WO2015029588A1 - 画像処理システム、画像処理方法及びプログラム - Google Patents
画像処理システム、画像処理方法及びプログラム Download PDFInfo
- Publication number
- WO2015029588A1 WO2015029588A1 PCT/JP2014/067693 JP2014067693W WO2015029588A1 WO 2015029588 A1 WO2015029588 A1 WO 2015029588A1 JP 2014067693 W JP2014067693 W JP 2014067693W WO 2015029588 A1 WO2015029588 A1 WO 2015029588A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- image
- stay
- persons
- processing system
- input
- Prior art date
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/194—Segmentation; Edge detection involving foreground-background segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
- G06V20/53—Recognition of crowd images, e.g. recognition of crowd congestion
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30196—Human being; Person
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30242—Counting objects in image
Definitions
- Some aspects according to the present invention relate to an image processing system, an image processing method, and a program.
- Patent Document 1 determines whether or not there is a motion in each partial area in the input image, and determines whether or not there is a person from the texture information for each partial area, thereby Discloses a congestion estimation device that can determine whether or not a message exists.
- Patent Documents 2-4 also disclose related technologies.
- one object of the present invention is to provide an image processing system, an image processing method, and a program that can suitably detect staying of a plurality of persons. To do.
- An image processing system is based on generation means for generating a background image based on input images taken at a plurality of times by an imaging device, and input images taken within a first time width from the processing time.
- a plurality of persons in the input image based on the difference between the generated first background image and the second background image generated based on the input image taken within the second time width from the processing time
- First detection means for detecting staying in a partial area in which an image can appear
- second detection means for detecting one or more persons appearing in the partial area of the input image
- stay detection results and person detection results
- a third detecting means for detecting staying of a plurality of persons in the partial area.
- An image processing method includes a step of generating a background image based on input images taken at a plurality of times by a photographing apparatus, and a generation based on an input image taken within a first time width from the processing time.
- a plurality of persons in the input image based on the difference between the first background image thus generated and the second background image generated based on the input image captured within the second time width from the processing time.
- a plurality of persons in the partial area based on the step of detecting staying in the partial area that can be reflected, the step of detecting one or more persons appearing in the partial area of the input image, and the detection result of the stay and the detection result of the person
- the image processing system performs the step of detecting the stagnation.
- a program according to the present invention is generated based on a process of generating a background image based on input images taken at a plurality of times by a photographing device and an input image photographed within a first time width from the processing time. Based on the difference between the first background image and the second background image generated based on the input image taken within the second time width from the processing time, a plurality of persons in the input image can be reflected. Based on a process for detecting staying in a partial area, a process for detecting one or more persons appearing in the partial area of the input image, and a stay detection result and a person detection result, a plurality of persons stay in the partial area And causing the computer to execute processing for detecting
- “part”, “means”, “apparatus”, and “system” do not simply mean physical means, but “part”, “means”, “apparatus”, “system”. This includes the case where the functions possessed by "are realized by software. Further, even if the functions of one “unit”, “means”, “apparatus”, and “system” are realized by two or more physical means or devices, two or more “parts” or “means”, The functions of “device” and “system” may be realized by a single physical means or device.
- an image processing system an image processing method, and a program capable of suitably detecting the stay of a plurality of persons.
- FIG. 1 is a functional block diagram illustrating a schematic configuration of an image processing system according to a first embodiment. It is a flowchart which shows the flow of a process of the image processing system shown in FIG. It is a flowchart which shows the flow of a process of the image processing system shown in FIG.
- FIG. 1 It is a flowchart which shows the flow of a process of the image processing system shown in FIG. It is a block diagram which shows the structure of the hardware which can mount the image processing system shown in FIG. It is a functional block diagram which shows schematic structure of the image processing system which concerns on 2nd Embodiment.
- the image processing system is for detecting staying (hanging) of a plurality of persons from an image captured by an imaging device such as a surveillance camera.
- FIG. 1 is a diagram showing a specific example of an image photographed by the photographing apparatus.
- people P1 to P5 are shown.
- the detection window W including the persons P1 to P3 is specified as a hanging area.
- a staying score calculated based on the detection result of the staying object region and a person with respect to the detection window W that can be set in the input image After calculating the congestion score calculated based on the detection result, a hang-up score for detecting hang-up is calculated using the stay score and the congestion score.
- the image processing system determines that hang-up has occurred in the hang-up detection window W. For example, as shown in FIG. Is shown on the video to notify the user of the occurrence of hang-up.
- the “crowd score” in the present embodiment can also be referred to as “crowd density”, for example.
- the detection window W is set to a size that includes a plurality of persons (in the example of FIG. 1, it is set to a size that can contain three persons).
- a large number of detection windows W can be set in the input image.
- Each detection window W may overlap each other.
- the size of the detection window W is set based on the position of the photographing apparatus, the size of the person, the number of persons included, and the like. Thereby, the detection window W is set large on the lower side of the input image (corresponding to the vicinity of the imaging device in the three-dimensional space to be imaged), and the upper side of the input image (in the three-dimensional space to be imaged, the imaging device The detection window W is set to be small in the case of distant).
- the retention score used when the image processing system according to the present embodiment calculates the hang-up score may be set such that the weight of the lower region W1 in the detection window W is increased and the weight of the upper region W2 is decreased.
- the stay detection result in the lower region W1 in the detection window W has a greater influence on the hang-up score than the stay detection result in the upper region W2. This is because, generally, when a human is stationary, the lower half of the body moves less than the upper half of the body, so that the lower region W1 can detect the stay more appropriately.
- the congestion score of the detection window W may be calculated by increasing the weight of the upper region W2 than the lower region W1. This is because, in the detection window W set according to the size of the person, if the detected murmur exists in the detection window W, the head and upper body should be more present above the detection window W. It is.
- FIG. 2 shows a specific example of coefficients (weights) to be multiplied with the stay score and the congestion score when calculating the hang-up score.
- 1 is set for the stay score in the lower region W2
- 0 is set for the stay score in the upper region W1
- 0 is set for the congestion score in the lower region W2
- the upper region W2 is set.
- 1 is set for the stay score at.
- the coefficient setting method is not limited to this.
- the retention score may be set in any way as long as the value in the lower region W1 has a larger influence than the value in the upper region W2, and it is necessary to set it in a binary manner as in the example of FIG. There is no. The same applies to the coefficient for the congestion score.
- the congestion score is calculated only for the upper region W2, and the stay score is calculated only for the lower region W1. Even if it is calculated, the same result can be obtained.
- the congestion score or the staying score may be calculated for only one of the upper region W1 and the lower region W2.
- a background image is created by averaging a large number of captured images taken for a long period of time (hereinafter also referred to as a long time window).
- a background image obtained by averaging captured images taken in a shorter period (hereinafter also referred to as a short time window) than that, an object that appears longer in the shorter period can be obtained. It is conceivable to detect as a staying object.
- an average image background image
- the influence of moving objects (including persons) that are immediately out of frame in the image is reduced. Since it can suppress, it becomes easy to extract a stationary object.
- an image that is generated based on a plurality of images and is suppressed from being affected by a moving object is called a background image.
- a moving object region hereinafter also referred to as “foreground region” and other still regions (hereinafter also referred to as “background regions”) are taken from the captured image.
- a still image of each captured image is averaged in the time direction to generate a background image.
- the human flow generation region becomes the foreground region, and thus the influence of the human flow on the generated background image can be suppressed.
- the background image of the long time window and the background image of the short time window are compared, it is possible to detect the staying object suitably.
- FIGS. 3 to 5 a specific example of the flow of processing will be described with reference to FIGS. 3 to 5.
- FIG. 3 is a diagram showing specific examples of input images taken from time t-4 to time t. Here, no one is shown in the images at time t-4 and time t-3, and a person is shown in the images at time t-2 to time t. The person is moving at time t-2, but the person is stopped at time t-1 and time t.
- each input image as shown in FIG. 3 is divided into a moving area (foreground area) and a stationary area (background area).
- FIG. 2 is a diagram illustrating an example of an image when only a still area is extracted from the input image of FIG.
- the area including the person is excluded from the image as the moving area.
- the input image remains unchanged.
- the image processing system generates a background image for a plurality of time windows from each image from time t-4 to time t from which only a still region is extracted.
- This background image is, for example, an average value, a median value, or a mode value of pixel values for each pixel in the image with respect to a still region of each image captured within a past fixed period from the current time t. It can be generated by obtaining. If a background image is generated for each of the short time window and the long time window, the image processing system compares the two, extracts pixels whose differences exceed the threshold, and identifies the pixel area as a staying area. To do.
- the stay score can be calculated as the size of the stay area (for example, the number of pixels).
- a person is not displayed in the background image for the long time window. This is because, for example, the mode value or median value of the pixel value is generated for each pixel when the background image is generated. This is because the influence of a person who is shown only for a short time is weakened.
- the image processing system according to the present embodiment appropriately detects a staying object as shown by the lowermost image in FIG. 5 by the processing shown in FIG.
- FIG. 6 is a block diagram illustrating a system configuration of the image processing system 1.
- the image processing system 1 includes an image input unit 601, a foreground / background separation unit 603, a background image generation unit 605, a background image storage unit 607, a background image comparison unit 609, a still area determination unit 611, a stay score calculation unit 613, a database ( DB) 615, person detection unit 621, congestion score calculation unit 623, hang-up score calculation unit 625, and output unit 627.
- an image input unit 601 a foreground / background separation unit 603, a background image generation unit 605, a background image storage unit 607, a background image comparison unit 609, a still area determination unit 611, a stay score calculation unit 613, a database ( DB) 615, person detection unit 621, congestion score calculation unit 623, hang-up score calculation unit 625, and output unit 627.
- DB database
- the image input unit 601 sequentially receives input of frame images included in video input from a photographing device such as a monitoring camera (not shown). That is, each frame image is an image having a different shooting time.
- the image input unit 601 may receive an input of a frame image obtained by decoding video data stored in an HDD (Hard Disk Drive) (not shown) or a VCR (Video Cassette Recorder).
- HDD Hard Disk Drive
- VCR Video Cassette Recorder
- the foreground / background separation unit 603 sequentially separates the foreground region and the background region from the input image input from the image input unit 601 using, for example, a background difference method or an optical flow.
- the foreground area is an area with motion in the image
- the background area is an area without motion (still).
- the foreground / background separation unit 603 identifies a moving block after comparing with the previous frame in units of macroblocks, for example. More specifically, for example, it may be specified by examining where a macroblock similar to each macroblock (a set of pixels) in the image to be processed is located in the immediately preceding image (block matching) or the like. it can.
- the foreground / background separation unit 603 identifies the moving object by comparing the difference between the background image acquired in advance and the image to be processed.
- the background image generation unit 605 includes a background region (an image of a still region) extracted by the foreground / background separation unit 603 and an image captured in a plurality of predetermined time windows stored in the background image storage unit 607.
- a background image is generated using the background region according to.
- the background image generation unit 605 calculates the average value, median value, or mode value of the pixel values for the time window obtained at each pixel position related to the background area of each image, thereby obtaining the background image. Can be generated.
- the background image storage unit 607 stores the images of the background area of each input image, which are sequentially extracted by the foreground / background separation unit 603, for a predetermined time.
- the background image storage unit 607 stores the background image in each time window generated by the background image generation unit 605 in accordance with the processing in the background image generation unit 605.
- the background image comparison unit 609 compares the background images generated by the background image generation unit 605 for each time window. More specifically, by comparing a background image generated from the longest time window (a background image assumed to be composed of a true background) and a background image generated from a shorter time window It is possible to detect a stationary object (a staying object) that has been stationary for a certain period of time. At this time, it may be possible to classify and detect stationary objects according to the stationary time length by generating a background image with a plurality of time window backgrounds.
- a background image comparison method by the background image comparison unit 609 for example, a method using an absolute value of a pixel value difference between background images, or a pixel in a rectangular area while operating a small-sized rectangular area on the image.
- a method for calculating a correlation between values, a method for calculating a histogram distance of pixel values in a rectangular area, and the like can be considered.
- the method using the rectangular area for example, a fixed size such as a macro block may be set, or a detection target object (person) is set using camera parameters (set according to the installation position of the camera).
- a different size is set for each location on the image (for example, the image area in which the vicinity of the image capturing device is shown has a larger rectangular area. A rectangular area may be made smaller in the image area in which the far field is shown.
- the still region determination unit 611 identifies a pixel having a difference that exceeds a threshold, and determines a pixel region including such a pixel as a stay region.
- the still region determination unit 611 identifies the stay region for the entire input image, but is not limited to this, and stays in at least the region including the detection window W to be processed. What is necessary is just to specify an area
- the stay score calculation unit 613 calculates the stay score of the detection window W to be processed using the stay region determined by the still region determination unit 611.
- a calculation method of the staying score as described above, for example, it may be calculated based on the number of pixels in the staying area in the detection window W.
- information related to the position and size of the detection window W is set in advance in the DB 615 as the divided region information 617.
- a plurality of detection windows W can be set for the input image. More specifically, for example, a large number of detection windows W are set so as to overlap each other over the entire input image. It is possible.
- the size of the detection window W may be a size that can include more than the number of people specified as hangouts.
- the detection window W is large corresponding to the vicinity of the photographing apparatus in accordance with the positional relationship with the photographing apparatus (for example, set as a camera parameter (not shown)), and the detection window corresponding to a distance from the photographing apparatus. It is conceivable to set W small.
- the residence score is multiplied by a coefficient set so that the influence of the lower region W1 is greater than that of the upper region W2.
- the coefficient is stored as coefficient information 619 in the DB 615.
- the person detection unit 621 detects a person from the image input from the image input unit 601.
- There are various methods for detecting a person For example, after preparing in advance with a learning device that has learned features on the image of the head, upper body, whole body, or crowd patch, A method of detecting a part or upper body from an input image is conceivable.
- the person detection unit 621 detects a person in the entire input image, but is not limited to this.
- the person detection is performed at least in a region including the detection window W to be processed. Just do it.
- the congestion score calculation unit 623 calculates the congestion score of the detection window W to be processed based on the person detection result determined by the person detection unit 621. As a calculation method of the congestion score, for example, it may be calculated based on the number of persons detected in the detection window W. Here, as described above, a coefficient set so that the influence of the upper region W2 is greater than that of the lower region W1 is calculated as the congestion score. The coefficient is stored as coefficient information 619. Note that the congestion score calculation unit 623 may store the number of persons in the detection window W that is the processing target in time series and detect the increase or decrease.
- the hang-up score calculation unit 625 calculates a hang-up score in the processing target detection window W using the stay score calculated by the stay score calculation unit 613 and the congestion score calculated by the congestion score calculation unit 623.
- Various methods of calculating the hang-up score are conceivable. For example, a method of setting a value obtained by multiplying the residence score multiplied by the coefficient and the congestion score as the hang-up score can be considered.
- the output unit 627 outputs the hangup detection result based on the hangup score obtained by the hangup score calculation unit 625.
- the hang-up score for the detection window W may be displayed as a numerical value (if a plurality of detection windows W are set, the hang-up score may be displayed respectively).
- a detection window W corresponding to the hang-up score (for example, an image showing the detection window W having a thickness corresponding to the size of the hang-up score) may be superimposed on the input image. It is also conceivable to notify the user of information corresponding to the hanging score by voice or the like.
- the output unit 627 gradually increases the number of people in the predetermined area (the human being gradually increases in the predetermined area). May be output separately.
- the output unit 627 may notify the user only when the hang-up score exceeds the threshold value. Alternatively, the output unit 627 may be notified to the user according to the duration that the hang-up score exceeds the threshold.
- the threshold value of the hang-up score may be set in advance, or may be set by user input.
- the output unit 627 may also output the information.
- FIGS. 7 to 9 are flowcharts showing a processing flow of the image processing system 1 according to the present embodiment.
- Each processing step to be described later can be executed in any order or in parallel as long as there is no contradiction in processing contents, and other steps can be added between the processing steps. good. Further, a step described as a single step for convenience can be executed by being divided into a plurality of steps, and a step described as being divided into a plurality of steps for convenience can be executed as one step.
- the stay score calculation unit 613 calculates a stay score for the detection window W of the image input from the image input unit 601 (S701).
- the processing flow of the image processing system 1 leading to the residence score calculation will be described later with reference to FIG.
- the congestion score calculation unit 623 calculates a congestion score for the detection window W of the image input from the image input unit 601 (S703).
- the flow of processing of the image processing system 1 leading to congestion score calculation will be described later with reference to FIG. Note that the processing order of S701 and S703 may be reversed, or the processes may be performed in parallel.
- the hang-up score calculation unit 625 calculates a hang-up score based on the stay score and the congestion score (S705).
- the output unit 627 outputs the sag detection result related to the detection window W based on the calculated sag score (S707).
- the image input unit 601 is obtained by reading out an image photographed by a photographing device such as a video camera or video data recorded with a video photographed by the photographing device, and then decoding the video data. An image input is received (S801).
- a photographing device such as a video camera or video data recorded with a video photographed by the photographing device
- the foreground / background separation unit 603 separates the image input from the image input unit 601 into a stationary background region and a foreground region in which movement has occurred (S803).
- the generated image of the background area is stored in the background image storage unit 607, for example.
- the background image generation unit 605 generates a background image by using the background area specified by the foreground / background separation unit 603 among the images taken within the preset time window (S805). At this time, the background image generation unit 605 generates background images for a plurality of time windows in accordance with the residence time of the person related to the extracted hangout.
- the background image comparison unit 609 compares the background images of the respective time windows generated by the background image generation unit 605 (S807), and the still area determination unit 611 determines an area where the difference between the background images is equal to or greater than a threshold value. It is specified as a staying area (S809).
- the stay score calculation unit 613 calculates a stay score based on the size of the stay area in the detection window W to be processed (S811). At this time, the staying score may be multiplied by a predetermined coefficient based on the position of the pixel included in the staying area in the detection window W.
- the image input unit 601 is obtained by reading out an image photographed by a photographing device such as a video camera or video data recorded with a video photographed by the photographing device, and then decoding the video data.
- An image input is received (S901). Note that this process can also serve as S801 in the process of FIG.
- the person detection unit 621 detects a person from the input image (S903).
- a person detection method may be head detection or upper body detection.
- the congestion score calculation unit 623 calculates a congestion score in the detection window W that is a processing target based on the person detection result detected by the person detection unit 621 (S905).
- the congestion score can be calculated based on the number of persons detected in the detection window W, for example.
- the congestion score may be multiplied by a predetermined coefficient based on the position of the detected person in the detection window W or the like.
- the image processing system 1 includes a processor 1001, a memory 1003, a storage device 1005, an input interface (I / F) 1007, a data I / F 1009, a communication I / F 1011, and a display device 1013.
- a processor 1001 a memory 1003, a storage device 1005, an input interface (I / F) 1007, a data I / F 1009, a communication I / F 1011, and a display device 1013.
- the processor 1001 controls various processes in the image processing system 1 by executing a program stored in the memory 1003.
- the processing related to the calculation unit 623, the hang-up score calculation unit 625, and the output unit 627 can be realized as a program that is temporarily stored in the memory 1003 and mainly operates on the processor 1001.
- the memory 1003 is a storage medium such as a RAM (Random Access Memory).
- the memory 1003 temporarily stores a program code of a program executed by the processor 1001 and data necessary for executing the program. For example, in the storage area of the memory 1003, a stack area necessary for program execution is secured.
- the storage device 1005 is a non-volatile storage medium such as a hard disk or a flash memory.
- the storage device 1005 includes an operating system, an image input unit 601, a foreground / background separation unit 603, a background image generation unit 605, a background image comparison unit 609, a still area determination unit 611, a stay score calculation unit 613, a person detection unit 621, Various programs for realizing the congestion score calculation unit 623, the hang-up score calculation unit 625, and the output unit 627, various data including the background image storage unit 607 and the DB 615, and the like are stored. Programs and data stored in the storage device 1005 are referred to by the processor 1001 by being loaded into the memory 1003 as necessary.
- the input I / F 1007 is a device for receiving input from the user. Specific examples of the input I / F 1007 include a keyboard, a mouse, and a touch panel. The input I / F 1007 may be connected to the image processing system 1 via an interface such as USB (Universal Serial Bus), for example.
- USB Universal Serial Bus
- the data I / F 1009 is a device for inputting data from the outside of the image processing system 1.
- Specific examples of the data I / F 1009 include a drive device for reading data stored in various storage media. It is conceivable that the data I / F 1009 is provided outside the image processing system 1. In that case, the data I / F 1009 is connected to the image processing system 1 via an interface such as a USB.
- the communication I / F 1011 is a device for data communication with an external device of the image processing system 1, for example, a video camera or the like by wire or wireless.
- the communication I / F 1011 may be provided outside the image processing system 1. In that case, the communication I / F 1011 is connected to the image processing system 1 via an interface such as a USB.
- the display device 1013 is a device for displaying various types of information such as video as shown in FIG.
- Specific examples of the display device 1013 include a liquid crystal display and an organic EL (Electro-Luminescence) display.
- the display device 1013 may be provided outside the image processing system 1. In that case, the display device 1013 is connected to the image processing system 1 via, for example, a display cable.
- the image processing system 1 uses the stay score calculated based on the stay area in the input image and the congestion score calculated based on the person detected from the input image. Used to detect stagnation of multiple persons. Thereby, while many objects staying for a fixed time or more are contained, the area
- a background image generated from images taken over a plurality of times is used, so that the stay area of a plurality of persons is temporarily blocked by other persons who move in front of the stay area. Even in such a case, it is possible to suitably detect the hang-up while reducing the influence thereof.
- FIG. 11 is a block diagram illustrating a functional configuration of the image processing system 1100.
- the image processing system 1100 includes a generation unit 1110, a first detection unit 1120, a second detection unit 1130, and a third detection unit 1140.
- the generation unit 1110 generates a background image based on input images captured at a plurality of times by the imaging device.
- the first detection unit 1120 applies the first background image generated based on the input image captured within the first time width from the processing time and the input image captured within the second time width from the processing time. Based on the difference from the second background image generated on the basis of the second background image, stagnation is detected in a partial area in the input image where a plurality of persons can be reflected.
- the second detection unit 1130 detects one or more persons appearing in the partial area of the input image.
- the third detection unit 1140 detects the stagnation of a plurality of persons in the partial area based on the stagnation detection result and the person detection result.
- a generating unit configured to generate a background image based on input images captured at a plurality of times by the imaging device; a first background image generated based on an input image captured within a first time width from the processing time; Based on the difference from the second background image generated based on the input image photographed within the second time width from the processing time, it stays in the partial area in the input image where a plurality of persons can appear
- a plurality of persons in the partial area based on the first detection means for detecting the second detection means, the second detection means for detecting one or more persons appearing in the partial area of the input image, and the detection result of the stay and the detection result of the person
- An image processing system comprising: third detection means for detecting staying in the apparatus.
- Appendix 2 The image processing system according to appendix 1, wherein the background image is generated based on a static region having no motion among input images.
- Appendix 6 The image processing system according to appendix 5, wherein the output means changes a notification method according to a degree of stay of the plurality of persons.
- Appendix 7 The image processing system according to appendix 5, wherein the output means notifies based on a time when the degree of stay of the plurality of persons exceeds a threshold value.
- Appendix 8 The image processing system according to appendix 7, wherein the threshold value can be set by a user.
- the second detection means detects an increase or decrease in the number of people in the partial area, and the output means gradually gathers persons in a predetermined area based on the detection result of the increase or decrease in the number of persons by the second detection means.
- the image processing system according to any one of supplementary notes 5 to 8, wherein information indicating that the information is separately provided is separately notified.
- Appendix 11 The image processing method according to appendix 10, wherein the background image is generated based on a static region having no motion in the input image.
- Appendix 12 12. The image processing method according to appendix 10 or appendix 11, wherein the size of the partial area is set according to a position of the photographing device that captures an input image and the partial area in the input image.
- Appendix 13 Any one of appendix 10 to appendix 12, wherein the stay detection result in the lower region in the partial region has a greater influence on the stay detection result of the plurality of persons than the stay detection result in the upper region. The image processing method as described.
- Appendix 14 The image processing method according to any one of appendix 10 to appendix 13, further comprising a step of notifying a detection result of the stay of the plurality of persons in the partial area.
- Appendix 16 The image processing method according to appendix 14, wherein notification is made based on a time when the degree of stay of the plurality of persons exceeds a threshold value.
- Appendix 17 The image processing method according to appendix 16, wherein the threshold value can be set by a user.
- Appendix 18 Additional information for detecting an increase or decrease in the number of persons in the partial area and separately indicating information indicating that persons are gradually gathering in the predetermined area based on the detection result of the increase or decrease in the number of persons by the second detection unit, 18.
- the image processing method according to any one of appendix 17.
- a process for detecting, a process for detecting one or more persons appearing in the partial area of the input image, and a process for detecting a stay of a plurality of persons in the partial area based on a stay detection result and a person detection result A program to be executed.
- Appendix 20 The program according to appendix 19, wherein the background image is generated based on a static area having no motion in the input image.
- Appendix 21 The program according to appendix 19 or appendix 20, wherein the size of the partial area is set according to a position of the photographing apparatus that captures the input image and the partial area in the input image.
- Appendix 22 Any one of appendix 19 to appendix 21, wherein the stay detection result in the lower region of the partial region has a greater influence on the stay detection result of the plurality of persons than the stay detection result in the upper region. The listed program.
- Appendix 23 The program according to any one of appendix 19 to appendix 22, further causing the computer to execute an output process for informing the detection result of the stay of the plurality of persons in the partial area.
- Appendix 24 The program according to appendix 23, wherein the notification method is changed according to the degree of stay of the plurality of persons.
- Appendix 26 The program according to appendix 25, wherein the threshold value can be set by a user.
- Addendum 23 which detects an increase or decrease in the number of persons in the partial area and separately notifies information indicating that persons are gradually gathering in the predetermined area based on the detection result of the increase or decrease in the number of persons by the second detection unit. Thru / or the program according to any one of appendix 26.
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
Description
その他、特許文献2-4も、関連技術を開示している。
図1乃至図10は、第1実施形態を説明するための図である。以下、これらの図を参照しながら、以下の流れに沿って本実施形態を説明する。まず、「1.1」で本実施形態における複数人物の滞留(以下、「たむろ」ともいう。)の検出方法の概要を説明する。その後、「1.2」で本実施形態に係る画像処理システムの機能構成の概要を、「1.3」で画像処理システムの処理の流れを説明する。「1.4」では、画像処理システムを実現可能なハードウェア構成の具体例を示す。最後に、「1.5」以降で、本実施形態に係る効果などを説明する。
(1.1.1 たむろの検出方法)
本実施形態に係る画像処理システムは、例えば監視カメラ等の撮影装置で撮影される映像から、複数人物の滞留(たむろ)を検出するためのものである。
以下、滞留スコアの算出方法の概要を、図3乃至図5を参照しながら説明する。
本実施形態に係る画像処理システムは、図5に示す処理により、図5の最下部の画像で示すように、適切に滞留物体を検出する。
以下、図6を参照しながら、本実施形態に係る画像処理システム1のシステム構成を説明する。図6は、画像処理システム1のシステム構成を示すブロック図である。
なお、混雑スコア算出部623は、処理対象である検知窓Wにおける人数を時系列的に記憶しておき、その増減を検出するようにしても良い。
以下、画像処理システム1の処理の流れを、図7乃至図9を参照しながら説明する。図7乃至図9は、本実施形態に係る画像処理システム1の処理の流れを示すフローチャートである。
まず、全体の処理の流れを図7を参照しながら説明する。
滞留スコア算出部613は、画像入力部601から入力された画像の検知窓Wに対する滞留スコアを算出する(S701)。滞留スコア算出に至る画像処理システム1の処理の流れは、図8を参照しながら後述する。
続いて、図8を参照しながら、滞留スコアの算出に係る画像処理システム1の処理の流れを説明する。当該処理は、図7のS701に相当する。
続いて、図9を参照しながら、混雑スコアの算出にかかる画像処理システム1の処理の流れを説明する。当該処理は、図7のS703に相当する。
人物検出部621は、入力画像から人物を検出する(S903)。前述の通り、人物の検出方法としては、頭部検出や上半身検出などが考えられる。
以下、図10を参照しながら、上述してきた画像処理システム1をコンピュータにより実現する場合のハードウェア構成の一例を説明する。なお、画像処理システム1の機能は、複数の情報処理装置により実現することも可能である。
以上説明したように、本実施形態に係る画像処理システム1は、入力画像内の滞留領域にもとづいて算出される滞留スコアと、入力画像から検出される人物に基づいて算出される混雑スコアとを用いて、複数人物の滞留(たむろ)を検出する。これにより、一定時間以上滞留する物体が多く含まれると共に、人物が多く存在している領域をたむろとして好適に検知することができる。
以下、第2実施形態を、図11を参照しながら説明する。図11は、画像処理システム1100の機能構成を示すブロック図である。図11に示すように、画像処理システム1100は、生成部1110と、第1検出部1120と、第2検出部1130と、第3検出部1140とを含む。
生成部1110は、撮影装置により複数の時刻に撮影された入力画像に基づいて背景画像を生成する。
第2検出部1130は、入力画像の部分領域に映る1以上の人物を検出する。
第3検出部1140は、滞留の検出結果及び人物の検出結果に基づき、部分領域における複数人物の滞留を検出する。
このように実装することで、本実施形態に係る画像処理システム1100によれば、複数人物の滞留を好適に検出することができる。
なお、前述の実施形態の構成は、組み合わせたり或いは一部の構成部分を入れ替えたりしてもよい。また、本発明の構成は前述の実施形態のみに限定されるものではなく、本発明の要旨を逸脱しない範囲内において種々変更を加えてもよい。
撮影装置により複数の時刻に撮影された入力画像に基づいて背景画像を生成する生成手段と、処理時刻から第1の時間幅内に撮影された入力画像に基づき生成された第1の背景画像と、前記処理時刻から第2の時間幅内に撮影された入力画像に基づき生成された第2の背景画像との差分に基づき、入力画像内の、複数の人物が映り得る部分領域に対して滞留を検出する第1の検出手段と、入力画像の前記部分領域に映る1以上の人物を検出する第2の検出手段と、滞留の検出結果及び人物の検出結果に基づき、前記部分領域における複数人物の滞留を検出する第3の検出手段とを備える画像処理システム。
前記背景画像は、入力画像のうち、動きのない静止領域に基づいて生成される、付記1記載の画像処理システム。
前記部分領域の大きさは、入力画像を撮影する前記撮影装置の位置と、入力画像内の前記部分領域とに応じて設定される、付記1又は付記2記載の画像処理システム。
前記部分領域内の下部領域における滞留の検出結果の方が、上部領域における滞留の検出結果よりも、前記複数人物の滞留の検出結果に大きな影響を与える、付記1乃至付記3のいずれか1項記載の画像処理システム。
前記部分領域における前記複数人数の滞留の検出結果を報知する出力手段を更に備える、付記1乃至付記4のいずれか1項記載の画像処理システム。
前記出力手段は、前記複数人数の滞留の度合いに応じて報知方法を変える、付記5記載の画像処理システム。
前記出力手段は、前記複数人数の滞留の度合いが閾値を超えた時間に基づいて報知する、付記5記載の画像処理システム。
前記閾値はユーザが設定可能である、付記7記載の画像処理システム。
前記第2の検出手段は、前記部分領域における人数の増減を検出し、前記出力手段は、前記第2の検出手段による前記人数の増減の検出結果に基づき、所定領域において徐々に人物が集合していることを示す情報を別途報知する、付記5乃至付記8のいずれか1項記載の画像処理システム。
撮影装置により複数の時刻に撮影された入力画像に基づいて背景画像を生成するステップと、処理時刻から第1の時間幅内に撮影された入力画像に基づき生成された第1の背景画像と、前記処理時刻から第2の時間幅内に撮影された入力画像に基づき生成された第2の背景画像との差分に基づき、入力画像内の、複数の人物が映り得る部分領域に対して滞留を検出するステップと、入力画像の前記部分領域に映る1以上の人物を検出するステップと、滞留の検出結果及び人物の検出結果に基づき、前記部分領域における複数人物の滞留を検出するステップとを画像処理システムが行う、画像処理方法。
前記背景画像は、入力画像のうち、動きのない静止領域に基づいて生成される、付記10記載の画像処理方法。
前記部分領域の大きさは、入力画像を撮影する前記撮影装置の位置と、入力画像内の前記部分領域とに応じて設定される、付記10又は付記11記載の画像処理方法。
前記部分領域内の下部領域における滞留の検出結果の方が、上部領域における滞留の検出結果よりも、前記複数人物の滞留の検出結果に大きな影響を与える、付記10乃至付記12のいずれか1項記載の画像処理方法。
前記部分領域における前記複数人数の滞留の検出結果を報知するステップを更に備える、付記10乃至付記13のいずれか1項記載の画像処理方法。
前記複数人数の滞留の度合いに応じて報知方法を変える、付記14記載の画像処理方法。
前記複数人数の滞留の度合いが閾値を超えた時間に基づいて報知する、付記14記載の画像処理方法。
前記閾値はユーザが設定可能である、付記16記載の画像処理方法。
前記部分領域における人数の増減を検出し、前記第2の検出手段による前記人数の増減の検出結果に基づき、所定領域において徐々に人物が集合していることを示す情報を別途報知する、付記14乃至付記17のいずれか1項記載の画像処理方法。
撮影装置により複数の時刻に撮影された入力画像に基づいて背景画像を生成する処理と、処理時刻から第1の時間幅内に撮影された入力画像に基づき生成された第1の背景画像と、前記処理時刻から第2の時間幅内に撮影された入力画像に基づき生成された第2の背景画像との差分に基づき、入力画像内の、複数の人物が映り得る部分領域に対して滞留を検出する処理と、入力画像の前記部分領域に映る1以上の人物を検出する処理と、滞留の検出結果及び人物の検出結果に基づき、前記部分領域における複数人物の滞留を検出する処理とをコンピュータに実行させるプログラム。
前記背景画像は、入力画像のうち、動きのない静止領域に基づいて生成される、付記19記載のプログラム。
前記部分領域の大きさは、入力画像を撮影する前記撮影装置の位置と、入力画像内の前記部分領域とに応じて設定される、付記19又は付記20記載のプログラム。
前記部分領域内の下部領域における滞留の検出結果の方が、上部領域における滞留の検出結果よりも、前記複数人物の滞留の検出結果に大きな影響を与える、付記19乃至付記21のいずれか1項記載のプログラム。
前記部分領域における前記複数人数の滞留の検出結果を報知する出力処理を更にコンピュータに実行させる、付記19乃至付記22のいずれか1項記載のプログラム。
前記複数人数の滞留の度合いに応じて報知方法を変える、付記23記載のプログラム。
前記複数人数の滞留の度合いが閾値を超えた時間に基づいて報知する、付記23記載のプログラム。
前記閾値はユーザが設定可能である、付記25記載のプログラム。
前記部分領域における人数の増減を検出し、前記第2の検出手段による前記人数の増減の検出結果に基づき、所定領域において徐々に人物が集合していることを示す情報を別途報知する、付記23乃至付記26のいずれか1項記載のプログラム。
Claims (11)
- 撮影装置により複数の時刻に撮影された入力画像に基づいて背景画像を生成する生成手段と、
処理時刻から第1の時間幅内に撮影された入力画像に基づき生成された第1の背景画像と、前記処理時刻から第2の時間幅内に撮影された入力画像に基づき生成された第2の背景画像との差分に基づき、入力画像内の、複数の人物が映り得る部分領域に対して滞留を検出する第1の検出手段と、
入力画像の前記部分領域に映る1以上の人物を検出する第2の検出手段と、
滞留の検出結果及び人物の検出結果に基づき、前記部分領域における複数人物の滞留を検出する第3の検出手段と
を備える画像処理システム。 - 前記背景画像は、入力画像のうち、動きのない静止領域に基づいて生成される、
請求項1記載の画像処理システム。 - 前記部分領域の大きさは、入力画像を撮影する前記撮影装置の位置と、入力画像内の前記部分領域とに応じて設定される、
請求項1又は請求項2記載の画像処理システム。 - 前記部分領域内の下部領域における滞留の検出結果の方が、上部領域における滞留の検出結果よりも、前記複数人物の滞留の検出結果に大きな影響を与える、
請求項1乃至請求項3のいずれか1項記載の画像処理システム。 - 前記部分領域における複数人数の滞留の検出結果を報知する出力手段
を更に備える、請求項1乃至請求項4のいずれか1項記載の画像処理システム。 - 前記出力手段は、前記複数人数の滞留の度合いに応じて報知方法を変える、
請求項5記載の画像処理システム。 - 前記出力手段は、前記複数人数の滞留の度合いが閾値を超えた時間に基づいて報知する、
請求項5記載の画像処理システム。 - 前記閾値はユーザが設定可能である、
請求項7記載の画像処理システム。 - 前記第2の検出手段は、前記部分領域における人数の増減を検出し、
前記出力手段は、前記第2の検出手段による前記人数の増減の検出結果に基づき、所定領域において徐々に人物が集合していることを示す情報を別途報知する、
請求項5乃至請求項8のいずれか1項記載の画像処理システム。 - 撮影装置により複数の時刻に撮影された入力画像に基づいて背景画像を生成するステップと、
処理時刻から第1の時間幅内に撮影された入力画像に基づき生成された第1の背景画像と、前記処理時刻から第2の時間幅内に撮影された入力画像に基づき生成された第2の背景画像との差分に基づき、入力画像内の、複数の人物が映り得る部分領域に対して滞留を検出するステップと、
入力画像の前記部分領域に映る1以上の人物を検出するステップと、
滞留の検出結果及び人物の検出結果に基づき、前記部分領域における複数人物の滞留を検出するステップと
を画像処理システムが行う、画像処理方法。 - 撮影装置により複数の時刻に撮影された入力画像に基づいて背景画像を生成する処理と、
処理時刻から第1の時間幅内に撮影された入力画像に基づき生成された第1の背景画像と、前記処理時刻から第2の時間幅内に撮影された入力画像に基づき生成された第2の背景画像との差分に基づき、入力画像内の、複数の人物が映り得る部分領域に対して滞留を検出する処理と、
入力画像の前記部分領域に映る1以上の人物を検出する処理と、
滞留の検出結果及び人物の検出結果に基づき、前記部分領域における複数人物の滞留を検出する処理と
をコンピュータに実行させるプログラム。
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2015534061A JP6414066B2 (ja) | 2013-08-27 | 2014-07-02 | 画像処理システム、画像処理方法及びプログラム |
US14/915,109 US9934576B2 (en) | 2013-08-27 | 2014-07-02 | Image processing system, image processing method, and recording medium |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2013176082 | 2013-08-27 | ||
JP2013-176082 | 2013-08-27 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2015029588A1 true WO2015029588A1 (ja) | 2015-03-05 |
Family
ID=52586166
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2014/067693 WO2015029588A1 (ja) | 2013-08-27 | 2014-07-02 | 画像処理システム、画像処理方法及びプログラム |
Country Status (3)
Country | Link |
---|---|
US (1) | US9934576B2 (ja) |
JP (1) | JP6414066B2 (ja) |
WO (1) | WO2015029588A1 (ja) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2016206995A (ja) * | 2015-04-23 | 2016-12-08 | キヤノン株式会社 | 画像処理装置、画像処理方法、及びプログラム |
JP2018142061A (ja) * | 2017-02-27 | 2018-09-13 | Kddi株式会社 | 顔検出装置及びプログラム |
Families Citing this family (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP6270433B2 (ja) * | 2013-11-26 | 2018-01-31 | キヤノン株式会社 | 情報処理装置、情報処理方法、情報処理システム |
EP3198572B1 (en) * | 2014-09-25 | 2020-11-11 | Guerzoni, Filippo | Surveillance method, device and system |
US10356317B2 (en) * | 2014-10-30 | 2019-07-16 | Technion Research & Development Foundation Limited | Wide-scale terrestrial light-field imaging of the sky |
US20160232672A1 (en) * | 2015-02-06 | 2016-08-11 | Qualcomm Incorporated | Detecting motion regions in a scene using ambient-flash-ambient images |
US10970896B2 (en) * | 2016-11-02 | 2021-04-06 | Canon Kabushiki Kaisha | Image processing apparatus, image processing method, and storage medium |
MY184063A (en) * | 2017-03-14 | 2021-03-17 | Mitsubishi Electric Corp | Image processing device, image processing method, and image processing program |
KR102187831B1 (ko) * | 2019-10-29 | 2020-12-07 | 곽찬우 | Cctv를 이용한 혼잡도 판단 시스템의 제어 방법, 장치 및 프로그램 |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2008217184A (ja) * | 2007-03-01 | 2008-09-18 | Giken Torasutemu Kk | 混雑レベル判別システム |
JP2010277547A (ja) * | 2009-06-01 | 2010-12-09 | Chuo Electronics Co Ltd | 混雑状況検出装置、方法、およびプログラム |
JP2013011950A (ja) * | 2011-06-28 | 2013-01-17 | Canon Inc | 画像処理装置、画像処理方法及びプログラム |
Family Cites Families (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP4852355B2 (ja) | 2006-06-26 | 2012-01-11 | パナソニック株式会社 | 放置物検出装置及び放置物検出方法 |
US7813528B2 (en) | 2007-04-05 | 2010-10-12 | Mitsubishi Electric Research Laboratories, Inc. | Method for detecting objects left-behind in a scene |
US8472715B2 (en) * | 2007-10-26 | 2013-06-25 | Panasonic Corporation | Situation determining apparatus, situation determining method, situation determining program, abnormality determining apparatus, abnormality determining method, abnormality determining program, and congestion estimating apparatus |
JP4966820B2 (ja) | 2007-10-29 | 2012-07-04 | パナソニック株式会社 | 混雑推定装置および方法 |
JP2010198566A (ja) | 2009-02-27 | 2010-09-09 | Nec Corp | 人数計測装置、方法及びプログラム |
-
2014
- 2014-07-02 WO PCT/JP2014/067693 patent/WO2015029588A1/ja active Application Filing
- 2014-07-02 US US14/915,109 patent/US9934576B2/en active Active
- 2014-07-02 JP JP2015534061A patent/JP6414066B2/ja active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2008217184A (ja) * | 2007-03-01 | 2008-09-18 | Giken Torasutemu Kk | 混雑レベル判別システム |
JP2010277547A (ja) * | 2009-06-01 | 2010-12-09 | Chuo Electronics Co Ltd | 混雑状況検出装置、方法、およびプログラム |
JP2013011950A (ja) * | 2011-06-28 | 2013-01-17 | Canon Inc | 画像処理装置、画像処理方法及びプログラム |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2016206995A (ja) * | 2015-04-23 | 2016-12-08 | キヤノン株式会社 | 画像処理装置、画像処理方法、及びプログラム |
JP2018142061A (ja) * | 2017-02-27 | 2018-09-13 | Kddi株式会社 | 顔検出装置及びプログラム |
Also Published As
Publication number | Publication date |
---|---|
JP6414066B2 (ja) | 2018-10-31 |
US20160210756A1 (en) | 2016-07-21 |
US9934576B2 (en) | 2018-04-03 |
JPWO2015029588A1 (ja) | 2017-03-02 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP6414066B2 (ja) | 画像処理システム、画像処理方法及びプログラム | |
JP6741130B2 (ja) | 情報処理システム、情報処理方法及びプログラム | |
US10713798B2 (en) | Low-complexity motion detection based on image edges | |
JP6525545B2 (ja) | 情報処理装置、情報処理方法及びコンピュータプログラム | |
JP3801137B2 (ja) | 侵入物体検出装置 | |
JP6436077B2 (ja) | 画像処理システム、画像処理方法及びプログラム | |
JP6652057B2 (ja) | 画像から移動体の滞留を検出するための画像処理システム、画像処理方法及びプログラム | |
JP6233624B2 (ja) | 情報処理システム、情報処理方法及びプログラム | |
US20150077568A1 (en) | Control method in image capture system, control apparatus and a non-transitory computer-readable storage medium | |
US8532191B2 (en) | Image photographing apparatus and method of controlling the same | |
Ha et al. | Foreground objects detection using multiple difference images | |
US20150146006A1 (en) | Display control apparatus and display control method | |
JP5360052B2 (ja) | 物体検出装置 | |
JP6551226B2 (ja) | 情報処理システム、情報処理方法及びプログラム | |
US20190230269A1 (en) | Monitoring camera, method of controlling monitoring camera, and non-transitory computer-readable storage medium | |
KR101460317B1 (ko) | 불안정한 카메라 환경에서의 이동 객체 검출 장치 및 방법 | |
JP2004228770A (ja) | 画像処理システム | |
JP6598952B2 (ja) | 画像処理装置及び方法及び監視システム | |
JP5268796B2 (ja) | 移動物体領域検出装置及び移動物体領域検出プログラム | |
JP7338174B2 (ja) | 物体検出装置および物体検出方法 | |
JP2011198261A (ja) | 対象物認識システム及び該システムを利用する監視システム、見守りシステム | |
JP2012128693A (ja) | 映像処理装置、映像処理方法およびプログラム | |
JP5834671B2 (ja) | 画像処理装置、画像処理方法およびプログラム | |
US20230222672A1 (en) | Image processing apparatus, image processing method, and non-transitory computer-readable storage medium | |
JP6308612B2 (ja) | 画像処理装置、画像処理方法、および画像処理プログラム |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 14841108 Country of ref document: EP Kind code of ref document: A1 |
|
ENP | Entry into the national phase |
Ref document number: 2015534061 Country of ref document: JP Kind code of ref document: A |
|
WWE | Wipo information: entry into national phase |
Ref document number: 14915109 Country of ref document: US |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 14841108 Country of ref document: EP Kind code of ref document: A1 |