human footprint information is extracted based on the output from the pressure-sensitive sensor when a person steps on and stored in the storage means. Then, the footprint information stored this time and the footprint information stored previously are compared to determine whether or not they match. If they match, the current footprint information and the previous footprint information belong to the same person. By excluding it from counting the number of passing people, it is possible to prevent double counting of the same person.

Japanese Patent Publication No. 8-23882 (page 3, Fig. 4) JP 11-175782 (paragraphs 0018 to 0024) JP 2000-259809 (paragraphs 0017, 0018, FIG. 4)

  In the techniques such as Patent Documents 1 and 2 described above, when the same person passes through the passage many times, the same person is counted each time, so the total number of passing persons can be counted. There is a problem that it is not possible to count the number of net passing people who do not overlap counting. On the other hand, in the technique such as Patent Document 3, since the double counting of the same person is prevented as described above, the net number of passing persons can be counted. However, for example, when applied to the store entrance for the purpose of counting the number of customers visiting the store, even if the store clerk frequently enters and exits the store to perform the work, every time the store clerk's footprint information and previously stored in the storage means Since it has to be determined whether or not they coincide with each other's footprint information, there is a problem that processing takes time. In particular, after storing many visitors' footprint information in the storage means, the number of objects to be compared with the store clerk's footprint information is increased, so that the processing takes a very long time.

  The present invention solves the above-mentioned problems, and the object of the present invention is to be able to count the number of people who pass through the same person without counting the same person and pass even if a specific person passes frequently. It is an object of the present invention to provide a people counting device that does not take time to count people.

  The number counting device according to the present invention includes an extraction means for extracting feature data indicating the characteristics of a passing person, a first storage means for previously storing feature data indicating the characteristics of a specific person, and a passage extracted by the extraction means. A second storage means for storing a plurality of feature data of the person who performs the process, and by comparing the feature data extracted from the passing person with the feature data stored in the first storage means, the person who passes is specified The extracted feature data and the feature data stored in the second storage means are collated only when it is determined that the user is not a specific person in the first determination. The second determination of whether or not there is the same person is performed, and when it is determined that there is no same person in the second determination, the number of passing persons is counted. As the specific person, a person who is known in advance to frequently pass through a specific place where the number of passing persons is counted is set.

  As described above, first, whether or not the passing person (passing person) is a specific person is determined, and only when the first determination determines that the passing person is not a specific person, the same as the passing person If the passer is a specific person by performing the second determination as to whether or not the person is in the passerby (the person who has already passed), the passer's feature data and the second storage means The second determination is no longer performed by comparing the already-passed person's feature data stored in (1). For this reason, even if a specific person frequently passes a specific place where the number of passing persons is counted, it is possible to prevent the passing person counting process from taking time. Moreover, only when it is determined in the second determination that the same person is not present, the number of passing persons can be counted by counting the number of passing persons as the number of passing persons.

  As a specific method of counting the number of passing people, for example, a passing number counting unit that counts the number of passing people is provided, and when it is determined in the second determination that there is no same person, the value indicated by the passing number counting unit is 1. Is added, and the value obtained as a result of the addition is set as the number of passing people. By doing in this way, 1 can be added to the value which a passing number counting means shows for every first passing person who is not a specific person or an already passed person, and a net passing person number can be counted reliably.

  In another number counting device of the present invention, in addition to the extracting means, the first storage means, and the second storage means described above, a total passing person count for detecting the passing person and counting the total passing person number. If it is determined that the person is a specific person in the first determination, 1 is subtracted from the value indicated by the total passing person counting means, and if it is determined that the person is not a specific person, the second determination is performed. When it is determined in the second determination that there is the same person, 1 is subtracted from the value indicated by the total passing person counting means, and the value obtained as a result of the subtraction is set as the passing person number. In this way, every time a specific person passes or every time the same person as an already-passed person passes, 1 is subtracted from the total passing person number indicated by the total passing person counting means, thereby ensuring the net passing person number. Can be counted.

  Further, in the above-described number counting device, the feature data extracted by the extraction unit is stored in the second storage unit only when it is determined that the person is not a specific person in the first determination. By doing so, the specific person's feature data is no longer stored in the second storage means, and thus it is possible to suppress an increase in the number of feature data to be subjected to collation and second determination with the passer's feature data. It is possible to reduce the time required for the collation and determination, so that it does not take much time for the passing person counting process.

  Further, another number counting device of the present invention includes the above-described extracting means, first storing means, second storing means, and total passing number counting means, and is not a specific person in the first determination. Only when the determination is made, the feature data extracted by the extraction means is stored in the second storage means, and one feature data stored in the second storage means is compared with the other feature data to find the same person. The second determination of whether or not the same person is present in the second determination is performed, and then the second determination is performed as one feature data for the same person. For other feature data determined to be present, the second determination is not performed as the first feature data, and it is determined by the first determination that the person is a specific person from the value indicated by the total passing person counting means. And the second decision By subtracting the number of judges that person is present, a pass number values of the result of the subtraction. In this way, the feature data of the same person included in the other feature data can be obtained by not performing the same judgment as other feature data that has been judged to be the same person in the second judgment. Since collation between data and the second determination are not performed, it is possible to reduce the number of times of the collation and determination, so that it does not take much time for the passing number counting process. Further, the total number of passing people is subtracted from the value indicated by the total passing number counting means by the number determined to be a specific person in the first determination and the number determined to be the same person in the second determination. Thus, the number of passing the specific person and the number of passing the same person as the already passed person are subtracted, so that the net number of passing persons can be counted reliably.

  Further, the above-described number counting device includes an imaging unit that images a passing person, and the extraction unit extracts facial data indicating the characteristics of the person's face from the image captured by the imaging unit as feature data. The first storage unit stores in advance face data indicating the features of the face of the specific person as feature data. In this way, by adopting face data as feature data, data can be easily extracted without hindering passage of passers-by by the imaging means and extraction means. Further, since the face data is data that is surely different for each person and can accurately identify each person, the first determination as to whether or not the passer is a specific person is the same as the passer. It is possible to accurately perform the second determination as to whether or not a person is among those who have already passed.

  According to the present invention, it is first determined whether or not the passing person is a specified person, and only when the passing person is not a specified person, whether or not the same person as the passing person is in the already passed person is determined. Therefore, even if a specific person frequently passes a specific place where the number of passing persons is counted, the process of counting the number of passing persons does not take time, and the same person as the passing person is present in the already passing persons. Since the number of passing people is counted only when there is not, it is possible to count the number of passing people without counting the same person repeatedly.

  FIG. 1 is a block diagram of a people counting apparatus according to the first embodiment of the present invention. In FIG. 1, a portion surrounded by a broken line is the number counting device 101 according to the present embodiment. Reference numeral 1 denotes a computer, which includes a control unit 11, a storage unit 12, an operation unit 13, and a display unit 14. The control unit 11 includes a CPU and a memory such as a RAM and a ROM, and controls each unit of the computer 1. The storage unit 12 includes a memory such as a RAM or a non-volatile memory, and a hard disk, and is provided with a passing number counter C1 that is an area for counting the number of passing persons. The operation unit 13 includes a keyboard and a mouse. The display unit 14 is composed of a CRT (cathode ray tube) and displays the number of passing people as will be described later. Reference numeral 2 denotes a monitoring camera that captures a moving image. The monitoring camera 2 is installed toward a passage at a specific place such as a store, a station, a facility, etc., images a person passing through the passage, and outputs the captured image to the computer 1. The control unit 11 of the computer 1 detects a human face from an image captured by the monitoring camera 2 and extracts face data indicating the feature of the face from the detected face as human feature data.

  Reference numeral 3 denotes a first storage device composed of a hard disk, which stores a specific person database 30. The specific person database 30 stores face data indicating the characteristics of the specific person's face in advance. As the specific person, a person who is known in advance to frequently pass through a specific place where the number of passing persons is counted is set. For example, if it is a store, it is a clerk, if it is a transportation facility such as an airport or a station, it is a staff member of the transportation facility, if it is a facility such as an amusement park or theme park, it is a staff of the facility. Reference numeral 4 denotes a second storage device composed of a hard disk, which stores a passer-by database 40. In the passer database 40, face data extracted from an image captured by the monitoring camera 2 by the control unit 11 is stored.

  Reference numeral 5 denotes a face image input device composed of a scanner or a digital camera. The face image input device 5 is connected to the computer 1 in order to capture an image obtained by capturing a specific person. When the control unit 11 of the computer 1 captures the image of the specific person from the face image input device 5, the control unit 11 detects the face of the specific person from the image and extracts face data indicating the feature of the face from the detected face. And the control part 11 registers the extracted specific person's face data in the specific person database 30 so that it may mention later.

  In the above configuration, the control unit 11 of the computer 1 constitutes an embodiment of the extraction means in the present invention. The surveillance camera 2 constitutes an embodiment of the imaging means in the present invention. The first storage device 3 storing the specific person database 30 constitutes an embodiment of the first storage means in the present invention, and the second storage device 4 storing the passer-by database 40 is the second storage device in the present invention. One embodiment of the storage means is configured. The passing number counter C1 constitutes one embodiment of passing number counting means for counting the passing number, and the display unit 14 constitutes one embodiment of notification means for notifying the passing number counted by the number counting device 101. To do.

  FIG. 2 is a diagram showing an example of the stored contents of the above-described specific person database 30. The specific person database 30 stores a specific person face feature amount 32 each time registration processing of the specific person's face data is performed by the control unit 11 described above, and identifies the registration date and time 33 and the specific person. Therefore, the ID number 34 assigned for each feature amount is stored. 31 is a registration number indicating the registration order.

  FIG. 3 is a diagram showing an example of the stored contents of the above-described passer database 40. Each time the face data of a person who has passed through a specific place (hereinafter referred to as a “passer”) is stored in the passer database 40, a facial feature 42 of the passer is stored. The passing date and time 43 and the ID number 44 assigned for each feature amount for identifying the passing person are stored. Reference numeral 41 denotes a storage number indicating the storage order.

  FIG. 4 is a flowchart showing a procedure for registering face data of a specific person in the specific person database 30. This procedure is executed by the control unit 11 of the computer 1 when, for example, a registration start button (not shown) is displayed on the display unit 14 and the registration start button is operated by the operation unit 13. . In FIG. 4, when the registration start button is operated after the face image input device 5 described above is connected to the computer 1, the control unit 11 acquires data of an image obtained by capturing a specific person from the face image input device 5. (Step S1). Next, the control unit 11 detects a face from the acquired image data (step S2). For this face detection, known methods such as a skin color detection method, a method for extracting a difference from a background image, and a method for extracting the likelihood of a face from pattern matching are used. When the face is detected, the control unit 11 detects feature points such as eyes, nose, mouth, and ears from the face image, and extracts feature amounts relating to the shape and position of each detected feature point (step S3). This feature amount extraction can also be performed by using a known method such as a method of extracting a feature amount from a grayscale image of eyes, nose, mouth, ears or the like by template matching. When the feature amount is extracted, the control unit 11 stores and registers the feature amount in the specific person database 30 as the specific person face data (step S4), and ends the process. As a result, the specific person database 30 stores the registration number 31, the registration date and time 33, and the ID number 34 together with the characteristic amount 32 of the specific person as shown in FIG.

  FIG. 5 is a flowchart showing a procedure for counting the number of passing persons according to the first embodiment of the present invention. In this procedure, for example, a counting start button (not shown) and a counting end button are displayed on the display unit 14 described above, and the control unit 11 of the computer 1 operates when the counting start button is operated by the operation unit 13 described above. When the count end button is operated by the operation unit 13, the control unit 11 ends. Instead of the above method, this procedure is performed after, for example, the operation unit 13 is operated to set the counting start time (store opening time, etc.) and the counting end time (store closing time, etc.). The controller 11 may be executed when the start time is reached, and may be ended when the count end time is reached.

  In FIG. 5, for example, when the above-described counting start button is operated, the control unit 11 clears (deletes all) the stored contents of the passer database 40 (step S11), and the passing number counter C1 of the storage unit 12 indicates. The value is cleared (initialized to 0) (step S12). Thereafter, the control unit 11 acquires image data output from the monitoring camera 2 (step S13), and detects a passer's face from the acquired image data (step S14). When the face is detected, the control unit 11 tracks the detected face (step S15), detects feature points such as eyes, nose, mouth, ears, and the like from the face image that is a still image obtained by tracking and detects the detected face. For each of the feature points, a feature amount related to their shape and position is extracted (step S16). By tracking, multiple face images can be acquired, but in this example, for each of the multiple face images, roughly grasp the position of the eyes, nose, mouth, ears, etc., and compare with the template, A score value indicating the degree of facialness is calculated from the comparison result, and the one face image with the highest score value, that is, the most face-like one with the eyes, nose, mouth, ears, etc., captured most firmly. A feature amount is extracted from the face image.

  When the feature amount is extracted as described above, the control unit 11 collates the extracted feature amount with the feature amount 32 (FIG. 2) of the specific person stored in the specific person database 30 (step S17), and this time It is determined whether or not the passing person is a specific person (step S18). Here, if the extracted feature quantity matches the feature quantity 32 of the specific person database 30, the control unit 11 determines that the specific quantity is specified (step S18: YES). If it judges in this way, control part 11 will transfer to Step S13 in order to process the next passing person, and will again perform Step S13-Step S18 in the procedure mentioned above.

  On the other hand, if the extracted feature quantity does not match the feature quantity 32 of the specific person database 30 in step S18, the control unit 11 determines that the specific person is not a specific person (step S18: NO). If it judges in this way, the control part 11 will collate the extracted feature-value with the feature-value 42 (FIG. 3) of the already passed person (henceforth already passed) memorize | stored in the passer-by database 40. (Step S19), it is determined whether or not the same person as the current passing person is among the passing persons (Step S20). Here, if the extracted feature value does not match the feature value 42 of the passer-by database 40, the control unit 11 determines that there is no same person (step S20: NO). Note that since the stored contents of the passer database 40 are cleared in step S11, the control unit 11 matches the extracted feature quantity in step S20 even when no feature quantity 42 is stored in the passer database 40. Since there is no feature value 42 in the passer-by database 40, it is determined that there is no same person (step S20: NO). If it judges in this way, the control part 11 will add 1 to the value which the passing person counter C1 shows (step S21), and memorize | store and memorize | store the extracted feature-value in the passer database 40 as passer's face data (step S21). Step S23). At this time, as shown in FIG. 3, the passer database 40 stores a storage number 41, a pass date / time 43, and an ID number 44 along with the passer's feature 42.

  On the other hand, when the extracted feature value matches the feature value 42 of the passer-by database 40 in step S20, the control unit 11 determines that there is the same person (step S20: YES). If it judges in this way, control part 11 will raise the same person detection flag in connection with the feature-value extracted this time (Step S22). The area where the same person detection flag is set may be provided in the passer-by database 40 or may be provided in the memory built in the control unit 11 or the storage unit 12. When the same person detection flag is set, the control unit 11 stores and saves the extracted feature amount in the passer database 40 as the passer's face data (step S23). Also at this time, as shown in FIG. 3, the passer database 40 stores the saver number 41, the pass date and time 43, and the ID number 44 along with the passer feature 42.

  When the feature amount is stored, the control unit 11 determines whether or not to finish counting the number of passing people (step S24). Here, for example, if the above-described counting end button has not been operated yet, the control unit 11 determines that the counting of the number of passing people is not finished (step S24: NO), and step S13 in order to process the next passing person. Then, step S13 to step S24 are executed again according to the procedure described above. Then, after repeatedly executing Step S13 to Step S24 several times, if the counting end button is operated in Step S24, the control unit 11 determines that the counting of the passing number of people is finished (Step S24: YES). The process is terminated.

  After the processing is completed as described above, the control unit 11 notifies the display unit 14 of the value indicated by the passing number counter C1 at this time as the passing number at the specific place. FIG. 6 is a diagram illustrating an example of a screen displayed on the display unit 14 at this time. In FIG. 6, the display screen 50 displays a counting place 51 where the number of passing people is counted, a counting date 52, a counting time 53, and a passing number 54 that is a value indicated by a passing number counter C1.

  As in the first embodiment described above, it is first determined whether or not the passer is a specific person. Only when it is determined that the passer is not a specific person, the same person as the passer When the passer is a specific person, the passer's feature amount is checked against the passer's feature amount stored in the passer database 40. No longer do. For this reason, even if a specific person frequently passes a specific place where the number of passing persons is counted, it is possible to prevent the passing person counting process from taking time. Further, only when it is determined that there is no same person in the same person determination, that is, every time a first passer who is not a specific person or a passing person passes, 1 is added to the value indicated by the passing person counter C1. Thus, it is possible to reliably count the number of net passing persons who do not repeatedly count the same person.

  Further, only when it is determined that the person is not a specific person in the specific person determination, the extracted feature amount is stored in the passer database 40, so that the specific person feature amount is not stored in the passer database 40. It is possible to suppress an increase in the number of feature quantities to be subjected to collation and determination of the same person and to reduce the time required for the collation and determination so that the passing person count process does not take much longer. be able to. Further, by adopting the facial feature amount as the human feature data, the surveillance camera 2 and the control unit 11 can easily extract data without obstructing the passage of the passer-by, and in addition, the facial feature Since the amount is data that is surely different for each person and can identify each person with high accuracy, the specific person determination and the same person determination can be performed accurately. Furthermore, by displaying the number of passing people counted by the above-described counting procedure on the display unit 14 and notifying it, it becomes possible to grasp the net passing number of people in a specific place at a glance.

  FIG. 7 is a block diagram of a people counting apparatus according to the second embodiment of the present invention. In the figure, the same parts as those described in FIG. In FIG. 7, the people counting device 102 according to the present embodiment is provided with a human detection sensor 6 in addition to the configuration provided in the people counting device 101 of FIG. 1. The human detection sensor 6 is composed of a sensor capable of detecting a person such as a photo sensor or a thermal sensor, detects a person passing through a passage in a specific place, and transmits a detection signal to the computer 1. . Each time the control unit 11 of the computer 1 receives a detection signal from the human detection sensor 6, the control unit 11 adds 1 to the value indicated by the total passing number counter C2 provided in a predetermined area of the storage unit 12, and passes through the passage. Count the total number of people who pass. The human detection sensor 6 and the total passing number counter C2 of the storage unit 12 constitute one embodiment of the total passing number counting means in the present invention.

  FIG. 8 is a flowchart showing a procedure for counting the number of passing persons according to the second embodiment of the present invention. In FIG. 8, for example, when the above-described counting start time set in advance is reached, the control unit 11 clears the stored contents of the passing person database 40 (step S31), and sets the value indicated by the total passing number counter C2 of the storing part 12. Clear (step S32). Thereafter, when the control unit 11 receives a detection signal indicating that a person passing through the passage at the specific location is detected from the human detection sensor 6 (step S33: YES), the control unit 11 sets 1 to the value indicated by the total passing number counter C2. Add (step S34). And the control part 11 acquires the image data output from the monitoring camera 2 (step S35), and detects a passer's face from the acquired image data (step S36). When the face is detected, the control unit 11 tracks the detected face (step S37), and extracts a feature amount from the tracked face image (step S38). And the control part 11 collates the extracted feature-value with the specific person's feature-value 32 memorize | stored in the specific person database 30 (step S39), and it is determined whether this time passer is a specific person. Determine (step S40). Here, if the extracted feature quantity matches the feature quantity 32 of the specific person database 30, the control unit 11 determines that the specific person is specified (step S40: YES). If it judges in this way, the control part 11 will subtract 1 from the value which the total passing number counter C2 shows (step S41), will transfer to step S33 in order to process the next passing person, and will again step S33-step S41 is executed according to the procedure described above.

  On the other hand, if the extracted feature quantity does not match the feature quantity 32 of the specific person database 30 in step S40, the control unit 11 determines that the specific person is not a specific person (step S40: NO). If it judges in this way, the control part 11 will collate the extracted feature-value with the feature-value 42 of the already-passed person memorize | stored in the passer-by database 40 (step S42), and the same person as the passer has already passed. It is determined whether or not the user is present (step S43). Here, if the extracted feature value and the feature value 42 of the passer-by database 40 do not match, the control unit 11 determines that there is no same person (step S43: NO). If it judges in this way, the control part 11 will memorize | store and memorize | store the extracted feature-value in the passer database 40 as a passer's face data (step S46).

  On the other hand, if the extracted feature value matches the feature value 42 of the passer-by database 40 in step S43, the control unit 11 determines that there is the same person (step S43: YES). If it judges in this way, control part 11 will raise the same person detection flag in relation to the extracted feature-value (Step S44), and will subtract 1 from the value which total passage number counter C2 shows (Step S45). Thereafter, the control unit 11 stores the extracted feature amount in the passer database 40 as passer's face data and saves it (step S46).

  When the feature amount is stored, the control unit 11 determines whether or not to finish counting the number of passing people (step S47). Here, for example, if the above-mentioned counting end time set in advance is not yet reached, the control unit 11 determines that the counting of the number of passing people is not finished (step S47: NO), and processes the next passing person. Then, the process proceeds to step S33, and steps S33 to S47 are executed again by the procedure described above. Then, after repeatedly executing step S33 to step S47 several times, if the count end time is reached in step S47, the control unit 11 determines to finish counting the number of passing people (step S47: YES), End the process. After the processing is completed, the control unit 11 displays the display screen 50 of FIG. 6 described above on the display unit 14, for example, and the number of passing people 54 in the screen 50 indicates the total passing number counter C2 at this time. Display the value as the number of people passing through a specific location.

  According to the second embodiment described above, when the passer is a specific person, the determination is performed by comparing the feature quantity of the passer with the feature quantity of the passer stored in the passer database 40. Therefore, even if the specific person frequently passes a specific place where the number of passing persons is counted, it is possible to prevent the passing person counting process from taking time. Also, when it is determined that the person is a specific person in the specific person determination (that is, every time the specific person passes), or when it is determined that there is the same person in the same person determination (that is, the same person as the already-passed person is By subtracting 1 from the total passing number indicated by the total passing number counter C2 every time it passes, the net passing number can be counted reliably.

  FIG. 9 is a block diagram of a people counting apparatus according to the third embodiment of the present invention. In the figure, the same parts as those described in FIG. 9, in addition to the configuration provided in the number counting device 101 of FIG. 1, the number counting device 103 according to the present embodiment detects the number of people passing through a passage at a specific place and counts the total number of passing people. A device 7 is provided. For example, the person counting device 7 detects a head from a device that detects and counts a moving body that passes under an infrared sensor installed on a ceiling portion of a passage, or an image captured by a camera from above a passing person. Known devices such as counting devices, devices that detect the head and the height of the head from the video taken by a 3D camera from above the passerby, and distinguish between humans and non-human dogs and cats It consists of counting equipment. The person counting device 7 constitutes one embodiment of the total passing person counting means in the present invention. The control unit 11 of the computer 1 acquires total passing number data indicating the total passing number from the connected number counting device 7. The storage unit 12 of the computer 1 is provided with a specific person passing counter C3 that is an area for counting the number of specific persons passing.

  FIG. 10 is a flowchart showing a procedure for storing the passer's face data in the passer database 40 according to the third embodiment of the present invention. Note that this procedure is executed by the control unit 11 of the computer 1 at a predetermined cycle, for example, from a preset storage start time (store opening time, etc.) to a storage end time (store closing time, etc.). . For example, the control unit 11 clears the stored content of the passer database 40 and clears the value indicated by the specific person pass counter C3 of the storage unit 12 before the storage start time. The number-of-people counting device 7 clears (initializes to 0) the total number of passing people, for example, before the storage start time, and counts the total number of passing people, for example, from the storage start time to the storage end time. .

  In FIG. 10, first, the control unit 11 acquires image data output from the monitoring camera 2 (step S51), and detects a passer's face from the acquired image data (step S52). When the face is detected, the control unit 11 tracks the detected face (step S53), and extracts a feature amount from the tracked face image (step S54). And the control part 11 collates the extracted feature-value with the specific person's feature-value 32 memorize | stored in the specific person database 30 (step S55), and it is determined whether this time passer is a specific person. Determination is made (step S56). Here, if the extracted feature quantity and the feature quantity 32 of the specific person database 30 do not match, the control unit 11 determines that it is not a specific person (step S56: NO). If it judges in this way, the control part 11 will memorize | store and preserve | save the extracted feature-value in passer database 40 as passer's face data (step S57), and will complete | finish a process. On the other hand, when the extracted feature quantity matches the feature quantity 32 of the specific person database 30 in step S56, the control unit 11 determines that the specific person is specified (step S56: YES). If it judges in this way, control part 11 will add 1 to the value which specific person passage counter C3 shows (Step S58), and will end processing. Thereafter, the control unit 11 repeatedly executes Steps S51 to S58 at a predetermined cycle as described above.

  FIG. 11 is a flowchart showing a procedure for counting the number of passing persons according to the third embodiment of the present invention. This procedure is executed by the control unit 11 of the computer 1 when, for example, a counting start button (not shown) displayed on the display unit 14 is operated or when a preset counting start time is reached. To do.

  In FIG. 11, for example, when the above-described counting start button is operated, the control unit 11 acquires the total passing number data from the number counting device 7 (step S61), and the value indicated by the specific person passing counter C3 from the storage unit 12 Is read (step S62). Subsequently, the control unit 11 calculates the number of passing persons by subtracting the number of specified persons passing indicated by the specified person passing counter C3 from the total number of passing persons indicated by the total number of passing person data (step S63). The control unit 11 temporarily stores the calculated number of passing people in a predetermined area of the memory built in the control unit 11 or a predetermined area of the storage unit 12.

  Next, the control part 11 reads the feature-value 42 of 1 from the passer-by database 40 (step S64). Note that the reading of the feature value 42 from the passer-by database 40 is performed, for example, one by one in order from the earlier or later stored order. When one feature amount 42 is read, the control unit 11 determines whether or not the same person detection flag is associated with the feature amount 42 (step S65). At this time, since only one feature 42 has been read from the passerby database 40 and no identical person detection flag has been set, the control unit 11 determines that there is no identical person detection flag (step S65: NO). If it judges in this way, the control part 11 will collate 1 feature-value 42 which was read, and the other feature-value 42 memorize | stored in the passer-by database 40 (step S66), and it will become 1 feature-value 42. It is determined whether or not the same person as the corresponding passer is among the passers corresponding to the other feature amount 42 (step S67). Here, if the one feature quantity 42 and the other feature quantities 42 do not match, the control unit 11 determines that there is no same person (step S67: NO), and proceeds to step S70. On the other hand, when one feature value 42 matches any one of the other feature values 42 in step S67, the control unit 11 determines that the same person exists (step S67: YES). The same person detection flag is set in association with another feature amount 42 corresponding to the same person that matches 42 (step S68). Then, when all the same person detection flags are set, the control unit 11 proceeds to step S70.

  If transfering it to step S70, the control part 11 will determine whether all the feature-values 42 were read from the passer-by database 40. FIG. At this time, since only one feature value 42 has been read from the passer-by database 40, the control unit 11 determines that all the feature values 42 have not been read (step S70: NO). If it judges in this way, the control part 11 will transfer to step S64, will read the next one feature-value 42 from the passer-by database 40, will be linked | related with the said feature-value 42, and it will be checked whether there exists the same person detection flag. Determination is made (step S65). At this time, if the one feature value 42 read out is another feature value 42 in which the same person detection flag is set in step S68 described above, the control unit 11 determines that there is the same person detection flag (step S65). : YES). If it judges in this way, the control part 11 will subtract 1 from the passing number of persons calculated in step S63 mentioned above (step S69), and will transfer to step S70. The control unit 11 updates the passing number by overwriting the passing number after the subtraction with the passing number before the subtraction.

  If it transfers to step S70 again, the control part 11 will judge whether all the feature-values 42 were read from the passer-by database 40, and if it judges that all the feature-values 42 were not read (step S70: NO), step S64 Then, step S64 to step S70 are repeatedly executed according to the procedure described above. Then, after repeatedly executing step S64 to step S70 several times, when the control unit 11 determines in step S70 that all the feature values 42 have been read from the passer-by database 40 (step S70: YES), the process ends. . After the processing is completed, the control unit 11 displays the display screen 50 of FIG. 6 described above on the display unit 14, for example, and the number of passing people at this time is displayed in the column of the number of passing people 54 in the screen 50. Display as number of people. In addition to the items shown in FIG. 6, the total number of passing persons acquired from the person counting device 7 and the number of specified persons passing by the specified person passing counter C <b> 3 may be displayed on the display screen 50.

  According to the third embodiment described above, when the passer is a specific person, the feature quantity of the specific person extracted in the storage procedure of FIG. 10 is not stored in the passer database 40, so the counting procedure of FIG. In this case, the same person determination and the verification for the determination are not performed, so that it is possible to prevent the passing person counting process from taking time. Further, another feature value 42 determined by the same person determination that the same person is present is included in the other feature value 42 by not performing the same determination as the one feature value 42 thereafter. Since the matching between the feature amounts 42 of the same person and the determination of the same person are not performed, the number of times of the matching and determination can be reduced, so that the passing person counting process does not take much time. Further, from the total number of passing persons acquired from the person counting device 7, the number of passing the specific person indicated by the specific person passing counter C3 and the characteristic amount 42 in which it is determined that the same person exists in the same person determination and the same person detection flag is set. By subtracting the number, the net number of people passing through can be counted reliably.

  The present invention can adopt various forms other than the embodiment described above. For example, in the second embodiment shown in FIG. 7 and FIG. 8, a case where the human detection sensor 6 and the described passing number counter C2 are provided to count the total number of passing people is taken as an example. The person counting device 7 shown in FIG. 9 may be provided. In this case, what is necessary is just to control so that the total number of passing people counted by the people counting device 7 can be changed by the control part 11. Moreover, in 3rd Embodiment shown in FIGS. 9-11, although the case where the number counting apparatus 7 was provided in order to count a total passing number of persons was mentioned as an example, it replaced with this and the human detection sensor 6 and A passing number counter C2 may be provided. In addition to the above, for example, every time a human face is detected from the image data from the monitoring camera 2, the total number of passing people may be counted.

  In the above embodiment, the case where the feature amount of the person's face is stored in the specific person database 30 and the passer-by database 40 is taken as an example, but the face image itself may be stored instead of the feature amount. Both the feature amount and the face image may be stored.

  Moreover, although the case where the display part 14 of the computer 1 was used as an example in the above embodiment as a notification means for notifying the counted number of passing people, the present invention is not limited to this. Other than this, as a notification means, for example, a display device that can display the number of passing people, such as an electric bulletin board, a voice output device that can output the number of passing people as a voice, such as an audio output circuit and a speaker, and a passage such as a printer You may make it use the printing apparatus etc. which can print the number of people on paper sheets. In other words, the notification means may be any means for notifying so that the counted number of passing people can be grasped.

  Furthermore, although the case where the feature-value which shows the feature of a person's face is employ | adopted as an example in the above embodiment as feature data which shows the feature of a person, this invention is not limited only to this. Other than this, for example, data indicating the characteristics of the silhouette of the whole body of the person or data indicating the characteristics of the footprints may be employed as the characteristic data. That is, the feature data may be data that can identify each person.

It is a block diagram of a people counting device concerning a 1st embodiment. It is a figure which shows an example of the memory content of a specific person database. It is a figure which shows an example of the memory content of a passer database. It is a flowchart which shows the registration procedure of a specific person's face data. It is a flowchart which shows the counting procedure which concerns on 1st Embodiment. It is a figure which shows an example of the screen displayed on a display part. It is a block diagram of the people counting device concerning a 2nd embodiment. It is a flowchart which shows the counting procedure which concerns on 2nd Embodiment. It is a block diagram of the people counting device concerning a 3rd embodiment. It is a flowchart which shows the preservation | save procedure concerning 3rd Embodiment. It is a flowchart which shows the counting procedure which concerns on 3rd Embodiment.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Computer 2 Monitoring camera 3 1st memory | storage device 4 2nd memory | storage device 6 Person detection sensor 7 Person count device 11 Control part C2 Total passing number counter 101,102,103 Person count apparatus

Claims (5)

Extraction means for extracting feature data indicating the characteristics of the passing person;
First storage means for storing in advance characteristic data indicating the characteristics of a specific person;
Second storage means for storing a plurality of passing person characteristic data extracted by the extraction means,
A first determination is made as to whether or not the passing person is a specific person by comparing the feature data extracted from the passing person with the feature data stored in the first storage means, and the first determination Only when it is determined that the person is not a specific person, the extracted feature data is compared with the feature data stored in the second storage means to determine whether or not the same person exists, A number counting device characterized in that when it is determined in the second determination that there is no same person, the number of passing people is counted. Extraction means for extracting feature data indicating the characteristics of the passing person;
First storage means for storing in advance characteristic data indicating the characteristics of a specific person;
Second storage means for storing a plurality of feature data of passing persons extracted by the extraction means;
A total number of passing people counting means for detecting the number of passing people and counting the total number of passing people,
A first determination is made as to whether or not the passing person is a specific person by comparing the feature data extracted from the passing person with the feature data stored in the first storage means, and the first determination 1 is subtracted from the value indicated by the total passing number counting means when it is determined that it is a specific person, and when it is determined that it is not a specific person, it is stored in the extracted feature data and the second storage means. When the second determination as to whether or not the same person is present is performed by comparing with the feature data, and it is determined in the second determination that the same person is present, 1 is determined from the value indicated by the total passing number counting means. The number counting device is characterized in that the value obtained as a result of the subtraction is used as the number of passing people. In the people counting device according to claim 1 or 2,
The number counting device, wherein the feature data extracted by the extraction means is stored in the second storage means only when it is determined that the person is not a specific person in the first determination. Extraction means for extracting feature data indicating the characteristics of the passing person;
First storage means for storing in advance characteristic data indicating the characteristics of a specific person;
Second storage means for storing a plurality of feature data of passing persons extracted by the extraction means;
A total number of passing people counting means for detecting the number of passing people and counting the total number of passing people,
A first determination is made as to whether or not the passing person is a specific person by comparing the feature data extracted from the passing person with the feature data stored in the first storage means, and the first determination Only when it is determined that the person is not a specific person, the feature data extracted by the extraction means is stored in the second storage means,
A second determination is made as to whether or not the same person exists by collating the one feature data stored in the second storage means with the other feature data, and the same person is determined in the second determination. For the other feature data determined to be absent, the second determination is performed as the first feature data, and for the other feature data determined to be the same person, the first Without performing the second determination as feature data,
Subtracting the number determined to be a specific person in the first determination and the number determined to be the same person in the second determination from the value indicated by the total passing number counting means A number counting device, characterized in that a result value is a passing number. In the people counting device in any one of Claims 1 thru | or 4,
An imaging means for imaging a passing person;
The extraction means extracts face data indicating the characteristics of a human face from the image captured by the imaging means as the feature data,
The number counting device according to claim 1, wherein the first storage means stores face data indicating a feature of a specific person's face in advance as the feature data.

Images