Information processing device, information processing method, and program

Abstract

Develop technology to calculate the time required for travel between two points. [Solution] The information processing device disclosed here comprises an acquisition unit, a first detection unit, a second detection unit, a unit for calculating per moving object, a determination unit, and a statistical value calculation unit. The acquisition unit acquires a plurality of first images taken at a first location at different timings and a plurality of second images taken at a second location at different timings. The first detection unit detects a plurality of moving objects from the plurality of first images as target moving objects. The second detection unit detects each of the plurality of target moving objects from the second images. The unit for calculating per moving object calculates the travel time between the first location and the second location for each target moving object based on the timing of the first and second images in which each target moving object was detected. The determination unit determines a weight coefficient for each target moving object according to a predetermined rule. The statistical value calculation unit calculates a statistical value of the travel time between the first location and the second location based on the travel time and weight coefficient for each target moving object.

Description

Technical Field

[0001] This disclosure relates to an information processing apparatus, an information processing method, and a program.

Background Art

[0002] There is a technology for calculating the time required for movement between two points. The related technology is disclosed in Patent Document 1. Patent Document 1 calculates the time required for movement between two points by dividing the distance between the two points by the speed (representative value) of the object.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] An example of the object of this disclosure is to develop a technology for calculating the time required for movement between two points.

Means for Solving the Problems

[0005] According to one aspect of an example of this disclosure, acquisition means for acquiring a plurality of first images captured at different timings at a first point and a plurality of second images captured at different timings at a second point; first detection means for detecting a plurality of moving objects as processing target moving objects from among the plurality of first images; second detection means for detecting each of the plurality of processing target moving objects from among the second images; moving object-by-moving object calculation means for calculating, for each of the processing target moving objects, the movement time between the first point and the second point based on the shooting timings of the first images and the second images in which each of the processing target moving objects is detected; A determination means for determining the weight coefficient of each of the moving objects to be processed according to a predetermined rule, A statistical value calculation means for calculating a statistical value of the travel time between the first point and the second point based on the travel time of each of the moving objects to be processed and the weight coefficient of each of the moving objects to be processed, An information processing device having the following is provided.

[0006] Furthermore, according to one aspect of this example disclosure, One or more computers, Multiple first images taken at different times at the first location and multiple second images taken at different times at the second location are acquired. Multiple moving objects are detected from among the multiple first images as the moving objects to be processed. Each of the multiple moving objects to be processed is detected from the second image, Based on the timing of the capture of the first and second images in which each of the moving objects to be processed is detected, the travel time between the first point and the second point is calculated for each of the moving objects to be processed. In accordance with the prescribed rules, the weight coefficient for each of the moving objects to be processed is determined, An information processing method is provided for calculating a statistical value of the travel time between the first point and the second point, based on the travel time of each of the moving objects to be processed and the weight coefficient of each of the moving objects to be processed.

[0007] Furthermore, according to one aspect of this example disclosure, Computers, Acquisition means for acquiring multiple first images taken at a first location at different timings and multiple second images taken at a second location at different timings. A first detection means for detecting multiple moving objects as target moving objects from among multiple first images, A second detection means for detecting each of the multiple moving objects to be processed from the second image, A calculation means for each of the moving objects to be processed, which calculates the travel time between the first point and the second point for each of the moving objects to be processed, based on the timing of the capture of the first and second images in which each of the moving objects to be processed is detected. A determination means for determining the weight coefficient of each of the moving objects to be processed according to a predetermined rule, A statistical value calculation means that calculates a statistical value of the travel time between the first point and the second point based on the travel time of each of the moving objects to be processed and the weight coefficient of each of the moving objects to be processed. A program is provided to enable it to function as such. [Effects of the Invention]

[0008] According to this example of disclosure, it is possible to develop technology for calculating the time required to travel between two points. [Brief explanation of the drawing]

[0009] [Figure 1] Figure 1 shows an example of a functional block diagram of an information processing device. [Figure 2] Figure 2 is a flowchart showing an example of the processing flow of an information processing device. [Figure 3] Figure 3 is a diagram illustrating an example of processing performed by an information processing device. [Figure 4] Figure 4 shows an example of the hardware configuration of an information processing device. [Figure 5] Figure 5 is a schematic diagram illustrating an example of the information processed by an information processing device. [Figure 6] Figure 6 is a diagram illustrating another example of processing performed by an information processing device. [Figure 7] Figure 7 is a flowchart showing another example of the processing flow of an information processing device. [Figure 8] Figure 8 is a flowchart showing another example of the processing flow of an information processing device. [Figure 9] Figure 9 is a flowchart showing another example of the processing flow of an information processing device. [Figure 10]FIG. 10 is a flowchart showing another example of the processing flow of the information processing apparatus. [Figure 11] FIG. 11 is a diagram for explaining another example of the processing of the information processing apparatus. [Figure 12] FIG. 12 is a diagram for explaining another example of the processing of the information processing apparatus.

Embodiments for Carrying Out the Invention

[0010] Hereinafter, embodiments of this disclosure will be described with reference to the drawings. In this disclosure, the drawings are associated with one or more embodiments. Also, in all the drawings, the same reference numerals are given to the same components, and the description thereof will be omitted as appropriate.

[0011] <<First Embodiment>> FIG. 1 is a functional block diagram showing an overview of the information processing apparatus 10. FIG. 2 is a flowchart showing an example of the processing flow executed by the information processing apparatus 10.

[0012] As shown in FIG. 1, the information processing apparatus 10 includes an acquisition unit 11, a first detection unit 12, a second detection unit 13, a moving body calculation unit 14 for each moving body, a determination unit 15, and a statistical value calculation unit 16. By these functional units, the processing shown in the flowchart of FIG. 2 is executed.

[0013] In S10, the acquisition unit 11 acquires a plurality of first images taken at different timings at the first location and a plurality of second images taken at different timings at the second location. In S11, the first detection unit 12 detects a plurality of moving bodies from among the plurality of first images as processing target moving bodies. In S12, the second detection unit 13 detects each of the plurality of processing target moving bodies from the second images. In S13, the moving body calculation unit 14 for each moving body calculates the moving time between the first location and the second location for each processing target moving body based on the imaging timings of the first images and the second images in which each processing target moving body is detected. In S14, the determination unit 15 determines the weight coefficient for each of the moving objects to be processed according to predetermined rules. In S15, the statistical value calculation unit 16 calculates a statistical value of the travel time between the first point and the second point based on the travel time of each object to be processed and the weight coefficient of each object to be processed.

[0014] Note that the process flow shown in the flowchart in Figure 2 is merely an example, and the order of processes can be changed as long as similar effects can be achieved.

[0015] By the way, the information processing device 10 performs the following processing in order to calculate the travel time between the first point and the second point (hereinafter sometimes simply referred to as "travel time").

[0016] As shown in Figure 3, the information processing device 10 detects a moving object (the moving object to be processed) from a first image taken at a first location. In the example in Figure 3, the moving object is a person. Then, the information processing device 10 detects the same moving object to be processed from a second image taken at a second location.

[0017] The process of detecting the target moving object from the second image is performed, for example, by comparing the appearance features of the target moving object detected from the first image with the appearance features of the moving object detected from the second image. In one example, if the similarity between the appearance features of the target moving object detected from the first image and the appearance features of the moving object detected from the second image is greater than or equal to a threshold, the information processing device 10 detects the moving object detected from the second image as the target moving object.

[0018] The information processing device 10 then calculates the difference in the timing of capturing the first image and the second image of the detected moving object as the travel time between the first and second points of the moving object. The information processing device 10 calculates the travel time for each of multiple moving objects using this process, and calculates a statistical value of the multiple calculated travel times as an estimate of the travel time between the first and second points. By using the statistical value of the travel time for each of multiple moving objects as an estimate of the travel time, rather than using the travel time of a single moving object as an estimate, the reliability of the travel time estimate is increased.

[0019] In this type of processing, the accuracy of detecting the target moving object from the first image into the second image can affect the reliability of the estimated movement time calculated. If this accuracy is low, there is a possibility that a different moving object may be mistakenly detected as the target moving object in the second image. In this case, the movement time calculated for that target moving object will be incorrect. As a result, the reliability of the estimated movement time (statistical value) calculated using such an incorrect movement time will be low.

[0020] One possible solution to this problem of reduced reliability is to set a higher similarity threshold in the matching process for detecting the target moving object from the second image described above. In other words, one could set a higher similarity threshold for determining that two objects are the same. However, if this method is adopted, the number of objects determined to be the same will decrease. That is, the number of target moving objects detected from the second image will decrease. As a result, the number of samples used to calculate the statistical value of movement time will decrease, and the reliability of the calculated estimate (statistical value) of movement time will decrease.

[0021] The information processing device 10 has a configuration that can solve these problems. In this embodiment, the restriction of "setting a high similarity threshold for determining that two moving objects are the same" is not imposed. Therefore, a threshold is set that allows a sufficient number of moving objects to be processed to be detected from the first and second images. In this case, the number of samples for calculating the statistical value of movement time becomes sufficient. As a result, the reliability of the estimated movement time (statistical value) is increased.

[0022] However, if the restriction of "setting a high similarity threshold for determining the same moving object" is not imposed, the inconvenience of mistakenly identifying different moving objects as the same object may occur. In other words, the inconvenience of mistakenly detecting different moving objects as the target moving object in the second image may occur. Furthermore, the samples (movement times) used to calculate statistical values ​​may include samples that may contain incorrect information. If multiple such samples are treated equally to calculate an estimate (statistical value) of movement time, the reliability of the result will be low.

[0023] Therefore, the information processing device 10 can assign weights to multiple samples and calculate an estimated (statistical) travel time considering these weights. Specifically, the information processing device 10 determines a weight coefficient for each of the multiple moving objects according to a predetermined rule. The predetermined rule determines a relatively large weight coefficient for moving objects that have a relatively high reliability in the matching process that detects the target moving object from the second image.

[0024] The information processing device 10 then uses this weighting coefficient in the process of calculating statistical values ​​for the travel time of each of the multiple target moving objects. Specifically, the information processing device 10 increases the weight of the travel time of target moving objects with relatively large weighting coefficients, and decreases the weight of the travel time of target moving objects with relatively small weighting coefficients, in order to calculate statistical values ​​for the travel time.

[0025] With this information processing device 10, it is possible to calculate an estimated travel time (statistical value) by relatively increasing the weight of samples (travel time) that have a relatively high reliability and are likely to contain correct information. Furthermore, with this information processing device 10, it is possible to calculate an estimated travel time (statistical value) by relatively decreasing the weight of samples (travel time) that have a relatively low reliability and are likely to contain incorrect information. As a result, the reliability of the calculated estimated travel time (statistical value) increases.

[0026] Such an information processing device 10 can be used to develop technology for calculating the time required to travel between two points.

[0027] <<Second Embodiment>> <Overview> The information processing device 10 of the second embodiment is a concrete implementation of the configuration of the information processing device 10 of the first embodiment. It will be described in detail below.

[0028] <Hardware Configuration> First, an example of the hardware configuration of the information processing device 10 will be described. Each functional unit of the information processing device 10 is realized by any combination of hardware and software. Those skilled in the art will understand that there are various variations in the implementation method and the device. The software includes programs that are pre-installed at the time of shipment of the device, as well as programs downloaded from recording media such as CDs (Compact Discs) or from servers on the Internet.

[0029] Figure 4 is a block diagram illustrating the hardware configuration of the information processing device 10. As shown in Figure 4, the information processing device 10 includes a processor 1A, memory 2A, input / output interface 3A, peripheral circuitry 4A, and bus 5A. Peripheral circuitry 4A includes various modules. The information processing device 10 does not necessarily have peripheral circuitry 4A. The information processing device 10 may also be composed of multiple physically and / or logically separated devices. In this case, each of the multiple devices may have the above hardware configuration.

[0030] Bus 5A is a data transmission path for the processor 1A, memory 2A, peripheral circuits 4A, and input / output interface 3A to send and receive data to and from each other. Processor 1A is a processing unit such as a CPU (Central Processing Unit) or GPU (Graphics Processing Unit). Memory 2A is a memory such as RAM (Random Access Memory) or ROM (Read Only Memory). Input / output interface 3A includes interfaces for acquiring information from input devices, external devices, external servers, external sensors, cameras, etc., and interfaces for outputting information to output devices, external devices, external servers, etc. Input / output interface 3A also includes interfaces for connecting to communication networks such as the Internet. Input devices include, for example, keyboards, mice, microphones, physical buttons, touch panels, etc. Output devices include, for example, displays, projectors, speakers, printers, mailers, etc. Processor 1A can issue commands to each module and perform calculations based on their calculation results.

[0031] <Functional Configuration> Next, the functional configuration of the information processing device 10 will be described in detail. Figure 1 is an example of a functional block diagram of the information processing device 10. As shown in the figure, the information processing device 10 includes an acquisition unit 11, a first detection unit 12, a second detection unit 13, a unit for calculating data per moving object 14, a determination unit 15, and a statistical value calculation unit 16.

[0032] The acquisition unit 11 acquires a plurality of first images taken at a first location at different timings and a plurality of second images taken at a second location at different timings.

[0033] "The first point and the second point" are the points from which the calculation of travel time between them is required. The first point and the second point are different locations from each other. The first point and the second point may be outdoors or indoors.

[0034] "Image 1" is an image taken at location 1. Location 1 is equipped with a camera that photographs location 1. Location 1 may be equipped with one camera or multiple cameras. Image 1 is generated when location 1 is photographed by the camera. The camera can detect visible light and create an image. The camera may also detect other electromagnetic waves such as infrared light and create an image.

[0035] The first camera can capture moving images. In this case, the acquisition unit 11 can acquire multiple frame images contained in the moving images as multiple first images captured at different timings at the first location. Alternatively, the first camera may capture still images at a predetermined number of timings. In this case, the acquisition unit 11 can acquire these multiple still images as multiple first images captured at different timings at the first location. The predetermined timing is, for example, the timing when a predetermined moving object is detected at the first location by a predetermined sensor.

[0036] The "second image" is an image taken at a second location. A second camera is installed at the second location to photograph the second location. There may be one second camera or multiple second cameras installed at the second location. The second image is generated when the second camera photographs the second location. The second camera can detect visible light and create an image. The second camera may also detect other electromagnetic waves such as infrared light and create an image.

[0037] The second camera can capture moving images. In this case, the acquisition unit 11 can acquire multiple frame images contained in the moving images as multiple second images taken at the second location at different timings. Alternatively, the second camera may capture still images at a predetermined number of timings. In this case, the acquisition unit 11 can acquire these multiple still images as multiple second images taken at the second location at different timings. The predetermined timing is, for example, the timing when a predetermined moving object is detected at the second location by a predetermined sensor.

[0038] The acquisition unit 11 can acquire the first image generated by the first camera and the second image generated by the second camera in real time. In this case, the information processing device 10 may be connected to the first camera and the second camera in a communicative manner. The acquisition unit 11 can acquire the first image and the second image transmitted from the first camera and the second camera. The acquisition unit 11 may acquire the first image and the second image in batch processing.

[0039] "Acquisition" includes at least one of the following: the device retrieving data or information stored in another device or storage medium (active acquisition), and the device inputting data or information output from another device into its own device (passive acquisition). Examples of active acquisition include making a request to another device and receiving a reply, and accessing and reading data from another device or storage medium. Examples of passive acquisition include receiving information that is delivered (or transmitted, push notification, etc.). Furthermore, acquisition may also involve selecting and acquiring data or information from among the received data or information, or selecting and receiving data or information that has been delivered.

[0040] The first detection unit 12 detects multiple moving objects from among multiple first images as moving objects to be processed.

[0041] A "moving object" is an object that moves. A moving object may be an object that possesses the function of moving on its own. Alternatively, a moving object may not possess the function of moving on its own, but may be an object that is moved by another object.

[0042] Examples of mobile entities that possess the function of moving on their own include, but are not limited to, people, animals, insects, vehicles (automobiles, motorcycles, bicycles, buses, trucks, ships, aircraft, submarines, etc.), and robots.

[0043] Examples of moving objects that do not possess the ability to move on their own but are moved by other objects include, but are not limited to, luggage, deliveries, goods, and manufactured products.

[0044] A set of moving objects to be detected as the target moving object is defined in advance. In one example, a person is defined as the target moving object. The first detection unit 12 detects the moving object defined as the target moving object from the first image. The first detection unit 12 can detect the target moving object from the first image based on the external characteristics of the target moving object. For example, the first detection unit 12 can detect the target moving object from the first image using an object detection model or classifier generated by machine learning.

[0045] The first detection unit 12 issues processing target mobile object identification information to the processing target mobile object detected from the first image, and can store information about the processing target mobile object linked to the processing target mobile object identification information in a predetermined storage device. The predetermined storage device may be provided within the information processing device 10, or it may be provided in an external device configured to communicate with the information processing device 10. The same premise regarding the predetermined storage device applies below.

[0046] Figure 5 schematically shows an example of information that the first detection unit 12 stores in a predetermined storage device. In Figure 5, multiple elements, including the identification information of the mobile object to be processed, the timing of passing through the first point, the timing of passing through the second point, the travel time, the weight coefficient, and the appearance features, are registered and linked to each other. The first detection unit 12 can register the identification information of the mobile object to be processed, the timing of passing through the first point, and the appearance features from among this information and link them to each other.

[0047] "Processing target mobile object identification information" is information that identifies multiple processing target mobile objects detected from the first image from each other.

[0048] The "timing of passing the first location" indicates the timing when each object to be processed passes the first location. The first detection unit 12 can register the timing of capturing the first image in which each object to be processed is detected as the timing of passing the first location. The capturing timing is indicated by date and time information. The capturing timing of each first image may be specified by image metadata or by other means. For example, if the acquisition unit 11 acquires the first image in real time, the acquisition date and time of each first image may be treated as the capturing timing of each first image.

[0049] Furthermore, when the first camera is capturing moving images, each object to be processed may be detected from among multiple frame images (first images). Even in this case, the first detection unit 12 issues one object identification piece for each object to be processed, and does not issue duplicate object identification pieces. In order to avoid issuing duplicate object identification pieces, it is necessary to identify the same object to be processed detected from among multiple first images. This identification can be achieved using widely known techniques. For example, the first detection unit 12 may identify the same object to be processed detected in different first images by tracking the movement of the object to be processed within the first image using widely known tracking techniques. Alternatively, the first detection unit 12 may identify the same object to be processed detected in different first images by authentication processing based on the characteristic features of the object's appearance.

[0050] When each target moving object is detected from among multiple first images, the first detection unit 12 can register a predetermined timing within the capture timing of each of the multiple first images as the first point passage timing. The predetermined timing may be, for example, the earliest timing, the latest timing, or an intermediate timing. Alternatively, the predetermined timing may be the timing at which the target moving object is detected in a pre-specified area within the first image.

[0051] "Appearance features" are the appearance features of the object being processed that appear in the first image. If the object being processed is a person, appearance features can include, for example, facial features, body features, gait features, clothing features, possession features, and attributes that can be estimated from the appearance (gender, age, nationality, etc.). Clothing features and possession features can include color features (type of color, color distribution, etc.), pattern types, and shapes. If the object being processed is not a person, the appearance features can be set to content appropriate for each object. In this case as well, appearance features can include color features (type of color, color distribution, etc.), pattern types, and shapes. Appearance features can be represented, for example, as a feature vector containing the values ​​of multiple elements.

[0052] Other elements shown in Figure 5 will be discussed later.

[0053] Returning to Figure 1, the second detection unit 13 detects each of the multiple target moving objects from the second image. The second detection unit 13 can detect each of the multiple target moving objects from the second image using the characteristic features of the appearance of each of the multiple target moving objects extracted from the first image.

[0054] In one example, the second detection unit 13 detects a moving object from the second image. The moving object detected here is of the same type as the moving object detected as the target moving object in the first image. The second detection unit 13 can detect the moving object from the second image using the same process as the process for detecting the moving object (target moving object) from the first image.

[0055] Next, the second detection unit 13 extracts feature quantities of the detected moving object's appearance from the second image. The feature quantities of the moving object's appearance extracted from the second image are of the same type as the feature quantities of the moving object's appearance extracted from the first image. The second detection unit 13 can extract the feature quantities of the moving object's appearance from the second image using the same process as the process for extracting the feature quantities of the moving object's appearance from the first image. The feature quantities of the moving object's appearance extracted from the second image can be represented, for example, as a feature vector containing the values ​​of multiple elements.

[0056] Next, the second detection unit 13 compares the feature quantities of the appearance of the moving object extracted from the second image with the feature quantities of the appearance of each of the multiple moving objects to be processed extracted from the first image.

[0057] In this matching process, for example, the second detection unit 13 calculates the similarity between the feature quantities of the appearance of the moving object extracted from the second image and the feature quantities of the appearance of each of the multiple target moving objects extracted from the first image. The calculation of the similarity is achieved using widely known techniques. Examples of techniques used for calculating the similarity include, but are not limited to, Euclidean distance, cosine similarity, and Manhattan distance.

[0058] Then, if there is a target moving object whose similarity to the appearance features of the moving object extracted from the second image is above a predetermined threshold, the second detection unit 13 detects that moving object extracted from the second image as the target moving object.

[0059] Setting the threshold higher can suppress the problem of misidentifying different moving objects as the target moving object in the second image. However, setting the threshold higher may result in the same moving object being misidentified, leading to a decrease in the number of target moving objects detected in the second image.

[0060] On the other hand, setting the threshold too low may lead to the misdetection of different moving objects as target moving objects in the second image. However, setting the threshold too low allows all target moving objects detected in the first image to be detected in the second image without fail, resulting in a larger number of target moving objects detected in the second image.

[0061] An appropriate threshold is set in advance, taking these trade-off relationships into consideration. For example, a threshold is set such that the number of moving objects to be processed detected from the second image becomes the desired number.

[0062] When the second detection unit 13 detects a moving object to be processed from the second image, it can store information about that moving object in a predetermined storage device, linking it to the identification information of that moving object.

[0063] Figure 5 schematically shows an example of information that the second detection unit 13 stores in a predetermined storage device. The second detection unit 13 can register the "second point passage timing" among the elements shown in Figure 5.

[0064] The "second point passage timing" indicates the timing when each target moving object passes the second point. The second detection unit 13 registers the timing of the capture of the second image detected by each target moving object as the second point passage timing. The capture timing is indicated by date and time information. The capture timing of each second image may be identified by image metadata or by other means. For example, if the acquisition unit 11 acquires the second image in real time processing, the acquisition date and time of each second image may be treated as the capture timing of each second image.

[0065] Furthermore, when the second camera is capturing moving images, each moving object to be processed can be detected from among multiple frame images (second images). In this case, the second detection unit 13 can register a predetermined timing among the capture timings of each of the multiple second images as the second point passage timing. The predetermined timing may be, for example, the earliest timing, the latest timing, or an intermediate timing. Alternatively, the predetermined timing may be the timing at which the moving object to be processed is detected in a pre-specified area within the second image.

[0066] Returning to Figure 1, the unit 14 for calculating the travel time between the first and second points is calculated for each mobile object to be processed, based on the timing of the capture of the first and second images detected for each mobile object.

[0067] For example, the unit 14 can identify the timing of each mobile object passing through a first point (see Figure 5), which has been registered by the first detection unit 12, as the timing of capturing the first image detected for each mobile object. Furthermore, the unit 14 can identify the timing of each mobile object passing through a second point (see Figure 5), which has been registered by the second detection unit 13, as the timing of capturing the second image detected for each mobile object.

[0068] The unit 14 can calculate the difference between the timing of each target mobile body passing a first point and the timing of passing a second point as the travel time between the first point and the second point for each target mobile body. The unit 14 can store the travel time calculated for each target mobile body in a predetermined storage device, linked to the target mobile body identification information of each target mobile body (see Figure 5).

[0069] Returning to Figure 1, the determination unit 15 determines the weight coefficient for each mobile object to be processed according to predetermined rules. The determination unit 15 can then store the weight coefficient determined for each mobile object to be processed in a predetermined storage device, linked to the mobile object identification information of each mobile object to be processed (see Figure 5).

[0070] Here, we will explain the process of determining the weight coefficients according to predetermined rules.

[0071] The determination unit 15 determines the weight coefficient of the target mobile body based on the feature quantities of the appearance of the target mobile body and the feature quantities of the appearance of each of the multiple comparison target mobile bodies. When determining the weight coefficient of the first target mobile body, which is any one of the multiple target mobile bodies, the determination unit 15 determines the weight coefficient of the first target mobile body based on the feature quantities of the appearance of the first target mobile body and the feature quantities of the appearance of each of the multiple comparison target mobile bodies. The comparison target mobile bodies are mobile bodies that are referenced when determining the weight coefficient of the target mobile body. Specific examples of comparison target mobile bodies will be described later.

[0072] The determination unit 15 makes the weight coefficient of the target mobile object located in a region where the appearance features are relatively dense in the distribution of appearance features of multiple comparison target mobile objects smaller than the weight coefficient of the target mobile object located in a region where the appearance features are relatively sparse. This process will be explained using Figure 6.

[0073] Figure 6 shows a feature distribution diagram where the feature quantities of the appearances of multiple comparison targets are plotted as black circles. In the feature distribution diagram of Figure 6, the feature quantities of the appearances of two target targets are plotted as white circles. Note that while Figure 6 shows the distribution of appearance features in a two-dimensional space, it may also be in a multi-dimensional space of three or more dimensions.

[0074] The appearance features corresponding to the white circles in Q are located in a denser region compared to the appearance features corresponding to the white circles in R. Therefore, the determination unit 15 makes the weight coefficient of the processed mobile object from which the appearance features corresponding to the white circles in Q have been extracted smaller than the weight coefficient of the processed mobile object from which the appearance features corresponding to the white circles in R have been extracted.

[0075] Furthermore, it is sufficient that the relative magnitudes of the weight coefficients of the multiple target moving objects satisfy the above-mentioned criteria, and various configurations can be used to determine the specific weight coefficients of the target moving objects.

[0076] In one example, the determination unit 15 first calculates a first value indicating whether the features of the appearance of the target mobile object are located in a relatively dense or sparse region within the distribution of feature quantities of the appearances of multiple comparison target mobile objects.

[0077] The first value can be calculated using various methods. For example, the determination unit 15 may calculate the number of comparison target mobile bodies whose similarity in appearance features to the processing target mobile body is equal to or greater than a threshold value as the first value. This similarity represents the distance between plotted positions in a feature distribution diagram as shown in Figure 6. This process is equivalent to counting the number of black circles corresponding to comparison target mobile bodies that are located within a circle of a predetermined radius centered on the white circle corresponding to each processing target mobile body shown in Figure 6. The radius of the circle corresponds to the threshold value of the similarity.

[0078] The larger this first value (count value) is, the more densely the features of the appearance of the first target moving object are located in the feature distribution map shown in Figure 6. Conversely, the smaller this first value (count value) is, the more sparsely the features of the appearance of the first target moving object are located in the distribution map shown in Figure 6.

[0079] The first value may be calculated by other means. For example, the determination unit 15 may identify a predetermined number of comparison target mobile objects from those with a high similarity in appearance features to the mobile object to be processed. The determination unit 15 can then calculate the similarity of appearance features between each of the identified predetermined number of comparison target mobile objects and the mobile object to be processed. The determination unit 15 can then calculate the statistical value of the calculated similarity as the first value. The statistical value may be the mean, maximum, minimum, mode, median, etc., but is not limited to these. A larger first value (statistical value of similarity) means that the appearance features of the mobile object to be processed are located in a denser area in the feature distribution map shown in Figure 6. A smaller first value (statistical value of similarity) means that the appearance features of the mobile object to be processed are located in a sparser area in the distribution map shown in Figure 6. The means for calculating the first value exemplified here are merely examples and are not limited to these.

[0080] After calculating the first value for each processing target mobile body, the determination unit 15 determines the weight coefficient for each processing target mobile body based on the first value for each processing target mobile body.

[0081] If the larger the first value, the denser the feature quantities of the moving object being processed are located in the feature quantity distribution map shown in Figure 6, then the determination unit 15 determines a weight coefficient that is smaller the larger the first value. In this case, the determination unit 15 determines a weight coefficient that is larger the smaller the first value.

[0082] In one example, a decision model for determining weight coefficients from a first value is pre-generated. The decision unit 15 can then use this decision model to determine the weight coefficients from the first value. The decision model may be a formula for calculating weight coefficients from the first value, or it may include a table showing the correspondence between the first value and the weight coefficients.

[0083] Next, I will explain the comparison target mobile object.

[0084] As mentioned above, the "comparison target mobile object" is a mobile object that is referenced when determining the weight coefficients of the mobile object to be processed. The comparison target mobile object is a mobile object of the same type as the mobile object to be processed.

[0085] When the determination unit 15 calculates the weight coefficient of the first target mobile body, which is any one of the multiple target mobile bodies, it can process at least one of the following first to sixth mobile bodies as the target mobile body for comparison.

[0086] (First moving object) A moving object detected from at least one of the first image and the second image taken within a predetermined time from the time when the first image in which the first object to be processed was detected is captured. (Second moving object) A moving object detected from at least one of the first image and the second image, taken within a predetermined time from the time when the second image in which the first moving object to be processed was detected was taken. (Third moving object) A moving object detected from at least one of the first image and the second image, which were taken in the same or a predetermined level or more similar shooting environment as the shooting environment of the first image in which the first object to be processed was detected. (Fourth moving object) A moving object detected from at least one of the first and second images, which were taken in the same or a predetermined level of similar shooting environment as the shooting environment of the second image in which the first moving object to be processed was detected. (Fifth moving object) A moving object detected from an image taken at a location different from the first and second locations, in a shooting environment that is the same as or at least similar to the shooting environment of the first image in which the first object to be processed was detected. (Sixth moving object) A moving object detected from images taken at a location different from the first and second locations, in a shooting environment that is the same as or at least similar to the shooting environment of the second image in which the first target moving object was detected.

[0087] The first and second mobile objects are mobile objects detected at the first and second locations at timings surrounding the detection of the first target mobile object at the first and second locations. The first target mobile object is mixed in among these first and second mobile objects.

[0088] By using such a first and second mobile object as comparison objects, and determining the weight coefficients for each target mobile object based on the relationship of similarity of the appearance features between these comparison objects, the weight coefficients can be appropriately determined.

[0089] In other words, for target mobile objects that have a high degree of similarity in appearance features with multiple comparison target mobile objects (located in a dense area in the feature distribution diagram shown in Figure 6) and are not easy to detect among multiple comparison target mobile objects, a relatively small weight coefficient can be determined. On the other hand, for target mobile objects that have a low degree of similarity in appearance features with multiple comparison target mobile objects (located in a sparse area in the feature distribution diagram shown in Figure 6) and are easy to detect among multiple comparison target mobile objects, a relatively large weight coefficient can be determined.

[0090] The third and fourth moving objects are moving objects detected from at least one of the past first and second images, which were taken in a shooting environment that is the same as or similar to a predetermined level or more to the shooting timing at which the first moving object to be processed was detected at the first and second locations.

[0091] "The first and second past images" can be, for example, images taken within a fixed period in the immediate vicinity. For example, the first and second past images may be images from the last few years, the last few months, the last few weeks, or the last few days.

[0092] "Shooting environment" is indicated by at least one of the following: location, time of day, day of the week, season, and weather.

[0093] Depending on the shooting environment defined by the elements described above, the characteristics of a person's appearance detected from a captured image may differ. For example, in a business district, there is a tendency for more people in suits to be detected, and a lower probability of people in flashy clothing being detected. Similarly, in a youth-oriented area, there is a tendency for more people in fashionable clothing to be detected, and a lower probability of people in suits being detected. Such tendencies can vary depending on the shooting environment, including location, time of day, day of the week, season, and weather.

[0094] Therefore, the characteristic features of the appearance of a moving object detected from images taken at multiple locations where the shooting environment is the same or similar to a predetermined level may show similar trends. Thus, by using the third and fourth moving objects defined above as comparison objects, and determining the weight coefficients of each processing target moving object based on the relationship of the similarity of the appearance features between these comparison objects, the weight coefficients can be appropriately determined.

[0095] In embodiments where the third and fourth mobile objects are used as comparison targets, the first and second images previously taken at the first and second locations are stored in a predetermined storage device, linked to the shooting environment of each image. The determination unit 15 then uses this information stored in the predetermined storage device to perform processing with the third and fourth mobile objects as comparison targets.

[0096] The fifth and sixth moving objects are moving objects detected from images taken at other locations in a shooting environment that is the same as or at least similar to the shooting timing at which the first target moving object was detected at the first and second locations. The other locations are different from the first and second locations.

[0097] As described above, the visual features of moving objects detected from images taken at multiple locations where the shooting environment is the same or similar to a predetermined level may show similar trends. Therefore, by using the fifth and sixth moving objects defined above as comparison objects, and determining the weight coefficients of each processing target moving object based on the relationship of the similarity of the visual features between these comparison objects, the weight coefficients can be appropriately determined.

[0098] In embodiments where the fifth and sixth mobile objects are used as comparison targets, images taken at locations other than the first and second locations are stored in a predetermined storage device, linked to the shooting environment of each image. The determination unit 15 then uses this information stored in the predetermined storage device to perform processing with the fifth and sixth mobile objects as comparison targets.

[0099] Here, we will explain the means for determining the condition that "the shooting environment is the same or similar to a predetermined level" relating to the third to sixth moving objects.

[0100] This means is implemented using widely known technology. An example is described below, but it is not limited to this example. First, the determination unit 15 quantifies (digitizes) each of the multiple elements that define the shooting environment (location, time of day, day of the week, season, and weather, etc.) according to predetermined rules, and generates an environment vector by arranging the quantified values ​​of the multiple elements. Then, if the similarity between the environment vectors is above a threshold, the determination unit 15 can determine that the two shooting environments are identical or similar to a predetermined level or higher. The similarity between environment vectors can be calculated using techniques such as Euclidean distance, cosine similarity, and Manhattan distance, but is not limited to these.

[0101] The determination unit 15 can identify the shooting environment of the first and second images in which the target moving object was detected by various means. In one example, the determination unit 15 can receive input from an operator indicating the shooting environment of the first and second images. In another example, the determination unit 15 can obtain weather information for the first and second locations at the time the first and second images were taken from an external server that provides weather information. In yet another example, the determination unit 15 can obtain information indicating the shooting location and shooting date and time from the first and second cameras.

[0102] Returning to Figure 1, the statistical value calculation unit 16 calculates a statistical value of the travel time between the first point and the second point based on the travel time of each target mobile body calculated by the per-mobile body calculation unit 14 and the weight coefficient of each target mobile body determined by the determination unit 15. For example, the statistical value calculation unit 16 can calculate a weighted average as the statistical value.

[0103] The statistical value calculation unit 16 may extract processing target moving objects that satisfy the extraction conditions from among the multiple processing target moving objects detected from the first image, and calculate a statistical value of the movement time between the first point and the second point based on the movement time and weight coefficient of the extracted processing target moving objects. The extraction conditions are, but are not limited to, "detected from the first image within a fixed time in the immediate vicinity". The predetermined time is, but is not limited to, several minutes, several tens of minutes, several hours, etc. By setting such extraction conditions, the inconvenience of using old samples in calculating the statistical value of movement time can be suppressed.

[0104] Next, an example of the processing flow of the information processing device 10 will be explained using the flowcharts in Figures 7 to 10. Note that the purpose here is to explain an example of the processing flow. Details of each process have been described above, so explanations will be omitted here as appropriate.

[0105] The flowchart in Figure 7 shows an example of the processing implemented by the acquisition unit 11 and the first detection unit 12.

[0106] When the information processing device 10 acquires a first image (S20), it detects a moving object from the first image as a target moving object (S21). The information processing device 10 then stores the information of the detected target moving object in a predetermined storage device (S22). In S22, the information processing device 10 issues target moving object identification information for the detected target moving object and stores information about that target moving object in the predetermined storage device, linked to the target moving object identification information. The information about the target moving object is the first point passage timing and appearance feature quantities in Figure 5. The information processing device 10 may leave the values ​​of elements other than the first point passage timing and appearance feature quantities in the information in Figure 5 blank. The values ​​of these elements will be filled in by a process described later.

[0107] The flowchart in Figure 8 shows an example of the processing implemented by the acquisition unit 11, the second detection unit 13, and the calculation unit 14 for each moving object.

[0108] When the information processing device 10 acquires a second image (S30), it detects a registered target mobile object from the second image (S31). A registered target mobile object is one for which a target mobile object identification information is issued in the process shown in Figure 7, and various information is stored in a predetermined storage device linked to that target mobile object identification information. The information processing device 10 detects a registered target mobile object from the second image using the characteristic features of the registered target mobile object's appearance. The information processing device 10 can store information about the target mobile object detected from the second image in a predetermined storage device linked to the target mobile object identification information of that target mobile object. This information about the target mobile object is the timing of passing through the second point in Figure 5.

[0109] Subsequently, the information processing device 10 calculates the travel time between the first point and the second point of the object to be processed detected from the second image and stores it in a predetermined storage device (S32). The information processing device 10 calculates the difference between the timing of passing the first point and the timing of passing the second point, which are stored in the predetermined storage device, as the travel time. The information processing device 10 can store the calculated travel time in a predetermined storage device, linked to the object identification information of the object to be processed (travel time in Figure 5).

[0110] The flowchart in Figure 9 shows an example of the process implemented by the decision unit 15.

[0111] The information processing device 10 determines whether there are any mobile objects among the registered mobile objects to be processed for which weight coefficients have not been determined (S40). For example, the information processing device 10 may perform the determination in S40 by referring to the information stored in a predetermined storage device as shown in Figure 5 and determining whether there is any mobile object identification information for which weight coefficients have not been registered.

[0112] If there is a mobile object among the registered mobile objects to be processed for which a weight coefficient has not been determined (Yes in S40), the information processing device 10 determines and registers the weight coefficient of that mobile object to be processed (S41). That is, the information processing device 10 determines the weight coefficient of that mobile object to be processed and stores the determined weight coefficient in a predetermined storage device, linked to the mobile object identification information of that mobile object to be processed (weight coefficient in Figure 5).

[0113] The flowchart in Figure 10 shows an example of the processing implemented by the statistical value calculation unit 16.

[0114] The information processing device 10 extracts a mobile object to be processed that satisfies the extraction criteria from among the registered mobile objects to be processed (S50). The extraction criteria are, for example, "detected from the first image within a predetermined time in the immediate vicinity," but are not limited to this. The predetermined time is, for example, several minutes, several tens of minutes, several hours, etc., but is not limited to these.

[0115] Subsequently, the information processing device 10 calculates and outputs statistical values ​​of the movement time based on the movement time and weight coefficient of the moving object to be processed extracted in S50 (S51).

[0116] The information processing device 10 can output statistical values ​​of the calculated travel time via various output devices. For example, an output device may be installed at the first location. Examples of such output devices include, but are not limited to, displays, projection devices, electronic billboards, and speakers. The information processing device 10 may also cause the output device installed at the first location to output the statistical values ​​of the calculated travel time. In this case, a person located at the first location can recognize an estimate of the time required to travel from there to the second location.

[0117] In addition, the information processing device 10 may output the calculated travel time statistics to an output device for administrators managing the facility including the first and second locations. In this case, the administrator can recognize the estimated time required to travel from the first location to the second location and, based on that information, understand the situation (congestion status, etc.) between the first and second locations. The output device in this case may be a display, projection device, electronic billboard, speaker, etc., installed in a monitoring center or the like. Alternatively, the output device in this case may be a mobile device such as a smartphone or tablet.

[0118] <Effects and Effects> According to the information processing device 10 of the second embodiment, the same effects and advantages as those of the information processing device 10 of the first embodiment can be achieved.

[0119] Furthermore, the information processing device 10 can determine the weight coefficient of the target mobile object based on the feature quantities of the object's appearance and the feature quantities of the appearance of each of the multiple comparison target mobile objects. The information processing device 10 can make the weight coefficient of the target mobile object located in a region where the feature quantities of the appearance are relatively dense in the distribution of the feature quantities of the appearance of the multiple comparison target mobile objects smaller than the weight coefficient of the target mobile object located in a region where the feature quantities of the appearance are relatively sparse.

[0120] As shown in Figure 6, a target mobile object (white circle labeled Q) located in a region where the appearance features are relatively dense in the distribution of appearance features of multiple comparison targets has a similar appearance to the other comparison targets, making it difficult to distinguish among them. The reliability of detecting such a target mobile object from the second image will be relatively low. On the other hand, a target mobile object (white circle labeled R) located in a region where the appearance features are relatively sparse in the distribution of appearance features of multiple comparison targets has a dissimilar appearance to the other comparison targets, making it easy to distinguish among them. The reliability of detecting such a target mobile object from the second image will be relatively high.

[0121] Based on this relationship, the information processing device 10 relatively reduces the weight coefficient of the processing target mobile object that is not easy to identify among the multiple comparison target mobile objects. Then, the information processing device 10 relatively increases the weight coefficient of the processing target mobile object that is easy to identify among the multiple comparison target mobile objects. By relatively increasing the weight coefficient of the processing target mobile object that is easy to identify and relatively decreasing the weight coefficient of the processing target mobile object that is not easy to identify, and calculating the statistics of movement time, it is possible to calculate statistics with high reliability.

[0122] Furthermore, the information processing device 10 can process at least one of the first to sixth mobile bodies described above as a comparison mobile body.

[0123] By using the first and second moving objects as comparison objects, the moving objects detected at the first and second locations at timings surrounding the detection timing of each target moving object at the first and second locations can be used as comparison objects. Each target moving object is mixed in with these first and second moving objects. By using these first and second moving objects as comparison objects and determining the weight coefficients of each target moving object based on the relationship of the similarity of the appearance features between these comparison objects, the weight coefficients can be appropriately determined. In other words, the weight coefficients of moving objects that can be easily identified within the set of moving objects including the first and second moving objects can be made relatively larger, and the weight coefficients of moving objects that cannot be easily identified within the set of moving objects including the first and second moving objects can be made relatively smaller. By using the weight coefficients determined in this way to calculate statistics on movement time, highly reliable statistics can be obtained.

[0124] Furthermore, by using the third to sixth moving objects as comparison moving objects, it is possible to select moving objects that tend to be present in the vicinity of each target moving object at the timing when it is detected at the first or second location as comparison moving objects. By using the third to sixth moving objects as comparison moving objects and determining the weight coefficients of each target moving object based on the relationship of the similarity of the appearance features between these comparison moving objects, the weight coefficients can be appropriately determined. In other words, the weight coefficients of moving objects that can be easily identified within the group of moving objects (the third to sixth moving objects) predicted to be present in the vicinity of each target moving object can be made relatively large. Conversely, the weight coefficients of moving objects that cannot be easily identified within the group of moving objects (the third to sixth moving objects) predicted to be present in the vicinity of each target moving object can be made relatively small. By using the weight coefficients determined in this way to calculate statistics on movement time, highly reliable statistics can be obtained.

[0125] <<Third Embodiment>> The information processing device 10 of the third embodiment differs from the second embodiment in that it has a predetermined rule for determining the weight coefficient of each target moving object. In the third embodiment, a reference feature quantity is defined in advance. The information processing device 10 then determines the weight coefficient of each target moving object based on the similarity between this reference feature quantity and the feature quantity of the appearance of each target moving object. This will be explained in detail below.

[0126] The determination unit 15 determines the weight coefficient for each mobile object to be processed based on the similarity between the feature quantities of the appearance of each mobile object to be processed and predefined reference feature quantities. The determination unit 15 makes the weight coefficient of the mobile object to be processed with a relatively high similarity smaller than the weight coefficient of the mobile object to be processed with a relatively low similarity. This process will be explained using Figure 11.

[0127] In Figure 11, the black circles represent the reference feature, and the white circles represent the appearance feature quantities of the two target moving objects, plotted in a two-dimensional coordinate system. Note that while Figure 11 plots these feature quantities in a two-dimensional space, they may also be plotted in a multi-dimensional space of three dimensions or more.

[0128] The appearance features corresponding to the white circles in Q are located closer to the black circles corresponding to the reference features than the appearance features corresponding to the white circles in R. A closer distance to the black circles means a higher degree of similarity with the appearance features corresponding to the black circles. For this reason, the decision unit 15 makes the weight coefficient of the processed mobile object from which the appearance features corresponding to the white circles in Q are extracted smaller than the weight coefficient of the processed mobile object from which the appearance features corresponding to the white circles in R are extracted.

[0129] A "reference feature" is a predefined feature. The reference feature is of the same type as the feature of the appearance of the moving object to be processed, which is extracted from the first and second images as described in the second embodiment. The reference feature can be represented, for example, as a feature vector containing the values ​​of multiple elements.

[0130] The reference features are the features of the appearance of the moving object that are common at the first and second locations. For example, an operator may determine such reference features and register them in the information processing device 10. Alternatively, the reference features may be determined by computer processing. In one example, the reference features can be determined by statistically processing the features of the appearance of the moving object detected from the first and second images taken at the first and second locations within a predetermined period in the past. For example, the reference features may be determined as appearance features whose frequency of occurrence (number of occurrences (number of occurrences) within a predetermined time period) is above a threshold.

[0131] The similarity between a reference feature and the feature of each target moving object can be calculated, for example, using a technique for calculating similarity between vectors. While such techniques as Euclidean distance, cosine similarity, and Manhattan distance can be used, they are not limited to these.

[0132] Here, a modified version of this embodiment will be described. In the modified version, multiple reference features are defined in advance and stored in a predetermined memory device. The shooting environment is registered in association with each of the multiple reference features. Each reference feature is a feature of the appearance of a moving object, which is a common appearance in images taken in the associated shooting environment.

[0133] In this modified example, the determination unit 15 identifies the shooting environment of the first and second images in which the target moving object was detected. The determination unit 15 then calculates the similarity between the shooting environments of the first and second images and the shooting environments associated with each of the pre-registered reference features. The determination unit 15 then selects the reference feature associated with the shooting environment that has the highest similarity to the shooting environments of the first and second images. Based on the similarity between the reference feature thus selected and the appearance features of each target moving object, the determination unit 15 can determine the weight coefficient for each target moving object.

[0134] Identifying the shooting environment of the first and second images, and calculating the similarity of the shooting environments, can be achieved using the method described in the second embodiment.

[0135] Other configurations of the information processing device 10 can be the same as those in the first and second embodiments.

[0136] According to the information processing device 10 of the third embodiment, the same effects and advantages as those of the information processing device 10 of the first and second embodiments can be achieved.

[0137] Furthermore, the information processing device 10 can determine the weight coefficient for each mobile object to be processed based on the similarity between the characteristic features of each mobile object's appearance and a predefined reference characteristic. Specifically, the information processing device 10 can make the weight coefficient for a mobile object to be processed with a relatively high similarity smaller than the weight coefficient for a mobile object to be processed with a relatively low similarity.

[0138] The reference features are the appearance features of a moving object that are commonplace at the first and second locations. As shown in Figure 11, moving objects to be processed (white circles labeled Q) whose appearance features have a relatively high similarity to these reference features are not easily distinguishable from other moving objects at the first and second locations. The confidence level of the result when detecting such moving objects from the second image will be relatively low. On the other hand, moving objects to be processed (white circles labeled R) whose appearance features have a relatively low similarity to the reference features are easily distinguishable from other moving objects at the first and second locations. The confidence level of the result when detecting such moving objects from the second image will be relatively high.

[0139] Based on this relationship, the information processing device 10 makes the weight coefficient of the object to be processed, which is an appearance feature with a relatively high similarity to the reference feature, smaller than the weight coefficient of the object to be processed with a relatively low similarity. By making the weight coefficient of the object to be processed that is easy to identify relatively larger and the weight coefficient of the object to be processed that is not easy to identify relatively smaller, and calculating the statistics of movement time, it is possible to calculate statistics with high reliability.

[0140] Furthermore, as shown in the modified example, the information processing device 10 can prepare multiple reference features and determine the weight coefficients for each target moving object using an appropriate reference feature selected from among them. Specifically, the information processing device 10 can select feature quantities of the appearance of each target moving object, which are common in the shooting environment of the image in which the detected target moving object is captured, through the characteristic processing described above. With such an information processing device 10, the weight coefficients for each target moving object can be appropriately determined using an appropriate reference feature according to the shooting environment at that time.

[0141] <<Fourth Embodiment>> The information processing device 10 of the fourth embodiment differs from the second and third embodiments in the predetermined rules for determining the weight coefficient of each moving object to be processed. In the fourth embodiment, if multiple moving objects moving together are detected simultaneously in both the first and second images, the weight coefficients of those moving objects are made relatively larger. This will be explained in detail below.

[0142] First, the technical concept of this embodiment will be explained. There are cases where multiple objects to be processed move together. Examples of such multiple objects to be processed include, but are not limited to, multiple people such as family members, friends, lovers, or colleagues.

[0143] Multiple moving objects to be processed that are moving together are likely to appear in both the first and second images simultaneously and be detected simultaneously from both images. When multiple moving objects to be processed that are moving together are detected simultaneously from both the first and second images, the reliability of the detection results from the first and second images is high. Therefore, based on this idea, the information processing device 10 relatively increases the weighting coefficient of the moving objects to be processed that are considered to have a high reliability in the detection results.

[0144] The determination unit 15 makes the weight coefficients of multiple target moving objects detected from the same first image and also from the same second image relatively larger than the weight coefficients of other target moving objects. In this process, multiple target moving objects detected from the same first image are identified as multiple target moving objects moving together. When multiple target moving objects moving together, identified in this way, are detected simultaneously from the same second image, the reliability of the result of detecting those multiple target moving objects from the second image is determined to be high. The weight coefficients of the multiple target moving objects with high reliability in the result detected from the second image are then relatively increased.

[0145] In one example, weight coefficients are predetermined for multiple target moving objects detected from the same first image and also from the same second image, as well as weight coefficients for other target moving objects. The determination unit 15 can then determine the weight coefficient for each target moving object according to these predetermined rules.

[0146] As another example, an additional value may be predetermined to be added to the weight coefficients of multiple target moving objects detected from the same first image and also from the same second image. The determination unit 15 may then determine the weight coefficients of each target moving object using the method described in the second and third embodiments, and then add the additional value to the target moving objects that satisfy the addition condition. The addition condition is that the object is one of multiple target moving objects detected from the same first image and also from the same second image.

[0147] Determining whether multiple target moving objects detected from the same first image and from the same second image can be achieved using various methods. For example, the first detection unit 12 links the target moving object identification information of multiple target moving objects detected from the same first image and stores it in a predetermined storage device. The second detection unit 13 also links the target moving object identification information of multiple target moving objects detected from the same second image and stores it in a predetermined storage device. The determination unit 15 then refers to this information to detect combinations of multiple target moving objects that are simultaneously detected in both the first and second images.

[0148] Now, a modified version of this embodiment will be described. In this modified version, the determination unit 15 may determine whether a plurality of moving objects to be processed that are simultaneously detected in both the first image and the second image satisfy predetermined companion conditions. Furthermore, the determination unit 15 may make the weight coefficients of the plurality of moving objects to be processed that are simultaneously detected in both the first image and the second image and satisfy the companion conditions relatively larger than the weight coefficients of the other moving objects to be processed.

[0149] In one example, weight coefficients for multiple objects to be processed that are detected simultaneously in both the first and second images and satisfy the accompanying conditions, as well as weight coefficients for other objects to be processed, are predetermined. The determination unit 15 can then determine the weight coefficient for each object to be processed according to these predetermined rules.

[0150] As another example, an additional value may be predetermined to be added to the weight coefficients of multiple processing target moving objects that are simultaneously detected in both the first and second images and satisfy the accompanying condition. The determination unit 15 may then determine the weight coefficient of each processing target moving object using the method described in the second and third embodiments, and then add that additional value to the processing target moving objects that satisfy the addition condition. The addition condition is that the object is simultaneously detected in both the first and second images and satisfies the accompanying condition.

[0151] The companion condition is a criterion for determining whether multiple moving objects to be processed, detected simultaneously in both the first and second images, are companions, that is, whether they are moving together. Multiple moving objects to be processed that satisfy the companion condition can be considered companions, i.e., moving together.

[0152] The accompanying conditions can be one or more of the following conditions connected by a predetermined logical operator. In the following conditions, multiple moving objects to be processed that are simultaneously detected in both the first and second images are referred to as "multiple moving objects to be judged". • In the first image, the distance between multiple moving objects to be judged is within the threshold. • In the second image, the distance between multiple moving objects to be judged is within the threshold. • In the first image, there is physical contact between multiple moving objects being judged. • In the second image, there is physical contact between multiple moving objects being judged. • In the first image, one of the multiple moving objects being judged is in the line of sight of another. • In the second image, one of the multiple moving objects being judged is in the line of sight of another.

[0153] The calculation of the distance between two objects in an image can be achieved using widely known image analysis techniques. The distance between two objects and its threshold may be expressed in pixels or by other methods. The presence or absence of contact between two objects in an image can be achieved using widely known image analysis techniques. The detection of a line of sight in an image, and the detection of the object at the end of the line of sight, can be achieved using widely known image analysis techniques.

[0154] Other configurations of the information processing device 10 can be the same as those of the first to third embodiments.

[0155] According to the information processing device 10 of the fourth embodiment, the same effects and advantages as those of the information processing device 10 of the first to third embodiments can be achieved.

[0156] Furthermore, as described above, the information processing device 10 can determine the weight coefficients of multiple moving objects to be processed using a distinctive method that assumes the case where multiple moving objects to be processed move together.

[0157] As shown in Figure 12, multiple moving objects to be processed that are moving together are likely to appear simultaneously in both the first and second images and be detected simultaneously from both images. When multiple moving objects to be processed that are moving together are detected simultaneously from both the first and second images, the reliability of the detection results from the first and second images is high. Therefore, based on this idea, the information processing device 10 relatively increases the weight coefficient of the moving objects to be processed that are considered to have a high reliability in the detection result. With such an information processing device 10, the weight coefficients of multiple moving objects to be processed can be appropriately determined. Furthermore, with such an information processing device 10, a highly reliable statistical value (an estimate of movement time) can be calculated using the appropriately determined weight coefficients of the moving objects to be processed.

[0158] Although this disclosure has been described above with reference to embodiments, this disclosure is not limited to the embodiments described above. Various modifications to the structure and details of this disclosure are possible, which can be understood by those skilled in the art within the scope of this disclosure. Furthermore, each embodiment can be combined with other embodiments as appropriate.

[0159] Furthermore, the flowchart used in the above explanation shows multiple steps (processes) in sequence. However, the execution order of the steps performed in each embodiment is not limited to the order in which they are described. In each embodiment, the order of the illustrated steps can be changed to the extent that it does not impede the content.

[0160] Some or all of the above embodiments may also be described as follows, but are not limited to the following. 1. Acquisition means for acquiring multiple first images taken at a first location at different timings and multiple second images taken at a second location at different timings, A first detection means for detecting multiple moving objects as target moving objects from among multiple first images, A second detection means for detecting each of the multiple moving objects to be processed from the second image, A calculation means for each moving object to be processed calculates the travel time between the first point and the second point for each moving object to be processed, based on the timing of the capture of the first and second images in which each moving object to be processed is detected. A determination means for determining the weight coefficient of each of the moving objects to be processed according to a predetermined rule, A statistical value calculation means for calculating a statistical value of the travel time between the first point and the second point based on the travel time of each of the moving objects to be processed and the weight coefficient of each of the moving objects to be processed, An information processing device having 2. The determination means is, The information processing apparatus according to claim 1, which determines the weight coefficient of the mobile object to be processed based on the characteristic features of the appearance of the mobile object to be processed and the characteristic features of the appearance of each of the multiple mobile objects to be compared. 3. The aforementioned determination means is: The information processing apparatus according to claim 2, wherein, in the distribution of appearance feature quantities of a plurality of comparison target moving objects, the weight coefficient of the processing target moving object located in a portion where the appearance feature quantities are relatively dense is made smaller than the weight coefficient of the processing target moving object located in a portion where the appearance feature quantities are relatively sparse. 4. The determination means is, When calculating the weight coefficient of the first processing target mobile body among a plurality of processing target mobile bodies, A moving object detected from at least one of the first image and the second image, taken within a predetermined time from the time the first image in which the first object to be processed was detected, A moving object detected from at least one of the first image and the second image, taken within a predetermined time from the time the second image in which the first object to be processed was detected, A moving object detected from at least one of the first image and the second image, which are taken in the same or a predetermined level or more similar shooting environment as the shooting environment in which the first moving object to be processed was detected. A moving object detected from at least one of the first image and the second image, which are captured in the same or a predetermined level of similar shooting environment as the shooting environment in which the first object to be processed was detected in the second image, A moving object detected from an image taken at a location different from the first and second locations in a shooting environment that is the same as or at least similar to the shooting environment of the first image in which the first object to be processed was detected, and A moving object detected from an image taken at a location different from the first and second locations in a shooting environment that is the same as or at least similar to the shooting environment of the second image in which the first object to be processed was detected, The information processing apparatus according to claim 2 or 3, which processes at least one of the moving bodies among them as the comparison target moving body. 5. The determination means is, An information processing apparatus according to any one of 1 to 4, which determines the weight coefficient for each of the mobile objects to be processed based on the similarity between the characteristic features of the appearance of each of the mobile objects to be processed and a predefined reference characteristic. 6. The determination means is, The information processing apparatus according to claim 5, wherein the weight coefficient of the processing target moving body having a relatively high similarity is made smaller than the weight coefficient of the processing target moving body having a relatively low similarity. 7. The determination means is, An information processing apparatus according to any one of 1 to 6, wherein the weight coefficients of a plurality of processing target moving objects detected from the same first image and also detected from the same second image are made relatively larger than the weight coefficients of other processing target moving objects. 8. The determination means is, The information processing apparatus according to claim 7, wherein the weight coefficients of a plurality of processing target moving bodies detected from the same first image and also detected from the same second image, which satisfy predetermined companion conditions, are made relatively larger than the weight coefficients of other processing target moving bodies. 9. One or more computers, Multiple first images taken at different times at the first location and multiple second images taken at different times at the second location are acquired. Multiple moving objects are detected from among the multiple first images as the moving objects to be processed. Each of the multiple moving objects to be processed is detected from the second image, Based on the timing of the capture of the first and second images in which each of the moving objects to be processed is detected, the travel time between the first point and the second point is calculated for each of the moving objects to be processed. In accordance with the prescribed rules, the weight coefficient for each of the moving objects to be processed is determined, An information processing method for calculating a statistical value of the travel time between the first point and the second point, based on the travel time of each of the moving objects to be processed and the weight coefficient of each of the moving objects to be processed. 10. Computers, Acquisition means for acquiring multiple first images taken at a first location at different timings and multiple second images taken at a second location at different timings. A first detection means for detecting multiple moving objects as target moving objects from among multiple first images, A second detection means for detecting each of the multiple moving objects to be processed from the second image, A calculation means for each of the moving objects to be processed, which calculates the travel time between the first point and the second point for each of the moving objects to be processed, based on the timing of the capture of the first and second images in which each of the moving objects to be processed is detected. A determination means for determining the weight coefficient of each of the moving objects to be processed according to a predetermined rule, A statistical value calculation means that calculates a statistical value of the travel time between the first point and the second point based on the travel time of each of the moving objects to be processed and the weight coefficient of each of the moving objects to be processed. A program that makes it function as such.

[0161] Some or all of the appendices 2 through 8, which are dependent on the information processing device described in appendice 1 above, may also be dependent on the information processing method in appendice 9 and the program in appendice 10 in the same dependent relationship as between appendice 1 and appendices 2 through 8. Furthermore, within the scope that does not depart from each of the embodiments described above, some or all of the configurations described as appendices can be realized in various hardware, software, various recording means for recording software, or systems. [Explanation of Symbols]

[0162] 10 Information Processing Devices 11 Acquisition Department 12 First detection unit 13 Second detection unit 14 Calculation unit for each moving object 15. Decision Section 16. Statistical Value Calculation Section 1A Processor 2A Memory 3A input / output I / F 4A Peripheral Circuits 5A bus

Claims

1. An acquisition means for acquiring multiple first images taken at different times at a first location and multiple second images taken at different times at a second location, A first detection means for detecting multiple moving objects as target moving objects from among multiple first images, A second detection means for detecting each of the multiple moving objects to be processed from the second image, A calculation means for each moving object to be processed calculates the travel time between the first point and the second point for each moving object to be processed, based on the timing of the capture of the first and second images in which each moving object to be processed is detected. A determination means for determining the weight coefficient of each of the moving objects to be processed according to a predetermined rule, A statistical value calculation means for calculating a statistical value of the travel time between the first point and the second point based on the travel time of each of the moving objects to be processed and the weight coefficient of each of the moving objects to be processed, An information processing device having

2. The aforementioned determination means is The information processing apparatus according to claim 1, which determines the weight coefficient of the mobile object to be processed based on the characteristic features of the appearance of the mobile object to be processed and the characteristic features of the appearance of each of the plurality of mobile objects to be compared.

3. The aforementioned determination means is The information processing apparatus according to claim 2, wherein, in the distribution of appearance feature quantities of a plurality of comparison target moving objects, the weight coefficient of the processing target moving object located in a portion where the appearance feature quantities are relatively dense is made smaller than the weight coefficient of the processing target moving object located in a portion where the appearance feature quantities are relatively sparse.

4. The aforementioned determination means is When calculating the weight coefficient of the first processing target mobile body among the multiple processing target mobile bodies, A moving object detected from at least one of the first image and the second image, taken within a predetermined time from the time the first image in which the first object to be processed was detected, A moving object detected from at least one of the first image and the second image, taken within a predetermined time from the time the second image in which the first object to be processed is detected, A moving object detected from at least one of the first image and the second image, which are taken in the same or a predetermined level or more similar shooting environment as the shooting environment in which the first image in which the first object to be processed was detected, A moving object detected from at least one of the first image and the second image, which are taken in the same or a predetermined level or more similar shooting environment as the shooting environment in which the first moving object to be processed was detected in the second image, A moving object detected from an image taken at a location different from the first and second locations in a shooting environment that is the same as or at least similar to the shooting environment of the first image in which the first object to be processed was detected, and A moving object detected from an image taken at a location different from the first and second locations in a shooting environment that is the same as or at least similar to the shooting environment of the second image in which the first object to be processed was detected, The information processing apparatus according to claim 2, wherein at least one of the moving bodies among them is processed as the comparison target moving body.

5. The aforementioned determination means is The information processing apparatus according to claim 1, which determines the weight coefficient for each of the mobile objects to be processed based on the similarity between the characteristic features of the appearance of each of the mobile objects to be processed and a predefined reference characteristic.

6. The aforementioned determination means is The information processing apparatus according to claim 5, wherein the weight coefficient of the processing target moving body having a relatively high similarity is made smaller than the weight coefficient of the processing target moving body having a relatively low similarity.

7. The aforementioned determination means is The information processing apparatus according to claim 1, wherein the weight coefficients of a plurality of processing target moving objects detected from the same first image and also detected from the same second image are made relatively larger than the weight coefficients of other processing target moving objects.

8. The aforementioned determination means is The information processing apparatus according to claim 7, wherein the weight coefficients of a plurality of processing target moving bodies detected from the same first image and also detected from the same second image, which satisfy predetermined companion conditions, are made relatively larger than the weight coefficients of the other processing target moving bodies.

9. One or more computers, Multiple first images taken at a first location at different times and multiple second images taken at a second location at different times are acquired. Multiple moving objects are detected from among the multiple first images as the moving objects to be processed. Each of the multiple moving objects to be processed is detected from the second image, Based on the timing of the capture of the first and second images in which each of the moving objects to be processed is detected, the travel time between the first point and the second point is calculated for each of the moving objects to be processed. In accordance with the prescribed rules, the weight coefficient for each of the moving objects to be processed is determined, An information processing method for calculating a statistical value of the travel time between the first point and the second point, based on the travel time of each of the moving objects to be processed and the weight coefficient of each of the moving objects to be processed.

10. Computers, Acquisition means for acquiring multiple first images taken at a first location at different timings and multiple second images taken at a second location at different timings. A first detection means for detecting multiple moving objects as target moving objects from among multiple first images, A second detection means for detecting each of the multiple moving objects to be processed from the second image, A calculation means for each of the moving objects to be processed, which calculates the travel time between the first point and the second point for each of the moving objects to be processed, based on the timing of the capture of the first and second images in which each of the moving objects to be processed is detected. A determination means for determining the weight coefficient of each of the moving objects to be processed according to a predetermined rule, A statistical value calculation means that calculates a statistical value of the travel time between the first point and the second point based on the travel time of each of the moving objects to be processed and the weight coefficient of each of the moving objects to be processed. A program that makes it function as such.

Images