Information processing devices, methods, programs, and storage media

Abstract

Set up a passage detection line that allows for more accurate passage detection. [Solution] An information processing device comprising: an acquisition unit for acquiring an image; a holding unit for holding trajectory information of a person in the image; and a determination unit for determining a passage detection line for detecting the passage of a person in the image acquired by the acquisition unit, wherein the determination unit determines the passage detection line based on at least one piece of trajectory information held in the holding unit.

Description

Technical Field

[0006] , , , ,

[0005] , , , ,

[0001] Relates to an information processing apparatus, method, program, and storage medium.

Background Art

[0002] Conventionally, there has been a need to measure the number of people passing through specific areas such as doors and gates in various facilities such as event venues, shopping malls, and theme parks, and utilize the results for purposes such as congestion alleviation and marketing. For such needs, a technology for counting the number of people passing through by video analysis has been utilized. The technology for counting the number of people passing through analyzes the video captured by an imaging device, detects and tracks the head of a person, etc., and detects that the trajectory has passed through a virtual passing detection line set on the video, and counts the number of detected passes.

[0003] As a technology related to counting the number of people passing through, Patent Document 1 discloses a method of grouping trajectories based on the moving direction and efficiently setting a passing detection line based on the result.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] Generally, in order to count the target persons without omission, it is desirable to set a passing detection line at a position where the trajectories detected and tracked of the head, etc. intersect as many times as possible.

[0006] However, manually setting the passage detection lines can be difficult because the way the lines are drawn varies depending on the person setting them, making it challenging to accurately detect the passage of people. For example, if the passage detection lines are drawn to match the height of a tall person, shorter people may not be detected, and conversely, if the lines are drawn to match the height of a short person, tall people may not be detected.

[0007] Therefore, the problem that the present invention aims to solve is to set up a passage detection line that enables more accurate passage detection. [Means for solving the problem]

[0008] To solve the above problems, an imaging device according to one aspect of the present invention includes an image acquisition unit that acquires an image, The device comprises a holding unit that holds trajectory information of a person in the image, and a determination unit that determines a passage detection line for detecting the passage of a person in the image acquired by the acquisition unit, wherein the determination unit determines the passage detection line based on at least one piece of trajectory information held in the holding unit. [Effects of the Invention]

[0009] According to the present invention, it is possible to set a passage detection line that enables more accurate passage detection. [Brief explanation of the drawing]

[0010] [Figure 1] Diagram showing the system configuration [Figure 2] Diagram showing the configuration of the information processing device. [Figure 3] Diagram showing the functional configuration of the information processing device. [Figure 4] Flowchart showing the process of collecting trajectory information [Figure 5] Flowchart showing the process for determining the passage detection line. [Figure 6] A diagram showing the trajectory of a person and the determined passage detection line. [Figure 7] Flowchart showing the process for determining the passage detection line. [Figure 8] A diagram showing the trajectory of a person and the determined passage detection line. [Figure 9] A diagram showing the trajectory of a person and the determined passage detection line. [Figure 10] Flowchart showing the process for determining the passage detection line. [Figure 11] A diagram showing the trajectory of a person and the determined passage detection line. [Modes for carrying out the invention]

[0011] The embodiments for carrying out the present invention will be described in detail below with reference to the attached drawings. The embodiments described below are examples of means for realizing the present invention and should be modified or changed as appropriate depending on the configuration of the apparatus to which the present invention is applied and various conditions, and the present invention is not limited to the embodiments described below. Furthermore, some of the embodiments described later may be combined as appropriate.

[0012] <Embodiment 1> (System Configuration) Figure 1 shows a people counting system according to this embodiment. The people counting system includes an imaging device 101, a network 102, an information processing device 103, a display device 104, and an input device 105.

[0013] The imaging device 101 captures an arbitrary monitoring area and transmits the captured image to the information processing device 103 via the network 102. The information processing device 103 can analyze the transmitted captured image and count the number of people on the captured image. Further, the information processing device 103 is connected to the display device 104 and the input device 105, and the display device 104 displays the captured image and the analysis result of the captured image (such as the number of counted people) on the display device. On the other hand, the input device 105 is a device for the user to operate the information processing device 103, such as a mouse or a keyboard. Therefore, in the people counting system according to the present embodiment, the user can confirm the output from the information processing device 103 via the display device 104 and operate the information processing device 103 via the input device 105. Since the information processing device 103 is connected to the imaging device 101 via the network as described above, it is also possible to operate the imaging device 101 based on an instruction from the user. For example, it is possible to change the imaging direction of the imaging device 101 or adjust the brightness of the image.

[0014] (Device Configuration) FIG. 2 is a diagram showing the device configuration of the information processing device according to the present embodiment. The information processing device 103 includes a CPU 201, a RAM 202, a ROM 203, an input I / F (interface) 204, an output I / F 205, and a network I / F.

[0015] The CPU 201, the RAM 202, and the ROM 203 are hardware for executing each function (software) of the information processing device, 103. That is, by loading various programs stored in the ROM 203 into the RAM 202 and having the CPU 201 execute using the RAM 202 as a workspace, the functions as software are achieved.

[0016] The input I / F 204, output I / F 205, and network I / F 206 are various interfaces for realizing communication between the information processing apparatus 103, the user, and the network. The input I / F 204 and output I / F 205 are terminals conforming to standards such as USB or HDMI (registered trademark), and the network I / F 206 is a communication terminal such as a LAN connector. Communication is performed with the display device 104, input device 105, network 102 (and the imaging device 101 beyond it) via these interfaces.

[0017] (Functional Configuration) FIG. 3 is a diagram showing the functional configuration of the information processing apparatus according to the present embodiment. The information processing apparatus 103 includes an input unit 301, a setting unit 302, a determination unit 303, an acquisition unit 304, a detection unit 305, a tracking unit 306, and a holding unit 307. Among these functional blocks, the functions realized by software are realized by executing a program by the CPU 201, RAM 202, ROM 203, etc. as described above.

[0018] The input unit 301 receives user input regarding settings and operations related to the information processing apparatus 103. While viewing the GUI displayed on the display device 104, the user can give instructions to change, save, and apply settings to the setting unit 302. The setting unit 302 stores, for example, settings (position, direction, length, etc.) regarding the passing detection line determined by the determination unit 303.

[0019] The acquisition unit 304 acquires an image captured by the imaging device 101 via the network 102. The acquired image is sent to the detection unit 305. The image acquired by the acquisition unit 304 does not necessarily have to be the live video of the imaging device 101, and may be, for example, a past recorded video. In that case, the recorded video stored in a storage unit (not shown) in the imaging device 101 or an external storage device (server, not shown) is acquired via the network 102.

[0020] The detection unit 305 performs a person detection process on the received image to detect a person in the image. The person detection process involves identifying the position of the person's body, head, or face within the image through image analysis. Known methods can be used for the person detection process, and it may be a pattern processing method using machine learning or a rule-based image processing method. The person detection process may also be performed for each of the multiple frames acquired from the imaging device 101. The position information of the person detected by the detection unit 305 is then sent to the tracking unit 306 while maintaining the time series between frames. Depending on the type of person detection process, the position of the person's body, head, or face may be detected, but basically, the detected position is considered to be the position of that person. One way to determine the detected position is, for example, to consider the center of the detection frame on the image as the detected position of that person.

[0021] The tracking unit 306 generates trajectory information of a person based on the position information and time series for each frame detected by the detection unit 305. Trajectory information is information that shows the change in a person's position between frames and indicates the person's trajectory. Trajectory information is generated as time-series data by, for example, associating the detected time or frame number with the detected position and storing it. If multiple people are detected, trajectory information may be generated for each person, and each may be generated as being associated with a unique identifier (ID). In each generated trajectory information, the position where the person was first detected is called the starting position, and the position where the person was last detected is called the ending position. Since trajectory information is time-series data, the first data basically indicates the starting position and the last data indicates the ending position, but labels may be added to identify the data indicating the starting position and ending position. The data generated by the tracking unit 306 is sent to the storage unit 307.

[0022] The storage unit 307 has the function of storing the trajectory information sent from the tracking unit 306, and the storage unit 307 stores the trajectory information in the RAM 202 or a storage device not shown. Trajectory information is stored for each person, and stores the trajectory of the person on the image. Although the trajectory information is time-series data, the temporal order between the trajectory information does not need to be considered, and it is sufficient that only the trajectory that the person took on the image is stored. For example, even if the trajectory information of person B is generated after a sufficient amount of time has passed since the trajectory information of person A was generated, the order between them is not considered and they can be treated as the same trajectory information. In other words, there does not need to be any difference in the trajectory information stored by the storage unit 307 compared to the case where person A and person B are detected at the same time and their trajectory information is generated at the same time.

[0023] The determination unit 303 determines information about the passage detection line, such as its position, direction, and length, based on the trajectory information held by the holding unit 307. The passage detection line is a reference line used by the detection unit 308, which will be described later, to detect the passage of a person. The passage detection line is a line set on the image and can be set to various shapes such as a straight line, a polyline, or a curve. The specific method by which the determination unit 303 determines the passage detection line will be described later. The information about the passage detection line determined by the determination unit 303 is set by the setting unit 302 and stored, for example, in RAM 202 or ROM 203.

[0024] The detection unit 308 detects whether a person in the image has passed through a detection line based on the passage detection line determined by the determination unit 303. More specifically, it observes changes in the positional information of the person detected in the image and determines whether or not they have passed through a passage detection line set on the image. The detection unit 308 may also have a function to count the number of people who have passed through the passage detection line, and the counted number is displayed, for example, on the display device 104. The detection unit 308 may also identify the direction in which the person in the image passed through the passage detection line. This makes it possible to count people who have entered and exited a particular room by comparing, for example, a person who passed through the passage detection line from a first direction with a person who passed through from a second direction different from the first direction. By comparing the counting results of people who have entered and people who have exited, it is also possible to count the number of people in a particular room. It should be noted that known methods can also be used for this passage detection method. For example, as an example of pattern processing using machine learning, it is possible to use an RNN (Recurrent Neural Network), which is suitable for time-series data, to track a person's trajectory and determine whether or not they have crossed a passage detection line.

[0025] Furthermore, the functions of the detection unit 305 and the tracking unit 306 among the various functional blocks of the information processing device 103 do not necessarily need to be incorporated into the information processing device 103. For example, the imaging device 101 may perform edge analysis processing such as person detection and tracking processing on the captured image. In that case, the software and analysis model that perform the person detection and tracking processing may be built on the cloud, and the imaging device 101 may receive the analysis results through the software on the cloud. If the functions of the detection unit 305 and the tracking unit 306 are provided externally, the acquisition unit 304 in the information processing device 103 will receive person trajectory information or person detection results along with the image.

[0026] (Operation instructions) Figures 4 and 5 are flowcharts showing the operation of the information processing device according to this embodiment. Each operation in the flowcharts shown in Figures 4 and 5 is realized by the CPU 201 loading a program stored in the ROM 203 and executing it using the RAM 202 as a workspace.

[0027] 《Collection and processing of trajectory information》 The flowchart shown in Figure 4 illustrates the operation of the information processing device according to this embodiment, specifically the operation of collecting trajectory information of a person detected on an image.

[0028] In step S401, the acquisition unit 304 acquires the image captured by the imaging device 101. In step S402, the detection unit 305 performs a person detection process and detects the position of the person's body, head, or face in the image acquired by the acquisition unit 304. The detected person's position information is sent to the tracking unit 306.

[0029] In step S403, the tracking unit 306 generates trajectory information of the person. As mentioned above, trajectory information is generated for each person detected by the detection unit 305, and the generated trajectory information is stored by the storage unit 307.

[0030] Step S405 determines whether the number of trajectory information records held by the holding unit 307 is sufficient. If it is determined that the number of trajectory information records is sufficient, the trajectory information collection process ends. If it is determined that the number is not sufficient, the generation and retention of trajectory information continues. Whether the number of trajectory information records is sufficient may be determined by whether a predetermined threshold has been reached, or by responding to a user's stop instruction.

[0031] 《Process for determining the passage detection line》 The flowchart shown in Figure 5 illustrates the operation of the information processing device according to this embodiment, specifically the operation of determining information regarding the passage detection line to be set on the image. Note that the operation shown in Figure 5 is performed after the trajectory information collection process shown in Figure 4 has been completed.

[0032] In step S501, the trajectory information held by the holding unit 307 is acquired. That is, the trajectory information stored in the ROM 203 or a storage device (not shown) is read and acquired. If no trajectory information exists, it is desirable to terminate this process and prompt the user to manually set the passage detection line, or to prompt the execution of the trajectory information collection process.

[0033] In step S502, noise information included in the trajectory information is removed. The trajectory information may include false detections of people or trajectories that do not need to be detected as passing through. Processing is performed to remove such noise. Noise information includes, for example, a person entering a door but immediately returning and exiting through the door, and it is desirable to remove trajectories of people moving in a direction different from normal movement. This can be done based on the endpoint position of the trajectory or the direction from the start position to the end position of the trajectory. It is also desirable to remove trajectories where the tracking of a person by the tracking unit 306 is interrupted midway, or where objects other than people are falsely detected for several frames. This can be done based on the distance from the start position to the end position of the trajectory, or the tracking time and number of frames. Specifically, if the distance from the start position to the end position is extremely short, or if the tracking time or number of frames is extremely small, the trajectory information should be removed.

[0034] In step S503, the start and end point information of each trajectory is extracted. As mentioned above, the start and end point are, in principle, the first and last data points of the trajectory information, but this is not always the case depending on the data format, so it is desirable to label the data indicating the start and end point in advance. If there is a sufficient number of trajectory information to set up passage detection lines, a sufficient number of start and end point points for the trajectories will be extracted accordingly to set up passage detection lines.

[0035] In step S504, auxiliary lines for determining the passage detection line are derived based on the extracted start and end point positions. Specifically, a first auxiliary line is derived connecting the highest starting point among the multiple extracted starting points on the image and the highest ending point among the multiple extracted ending points on the image. Furthermore, a second auxiliary line is derived connecting the lowest starting point among the multiple extracted starting points on the image and the lowest ending point among the multiple extracted ending points on the image. Here, the height of the starting and ending points is defined as being higher towards the top edge of the image and lower towards the bottom edge of the image in a coordinate system on the image plane. That is, the first auxiliary line is a straight line connecting the starting and ending points closest to the top edge of the image, and the second auxiliary line is a straight line connecting the starting and ending points closest to the bottom edge of the image. The method for selecting the start and end points used to determine the first and second auxiliary lines is not limited to this; the leftmost or rightmost point (the point closest to the left or right edge of the image) or a point in a specific area may also be used. Here, the specific area is an area specified by the user, and the user specifies the area on the image through which a person passes, based on the stored trajectory information. At least a portion of the edges constituting the user-specified area can be selected as the first or second auxiliary line. The first and second auxiliary lines may be derived by the determination unit 303 or by user instruction. However, when derived by user instruction, the trajectory information on the image is superimposed on the image and displayed on the display device 104 to present the trajectory information to the user. Furthermore, the user is guided to specify an area enclosing all the trajectory information on the superimposed image, or to draw the first and second auxiliary lines. The determination unit 303 then derives the first and second auxiliary lines in accordance with the user instruction.

[0036] In step S505, the determination unit 303 determines the passage detection line based on the auxiliary lines derived in step S504. For example, one method is to determine the passage detection line as a straight line that intersects the first and second auxiliary lines. In particular, if the passage detection line is a straight line passing through the center points of the two auxiliary lines, the passage detection line can be uniquely determined. However, this method is not limited to this, and the passage detection line can be determined based on the auxiliary lines from various perspectives. The length of the passage detection line may be determined as a line segment connecting the center points of the two auxiliary lines, or as a straight line extending that line segment to the edge of the image. There is also a method of determining the passage detection line based on the start and end points without determining the auxiliary lines, but this will be described later in Modification 1 and Modification 2. The information regarding the passage detection line determined by the determination unit 303 is sent to the setting unit 302, and the setting unit 302 sets the passage detection line for the captured image. The set information regarding the passage detection line is referenced by the detection unit 308 and used for the person passage detection process.

[0037] Figure 6 shows a schematic diagram of the method for determining the passage detection line described above. An entrance / exit for a person is provided in the center of the image, and the area around the entrance / exit is being imaged by the imaging device 101. The dashed arrows indicate the trajectory information of the person. The dashed straight lines indicate auxiliary lines, and the solid lines indicate passage detection lines.

[0038] Here, three pieces of trajectory information exist. From these trajectory pieces, a first auxiliary line is derived connecting the highest-positioned starting point and the highest-positioned ending point, and a second auxiliary line is derived connecting the lowest-positioned starting point and the lowest-positioned ending point. A straight line passing through the center point of the derived first and second auxiliary lines is then set on the image as a passage detection line. In this way, by determining the passage detection line based on the person's trajectory information, it is possible to set a passage detection line that can detect the passage of a person without fail.

[0039] (Variation 1) Here, we will describe a modified method for setting a passage detection line based on a person's trajectory information, which differs from the method described above. Figure 7 is a flowchart showing the operation of the information processing device according to this embodiment. The same reference numerals are used for operations similar to those in the flowchart of Figure 5, and their explanations are omitted. Similarly, the flowchart in Figure 7 is realized by the processor executing a program stored in memory.

[0040] In step S701, the determination unit 303 calculates the start range and end range based on the start and end positions extracted in step S504.

[0041] In step S702, the determination unit 303 determines the passage detection line based on the calculated start and end ranges. More specifically, the determination unit determines the passage detection line so that it passes between the start and end ranges.

[0042] Here, the method for calculating the start and end ranges, as well as the method for determining the passage detection line, will be explained in more detail with reference to Figures 8 and 9.

[0043] Figure 8 shows the method for determining the start and end point ranges and the passage detection line. Figure 8(a) is the original image used to determine the start and end point ranges and the passage detection line, and the trajectory information of a person is shown on the image.

[0044] As shown in Figure 8(b), the starting point range is the range in which a person's trajectory information may contain a starting point, and the ending point range is the range in which a person's trajectory information may contain an ending point. Basically, the starting point range is calculated by the lines connecting the starting points of the trajectory information, and the ending point range is calculated by the lines connecting the ending points of the trajectory information. Statistically, since the ending point of an image in trajectory information is often at the edge of the image, it is desirable to define the ending point range as a line connecting the ending points that traces the edge of the image. Alternatively, the ending point range may be defined as a line directly connecting the ending points. Furthermore, while the starting point range and ending point range indicate the range in which the starting point and ending point set based on the trajectory information may exist, the starting point range and ending point range can be expanded or contracted according to user instructions.

[0045] Here, the passage detection line is determined to separate the start and end point ranges. In other words, based on the positions of the start and end point ranges, the position and direction (angle) of the passage detection line are determined so that it is located between the start and end point ranges. In this way, by statistically identifying the ranges where the start and end points can exist from the trajectory information and placing the passage detection line between them, it becomes possible to determine a more accurate passage detection line.

[0046] (Modification 2) Figure 9 shows the method for determining the passage detection line when the trajectory information of the person in Figure 8 is further increased. Figure 9(a) shows the trajectory information, and Figure 9(b) shows the start point range, end point range, and passage detection line calculated and determined based on the trajectory information. Compared to Figure 8(a), Figure 9(a) shows an even greater variety of trajectory information.

[0047] As shown in Figure 9(b), even when the amount of trajectory information increases, it is possible to perform appropriate passage detection settings based on statistical data of the person's trajectory by deriving the start point range and the end point range and setting the passage detection line to be placed at the position between them.

[0048] <Embodiment 2> The information processing device according to this embodiment is characterized by updating the settings of the passage detection line while performing passage detection. In Embodiment 1, trajectory information was stored in advance, and then the passage detection line was determined. However, if a person moves along a trajectory that was not anticipated when the trajectory information was collected after the passage detection line has been determined, the passage detection line may not be able to detect the person's passage. Therefore, in this embodiment, a configuration is described in which the passage detection line is updated during the execution of passage detection so that the passage of a person can be detected with higher accuracy. Note that the device configuration and functional configuration of the information processing device according to this embodiment are the same as in Embodiment 1, so a description is omitted.

[0049] Figure 10 is a flowchart showing the operation of the information processing device according to this embodiment. The operation shown in this flowchart is performed by software and is realized by the execution of a program stored in memory by the processor, similar to Embodiment 1.

[0050] In step S1001, an image captured by the imaging device 101 is acquired. In step S1002, a process is executed to detect a person from the image acquired by the acquisition unit 304. In step S1003, a process is executed to track the detected person. The processes from step S1001 to step S1003 are the same as in Embodiment 1.

[0051] In step S1004, information regarding the passage detection line set by the setting unit 302 is acquired. In step S1005, the detection unit 308 uses the acquired information regarding the passage detection line to perform passage detection processing to determine whether the person being tracked by the tracking unit 306 has passed the passage detection line. Passage detection is performed, for example, by determining whether the trajectory resulting from the tracking process has crossed the passage detection line. If passage is detected, additional processing may be performed, such as saving a log, notifying the user, or counting the number of detected passages.

[0052] In step S1006, the trajectory information obtained through the tracking process is saved. The trajectory information is saved by the storage unit 307 and stored in the ROM 203 or a storage device (not shown). In step S1007, unnecessary noise information is removed from the trajectory information held by the storage unit 307. Note that the operation of steps S1006 and S1007 is the same as the operation shown in Embodiment 1.

[0053] In step S1008, it is determined whether it is time to update the detection line and whether there is sufficient trajectory information data. The detection update timing may be set in advance by the user. For example, it may be specified by time, such as every 10 minutes, or the update may be triggered when the number of detected passages or the number of added trajectory information (based on the number of trajectory information used in the latest passage detection line setting) exceeds a predetermined threshold. Alternatively, the timing may not be specified, and only whether there is sufficient trajectory information data may be determined. Whether there is sufficient trajectory information data is determined in the same way as in step S405 in Embodiment 1. If it is determined that there is not sufficient trajectory information data, the process returns to step S1001, and if there is sufficient trajectory information data, the process proceeds to step S1009.

[0054] The operations of steps S1009 to S1011 are the same as those of steps S503 to 505 in Embodiment 1, so their explanation will be omitted. In this embodiment, S1010 is the operation to determine the auxiliary line, but as shown in Modification 1 and Modification 2, the start point range and end point range may be calculated. In that case, in step S1011, the passage detection line is determined based on the start point range and end point range. In step S1012, the information regarding the passage detection line set in the setting unit 302 is updated according to the information regarding the passage detection line determined in step S1011.

[0055] By doing so, it becomes possible to perform passage detection processing while appropriately updating the passage detection line in response to changes in the accumulated trajectory information.

[0056] <Embodiment 3> The information processing device according to this embodiment aims to set passage detection lines that can accurately detect passage even when there are multiple doors. To this end, it is characterized by storing trajectory information for each group and determining the passage detection line based on the trajectory information for each group.

[0057] When an image captured contains multiple doors or gates, and people enter and exit through each door, the starting point of the trajectory differs significantly depending on the door. Therefore, even when determining the passage detection line using the methods of Embodiments 1 and 2, trajectories that cannot detect passage may appear. In this embodiment, a method for determining the passage detection line to a position and length that can detect passage with higher accuracy, even when there are multiple doors, will be described. Note that the device configuration and functional configuration of the information processing device according to this embodiment are the same as those of Embodiments 1 and 2, so a description will be omitted.

[0058] The operation of the information processing device according to this embodiment is the same as that shown in Embodiment 1 and Embodiment 2, but the process of retaining trajectory information and the process of determining passage detection lines differ in some respects. The process of dividing trajectory information into multiple groups and the process of determining passage detection lines will be described below with reference to Figure 11.

[0059] Figure 11(a) shows a method for grouping trajectory information. Unlike Embodiments 1 and 2, Figure 11(a) shows a scene with three entrances and exits being captured by the imaging device 101. The dashed arrows indicate the trajectory information of people.

[0060] The lateral position of the starting point of the trajectory information differs depending on which door the person enters through. Therefore, the trajectory information is divided into three groups based on the entrance / exit. The grouping is shown by the ellipse in Figure 11(a). One method of grouping is for the user to input the number of doors and select the starting point of the trajectory belonging to each group. Alternatively, the grouping can be done automatically using mathematical classification methods such as non-hierarchical cluster analysis based on the lateral coordinates of the starting point of the trajectory. In other words, the grouping can be done based on the distance between the starting point positions in each piece of trajectory information. This is based on the assumption that the distance between the starting point positions of trajectory information of people who entered through the same door is shorter than the distance to the starting point positions of trajectory information of people who entered through other doors.

[0061] Next, the process for determining the auxiliary lines and passage detection lines will be explained with reference to Figure 11(b). First, auxiliary lines are derived based on the trajectory information for each group. The method for determining the auxiliary lines is the same as in Embodiment 1 and Embodiment 2. Therefore, when auxiliary lines are derived for three groups, there will be six auxiliary lines.

[0062] When determining the passage detection line, it is determined based on these six auxiliary lines. For example, the passage detection line is defined as a straight line that passes through as many of the auxiliary lines as possible. More specifically, one method is to set the passage detection line as the straight line that is the shortest distance from the center point of each auxiliary line using the least squares method. As shown in Figure 11(b), it is also possible to define the passage detection line as a single straight line that passes through all the auxiliary lines, or to find the line segments that pass through each group of auxiliary lines and derive a line segment that connects them. In that case, the passage detection line may be a polyline or a curve instead of a straight line.

[0063] Furthermore, this embodiment can also be modified as shown in Modification 1 and Modification 2, by determining the endpoint range and starting point range for each group and setting passage detection lines to separate all starting point ranges and endpoint ranges.

[0064] In this way, by deriving auxiliary lines for each group and determining passage detection lines based on the derived auxiliary lines, it is possible to find passage detection lines that intersect the trajectories of all individuals, thereby enabling the setting of more accurate passage detection lines.

[0065] <Other Embodiments> The present invention can be realized by a process of reading and executing a program that implements one or more of the functions of Embodiment 1 described above. This program is supplied to a system or device via a network or storage medium, read and executed by one or more processors in the computer of that system or device. It can also be realized by a circuit (e.g., an ASIC) that implements one or more functions. [Explanation of Symbols]

[0066] 101 Imaging device 102 Network 103 Information Processing Device 301 Input section 302 Settings Section 303 Decision Section 304 Acquisition Department 305 Detection unit 306 Tracking part 307 Holding part 308 Detection Unit

Claims

1. Image acquisition unit, A holding unit that holds the trajectory information of the person in the aforementioned image, The acquisition unit has a determination unit that determines a passage detection line for detecting the passage of a person in the image acquired by the acquisition unit, The information processing apparatus is characterized in that the determination unit determines the passage detection line based on at least one trajectory information held in the holding unit.

2. A detection unit for detecting a person in an image acquired by the acquisition unit, It includes a tracking unit that generates trajectory information of a person detected by the detection unit, The information processing apparatus according to claim 1, characterized in that the holding unit holds the trajectory information generated by the tracking unit.

3. The at least one trajectory information held by the holding unit includes a starting position indicating the first location where the person was detected and an ending position indicating the last location where the person was detected. The information processing apparatus according to claim 1, characterized in that the determination unit determines the passage detection line based on the starting position and the ending position.

4. The determination unit calculates auxiliary lines for determining the passage detection line based on the starting position and the ending position, The information processing apparatus according to claim 3, characterized in that the passage detection line is determined based on the auxiliary line.

5. The aforementioned determination unit, From the plurality of trajectory information held in the holding unit, a first auxiliary line is formed connecting the starting point closest to the top edge of the image and the ending point closest to the top edge of the image, From the multiple trajectory information held in the holding unit, a second auxiliary line is calculated connecting the starting point closest to the bottom edge of the image and the ending point closest to the bottom edge of the image, among the multiple starting point and multiple ending point positions obtained. The information processing apparatus according to claim 1, characterized in that the passage detection line is determined based on the first auxiliary line and the second auxiliary line.

6. The information processing apparatus according to claim 5, characterized in that the determination unit determines a straight line passing through the center point of the first auxiliary line and the center point of the second auxiliary line as the passage detection line.

7. The at least one trajectory information held by the holding unit includes a starting position indicating the first location where the person was detected and an ending position indicating the last location where the person was detected. The determination unit calculates a starting range indicating the range in which the starting position may exist on the image based on the starting position, and calculates an ending range indicating the range in which the ending position may exist on the image based on the ending position. The information processing device according to claim 1, characterized in that it determines the passage detection line based on the starting range and the ending range.

8. The information processing apparatus according to claim 7, characterized in that the determination unit determines a passage detection line on the image so as to separate the starting point range and the ending point range.

9. The aforementioned storage unit stores multiple trajectory information of a person, divided into multiple groups. The information processing apparatus according to claim 1, characterized in that the determination unit determines the passage detection line based on the trajectory information held for each group.

10. Each of the plurality of trajectory information held by the holding unit includes a starting position indicating the first location where the person was detected and an ending position indicating the last location where the person was detected. The information processing device according to claim 9, characterized in that the plurality of trajectory information is grouped based on the distance between the starting points of the plurality of trajectory information.

11. The information processing apparatus according to any one of claims 1 to 10, characterized in that the determination unit determines the length and angle of the passage detection line on the image.

12. The information processing apparatus according to any one of claims 1 to 10, further comprising a detection unit that detects the passage of a person based on an image acquired by the acquisition unit and a passage detection line determined by the determination unit.

13. The information processing device according to claim 12, characterized in that the detection unit counts the number of people detected by the detection unit.

14. The process of acquiring images, A retention step of retaining the trajectory information of the person in the aforementioned image, The process includes a determination step for determining a passage detection line for detecting the passage of a person in the image acquired in the acquisition step, The method is characterized in that the determination step determines the passage detection line based on at least one trajectory information held in the holding step.

15. A program for causing a computer to perform the method described in claim 14.

16. A computer-readable storage medium storing the program described in claim 15.

Images