JPWO2018084191A1 - Congestion situation analysis system - Google Patents

Abstract

It is possible to analyze the congestion situation considering not only the number and movement of objects obtained from the image information of the imaging device but also human senses such as crampedness and stress that vary depending on the environmental state and the objects carried or used by people. Provide a possible congestion situation analysis system. In a congestion analysis system comprising an imaging device that captures an image of a region to be analyzed, and a congestion analysis device that processes the image captured by the imaging device and collects information in the analysis target region, the congestion analysis device Includes an object information analysis unit, and the object information analysis unit extracts a person / object information extraction unit that extracts information about a person and an object included in an image captured by the imaging device, and a person extracted by the person / object extraction unit. The type judgment means for judging the type of the object, the personal space of the person extracted by the person / object extraction means, and the occupancy area for each type of the object judged by the type judgment means, including the tightness felt by humans An estimated number of persons calculation means for calculating the estimated number of persons, and a congestion degree determination means for determining a degree of congestion close to a human sense based on the estimated number of persons calculated by the estimated number of persons calculation means. It is a sign.

Description

  The present invention relates to a congestion situation analysis system using video, and more particularly to a congestion situation analysis system capable of analyzing a congestion situation in consideration of a human stress level.

  In a congestion situation analysis system, for example, an object that enters an analysis target area is imaged using an imaging device such as a television camera (TV camera), and the number and movement of objects existing in the area from the captured image. To determine whether or not congestion occurs in the area. In addition, in a congestion situation analysis system, a congestion situation analysis system that automatically analyzes the system instead of manned monitoring by the user and informs the user of the congestion situation according to the situation has been conventionally studied (for example, patent documents). 1).

  In addition, as a technique for detecting an object that enters an analysis target region, for example, a detection technique using an image processing method called a background subtraction method has been widely used. Thus, the background difference method will be described with reference to FIGS. FIG. 8 is a diagram for explaining an outline of a procedure for detecting a region whose luminance value has changed by the background subtraction method, and FIG. 9 shows a typical processing procedure of a monitoring method applying the background subtraction method. It is a flowchart to show.

First, the procedure for detecting the luminance change region of the input image sequentially input from the imaging device by the background subtraction method will be described with reference to FIG.
In FIG. 8, 401 is an input image that is sequentially input from the imaging apparatus, and 402 is a background image that is prepared in advance and does not show the target object to be detected. When the two input images 401 and the background image 402 are input to the differentiator 405, the difference between the luminance values for each pixel is calculated by the differencer 405 to obtain a difference image 403.
Next, when the difference image 403 is input to the binarizer 406, the binarizer 406 sets each pixel of the difference image 403 to a threshold value Th (Th is experimentally determined, for example, Th = 20). The threshold value processing is performed to obtain a binarized image 404 with a pixel value less than the threshold Th as “0” and a pixel value greater than or equal to the threshold Th as “255”. As a result, the humanoid object 407 shown in the input image 401 is calculated as a region 408 in which a difference is generated by the differentiator 405 (changed luminance region of the input image), and the pixel value “255” is calculated by the binarizer 406. "Is detected as an image 409.

Next, a typical processing procedure of the object detection method applying the background difference method will be described with reference to FIG.
In FIG. 9, in an initialization processing step S501, initialization of external devices, variables, image memory, etc. for realizing the object detection method based on the background difference method is performed.
In the image input step S502, an input image having a width of 640 pixels and a height of 480 pixels, for example, is obtained from the imaging device.
In the difference processing step S503, the luminance value of each pixel of the input image obtained in step S502 (see the input image 401 in FIG. 8) and the reference background image prepared in advance (see the background image 402 in FIG. 8). Is calculated (difference image 403).
In the binarization processing step S504, the pixel value (difference value) of the difference image 403 obtained in step S503 is set to a pixel value that is less than a predetermined threshold Th (for example, Th = 20) is set to “0”. The pixel value of a pixel equal to or greater than Th is “255” (the pixel value of one pixel is calculated by 8 bits. That is, one pixel has a value from 0 to 255), and the binarized image (binarized image of FIG. 8) 404).

Next, in the labeling processing step S505, a block of pixels (see the image 409 in FIG. 8) having the pixel value “255” in the binarized image 404 obtained in step S504 is detected and numbered. So that they can be distinguished.
In the object presence determination step S506, for each of the numbered change areas, it is determined whether the change area matches the detection condition based on the size, area, and the like. If the change area matches the detection condition (YES) ), In the object information output step S507, the position and size of the object are output by a method as necessary, and the congestion status is determined using the total number of objects determined from the output information. .
If the detection condition is not met in step S506 (NO), the process branches to background update step S508 because no object exists.
In the background update step S508, the background image is updated using the input image obtained in step S502. As a background image update method, for example, there is a method in which a current background image is weighted and averaged by applying a certain weight (update rate) to a current background image, and new background images are sequentially generated.

  Such a congestion situation analysis system is used for various purposes. For example, it is used for the purpose of preventing incidents and accidents in facilities visited by an unspecified large number of people such as train stations, airports, stadiums, and shopping malls. Has been. In addition, the recording devices used in the system have various data recorded based on output information from long-term image data, and a function for predicting congestion in advance based on these accumulated data It is often equipped with.

JP 2013-122735 A

  As described above, in the conventional congestion situation analysis system, the degree of congestion is determined using video information from the imaging device. However, because the degree of congestion that people perceive depends on the environmental conditions, it is far from the congestion status determined by the congestion status analysis system, and even if it is determined that the image information is not crowded, You may feel uncomfortable.

For example, on a rainy day in the summer, most people move with an umbrella, but when analyzing the congestion status with the congestion status analysis system, the volume of the umbrella is very small compared to the person. It is judged that there is almost no difference in the image from the normal time (for example, at the time of fine weather in early spring), and the same congestion result as the normal time is output.
However, in reality, the floor in the station is wet, so it moves slowly and naturally because of falling, and the temperature and humidity are high and you have a wet umbrella. The environmental condition is worse than. For this reason, compared to the normal time, the stress level of the person increases, and the “personal space” that is uncomfortable when approaching others is more widespread. It will be difficult to walk just because there are people, and it will be crowded. Therefore, when it is determined that the congestion state analysis system is congested in this adverse environment, it is already recognized as a heavy congestion for stressed passers-by. For example, in such an environment, for example, Incidence of incidents and accidents such as violent incidents caused by wet umbrellas is higher than usual.
In this way, the degree of congestion felt by humans varies greatly depending on environmental conditions and incidental items, and unexpected incidents and accidents may occur.

  The present invention has been made in view of such circumstances, and not only the number and movement of objects obtained from image information of an imaging apparatus, but also the tightness and stress that vary depending on environmental conditions and objects carried or used by people. The purpose is to provide a congestion status analysis system that can analyze the congestion status in consideration of human senses.

  In order to achieve the above object, a congestion situation analysis system according to the present invention includes an imaging device that captures an image of a region to be analyzed, and processes the image captured by the imaging device to collect information in the analysis target region. In the congestion situation analysis system configured by the congestion analysis device, the congestion analysis device includes an object information analysis unit, and the object information analysis unit includes a person and an object included in an image captured by the imaging device. A person / object information extracting means for extracting the information, a type determining means for determining the type of the person and the object extracted by the person / object extracting means, a personal space of the person extracted by the person / object extracting means, And an estimated number calculating means for calculating an estimated number of persons including a cramp felt by a person using the occupied area for each type of object determined by the type determining means, and the estimated number calculating means And a congestion degree determination means for determining a congestion degree close to a human sense based on the estimated number of persons calculated in a stage.

  In order to achieve the above object, the congestion status analysis system of the present invention is the above-described congestion status analysis system, wherein the congestion analysis device includes an environment parameter generation unit, and the environment parameter generation unit includes at least Environmental parameter calculation means for calculating environmental parameters based on weather information and weather information from one weather sensor, and the object information analysis unit uses the environmental parameters calculated by the environmental parameter calculation means, and uses the estimated number of persons. An environmental number calculating means for calculating the number of environmental persons reflecting the environmental condition in the analysis target area is provided for the estimated number of persons calculated by the calculating means, and the congestion degree determining means is calculated by the environmental number calculating means. Determining the degree of congestion including the degree of human stress due to the environmental condition in the analysis target area based on the number of environmental persons It is characterized by.

  In order to achieve the above object, the congestion status analysis system of the present invention is the above-described congestion status analysis system, wherein the environmental parameter calculation means has an environment table and a weather table, and is acquired from the environment table. The environmental parameter indicating the discomfort level of the person in the analysis target area is calculated using the stress parameter obtained and the weather parameter acquired from the weather table.

  According to the present invention, not only the number and movement of objects obtained from the image information of the imaging device, but also the situation of congestion taking into account human senses such as crampedness and stress that vary depending on environmental conditions and objects carried or used by people. Can be analyzed.

It is a block diagram showing an example of a configuration of a congestion situation analysis system according to an embodiment of the present invention. It is a figure which shows an example of the object information table of the congestion condition analysis system which concerns on one Embodiment of this invention. It is a figure which shows an example of the environment table of the congestion condition analysis system which concerns on one Embodiment of this invention. It is a figure which shows an example of the weather table of the congestion condition analysis system which concerns on one Embodiment of this invention. It is a flowchart which shows an example of the process in the congestion condition analysis system which concerns on one Embodiment of this invention. It is a figure for demonstrating the input area in the congestion condition analysis system which concerns on one Embodiment of this invention, and the occupation area by the person and object in the said input image, and an object classification. It is a flowchart which shows an example of the process in the environmental parameter production | generation part of the congestion condition analysis system which concerns on one Embodiment of this invention. It is a figure for demonstrating the outline of the procedure which detects the area | region where the luminance value changed with the background difference method. It is a flowchart which shows the typical process sequence of the monitoring system which applied the background difference method.

Hereinafter, a congestion situation analysis system according to an embodiment of the present invention will be described.
A congestion situation analysis system according to an embodiment of the present invention includes an imaging device that captures an image of a region to be analyzed, a congestion analysis device that processes an image captured by the imaging device and collects information in the analysis target region, and In the congestion situation analysis system composed of: the congestion analysis device includes an object information analysis unit, and the object information analysis unit extracts a person / object information included in an image captured by the imaging device. Information extracting means, type determining means for determining the type of person and object extracted by the person / object extracting means, personal space of the person extracted by the person / object extracting means, and object type determined by the type determining means Based on the estimated number of people calculated by the estimated number of people calculated by the estimated number of people calculated by the estimated number of people including the cramped feeling felt by the person and the estimated number of people And a congestion degree determination means for determining the degree of the fluctuation, which varies depending not only on the number and movement of objects obtained from the image information of the imaging device, but also on the environmental state and the object carried or used by the person. The situation of congestion can be analyzed in consideration of human senses such as cramping and stress.

[Configuration of congestion analysis system]
A configuration of a congestion situation analysis system according to an embodiment of the present invention will be described with reference to FIGS. FIG. 1 is a block diagram showing an example of the configuration of a congestion situation analysis system according to an embodiment of the present invention. FIG. 2 is a diagram showing an example of an object information table of the congestion situation analysis system according to the embodiment of the present invention. FIG. 3 is a diagram showing an example of an environment table of the congestion situation analysis system according to the embodiment of the present invention. FIG. 4 is a diagram illustrating an example of a weather table of the congestion situation analysis system according to the embodiment of the present invention.

  As shown in FIG. 1, the congestion status analysis system of the present invention includes an imaging device 10, a congestion analysis device 20 having an object information analysis unit 30 and an environment parameter generation unit 40, a recording device 50, and a display device 60. It is configured. Further, the imaging device 10 is connected to an image input I / F (interface) (not shown) of the congestion analysis device 20, and the display device 60 and the recording device 50 are connected to an image output I / F (not shown) of the congestion analysis device 20. Connected to F (interface). The weather sensor 70 and the weather information acquisition device 80 are connected to an I / F (interface) of the congestion analysis device 20. The imaging device 10 may be connected to the congestion analysis device 20 via a network. In addition, the weather information acquisition device 80 may be connected to the congestion analysis device 20 via a network. In FIG. 1, the imaging apparatus 10 is configured as one unit, but a plurality of imaging apparatuses may exist.

The imaging device 10 is connected to an image input I / F (not shown) of the congestion analysis device 20, captures an image of a region to be analyzed, and the captured image data is input to the image input unit 31 of the congestion analysis device 20. Output.
The object information analysis unit 30 of the congestion analysis device 20 includes an image input unit 31, a motion information extraction unit 32, a person / object analysis unit 33, a congestion determination unit 34, and an object information table 35.
The image input unit 31 acquires image data captured from the imaging device 10, for example, an input image having a width of 640 pixels and a height of 480 pixels.
The motion information extraction unit 32 extracts an area in which an object exists from the input image input from the image input unit 31, an image showing a range in which a person or an object exists (hereinafter, an extracted image), and its region Create information.

  The person / object analysis unit 33 determines the total number of persons and the class of the object from the extracted image or the like input from the motion information extraction unit 32. In addition, the person / object analysis unit 33 uses the input image or the extracted image to extract, for example, “size”, “speed”, “traveling direction (when entering and leaving) from the person or object existing in the extracted image. ) "," Contour shape ", etc., for example," person only "," stroller "," accompanied by assisting dog "," carrying large luggage (suitcase or large sports bag) "," wheelchair use ”,“ Carrying white cane ”,“ with children ”and so on. The person / object analysis unit 33 refers to a person / object information table 35 to be described later with respect to the total number of persons, and recalculates the number of persons (estimated number) reflecting the type of object.

  As shown in FIG. 2, the person / object information table 35 includes a type 35 a, a type number 35 b, and an occupied area (width × depth) 35 c, and an occupied area 35 c “in the case of the type 35 a“ person ”. 120 × 120 ”is used as a reference (type number 35b“ 1.0 ”), and the number of types calculated by adding the occupied area that increases in the case of other types as an additional number is registered. In addition, the occupation area 35c of the standard type 35a “person” is assumed to be 60cm when the hand is extended, and the interpersonal distance that the fingertip reaches to the person is reached. The space to be set is set as, for example, a rectangle having a width of 120 cm and a depth of 120 cm.

  The congestion determination unit 34 calculates the number of people (hereinafter referred to as “environment number”) in consideration of the degree of human stress from the estimated number of people calculated by the person / object analysis unit 33 and the environmental parameters acquired from the environmental parameter generation unit 40. In addition, the congestion determination unit 34 determines the degree of congestion using the number of people in the environment, and outputs a determination result (for example, a quiet state, a normal state, a congestion state, etc.).

The environment parameter generation unit 40 of the congestion analysis apparatus 20 includes an environment analysis unit 41, an environment table 42, and a weather table 43, calculates environment parameters to be described later, and outputs them to the congestion determination unit 34.
The environment analysis unit 41 calculates the discomfort index based on the weather data acquired from the weather sensor 70, and based on the calculated discomfort index and the weather information acquired from the weather information acquisition device 80, the environment table 42 and the weather described later A stress parameter and a weather parameter are determined with reference to the table 43, an environmental parameter is calculated from the stress parameter and the weather parameter, and the environmental parameter is output to the congestion determination unit 34.

  As shown in FIG. 3, the environment table 42 includes a discomfort index 42a, a stress parameter 42b, and a general sense 42c, and a stress parameter 42b corresponding to the numerical value of the discomfort index 42a is set and registered. The stress parameter is a value indicating the degree of stress felt by humans with respect to the discomfort index. As shown in FIG. 3, the stress parameter increases in proportion to the increase in the discomfort index. That is, according to this environmental table 42, for example, if the discomfort index 42a is 76 or more, some people are said to be uncomfortable, so the stress parameter is 1.3 and the discomfort index 42a is 81 or more. If so, the stress parameter 42b is 1.5 because it is said that everyone becomes uncomfortable.

  As shown in FIG. 4, the weather table 43 includes a weather 43a, a weather parameter 43b, and a general sense 43c, and the weather parameter 43b corresponding to the weather 43a is set and registered. The weather parameter is a value indicating the degree of stress felt by humans with respect to the weather condition. That is, according to the weather parameter 43, for example, the weather parameter 43b is 1.0 when the weather 43a is fine, whereas the weather parameter 43b is 1.2 when the weather is rainy (less than 1 mm of precipitation). The weather parameter 43b in rainy weather (rainfall of 1mm or more) is 1.5, and there are more unpleasant factors such as possession of wet umbrellas, wearing boots and raincoats, and slow movement on wet floors and stairs as the weather gets worse Therefore, the parameter value is set to be large.

The meteorological sensor 70 outputs the meteorological data in the analysis target area to the environment analyzing unit 41 via the network in response to an instruction from the environment analyzing unit 41. The weather sensor 70 is, for example, a temperature sensor, a humidity sensor, or a rain sensor, a wind speed sensor, a wind direction sensor, a barometric pressure sensor, etc. when outdoors. In this embodiment, assuming that the analysis target area is indoors, meteorological data acquired from the indoor weather sensor 70 installed in the vicinity of the analysis target area is stored in the “temperature sensor” and “humidity sensor”. The information obtained from the two types is used. When the analysis target area is outdoors, the most suitable sensor is used outdoors.
In response to an instruction from the environment analysis unit 41, the weather information acquisition device 80 outputs the weather information of the analysis target area to the environment analysis unit 41 via the network. For example, the weather information acquisition device 80 periodically acquires the weather information of the analysis target area disclosed on the HP (homepage) of the Japan Meteorological Association and stores it in the storage unit.

[Operation of congestion analysis system]
Next, in the congestion situation analysis system according to one embodiment of the present invention, the flow chart of FIG. 5 shows the procedure and contents until the congestion result analyzed based on the captured object information and the imaging environment is displayed. It explains using. FIG. 5 is a flowchart showing an example of processing in the congestion situation analysis system according to the embodiment of the present invention.
As shown in FIG. 5, in the image input step S <b> 201, the image input unit 31 of the object information analysis unit 30 of the congestion analysis device 20 performs image data captured from the imaging device 10, for example, a width of 640 pixels and a height of 480 pixels. Get the input image.
In the object extraction step S202, the motion information extraction unit 32 of the object information analysis unit 30 extracts an area where an object exists from the input image acquired in step S201. The processing in this step corresponds to the processing from step S502 to step S508 described with reference to FIG. 9 of the prior art, and an image (hereinafter referred to as an extracted image) showing a range where a person or an object exists and its region information Create

  In the type analysis step S203, the person / object analysis unit 33 of the object information analysis unit 30 determines the total number of persons and the class of the object from the extracted image or the like acquired in step S202. The person / object analysis unit 33 uses the input image or the extracted image to extract, for example, “size”, “speed”, “traveling direction (when entering and leaving)” from the person or object existing in the extracted image. Based on multiple features such as “contour shape”, for example, “person only”, “stroller”, “accompanied by service dog”, “carrying large luggage (suitcase or large sports bag)”, “use of wheelchair” , “Carrying a white cane”, “with children” and the like. As a means of classification, SVM (Support Vector Machine), known discriminators using machine learning such as random forest and deep learning, and analysis methods can be used, and learning data etc. are prepared in advance if necessary. Keep it.

  In the estimated number calculation step S204, the person / object analysis unit 33 of the object information analysis unit 30 refers to the person / object information table 35 described above for the total number of persons and reflects the number of objects (estimation). Recalculation of the number of people). For example, referring to the object information table 35, the occupation area 35 c of the standard type 35 a “person” is 120 cm wide × 120 cm deep, and the type number 35 b is 1.0. Further, since the occupation area 35c of the type 35a “stroller” is (the main body only) 100 cm wide × 100 cm deep, the type number 35b assumes that the “person” is pushing the “stroller”. It is set to 8 people.

Therefore, an example of the estimated number calculation process performed in the estimated number calculation step S204 will be described with reference to FIG. FIG. 6 is a diagram for explaining an input image in the congestion state analysis system according to the embodiment of the present invention, and an occupied area by a person, an object, and an object type in the input image.
FIG. 6A shows an example of the input image acquired in the image input step S201. In FIG. 6A, passers-by 901 to 903 can be confirmed, and it is determined that there are three persons in the analysis target area in the image recognition process performed in the existing system. Further, in the type analysis step S203, the passerby 901 is the type 35a “person”, the passerby 902 is the type 35a “white cane (displayed as a black cane in the drawing)”, and the passerby 903 is the “Large luggage”.
Here, paying attention to the passer-by 903, since the passer-by 903 has a suitcase, it is determined as the type 35a “large baggage”, and therefore the object region obtained in the object extraction step S202 has a baggage. There are two types: “persons” and “large luggage” that is carried.

That is, in the estimated number of people calculation step S204, the size of the object is 170 cm (120 cm + 50 cm) × 120 cm in depth, and the occupied area 1003 of the object is larger than the occupied area 1001 of the type 35a “person”. . Using the person / object information table 35, this occupied area 1003 is converted into the number of people of a type based on the type 35a “person”. The result is 4 people.
Accordingly, in the estimated number of persons calculation step S204, as a result of adding up the occupied areas 1001 to 1003 corresponding to the types for all passers-by 901 to 903 in FIG. 6B, the input image in FIG. The estimated number of people is 4.2 (1.0 +1.4 +1.8). Thus, in this step, it is possible to calculate the estimated number of people taking into account that the area occupied during movement varies greatly depending on the type of object.

In the present embodiment, the process for calculating the estimated number of persons from the occupied area with the size of each type including “person” fixed is described. However, based on the extraction result, the size of each object is calculated. By calculating the estimated number of people using the area occupied by adding personal space, it is possible to calculate a more accurate estimated number of people.
Also, if it is determined from the extracted image that the density of objects is very high, it is very difficult to determine the type of individual person / object. Therefore, in the type analysis step S203, only the number of people that can be estimated from the extracted image is calculated. Thus, this may be determined as the estimated number of people, and the estimated number of people calculation step S204 may be skipped.

  Next, in the environmental parameter update determination step S205, the environmental analysis unit 41 of the environmental parameter generation unit 40 receives environmental parameters from a storage unit (not shown) of the object information analysis unit 30 (in this embodiment, from the weather sensor 70). It is determined whether it is time to update the stress parameter based on the weather data and the correction parameter based on the weather parameter based on the weather information from the weather information acquisition device 80. For example, if judgment conditions are set in advance and, for example, a set time (for example, 1 minute) has passed since the previous parameter update, or if the acquired weather information is different from the previously acquired information, Or, when the judgment conditions such as a sudden increase or decrease in the number of people within a certain time are satisfied (YES), the environmental parameter acquisition step S206 is performed, and when the above conditions are not satisfied (NO) The process proceeds to the number of people calculation step S207.

  In the environment parameter acquisition step S206, the environment analysis unit 41 outputs an environment parameter corresponding to the imaging environment in the analysis target area by the imaging device 10 to the congestion determination unit 34.

Therefore, processing performed by the environment parameter generation unit 40 will be described with reference to FIG. FIG. 7 is a flowchart showing an example of processing in the environment parameter generation unit of the congestion situation analysis system according to the embodiment of the present invention.
In the weather data acquisition step S <b> 301, the environment analysis unit 41 of the environment parameter generation unit 40 acquires weather data from the weather sensor 70.

In the discomfort index calculation step S302, the environment analysis unit 41 calculates a discomfort index (DI) using the weather data of the analysis target area acquired in step S301. Here, the discomfort index is an index that quantitatively represents the summer heat and heat, and is calculated by the following (Formula 1) when the temperature is Td [° C.] and the humidity is H [%].
Discomfort index (DI) = 0.81Td + 0.01H × (0.99Td-14.3)
+46.3 (Equation 1)

  In the stress parameter setting step S303, the environment analysis unit 41 determines a stress parameter with reference to the environment table 42 of FIG. 3 described above based on the discomfort index calculated in step S302.

  In the weather information collection step S304, the environment analysis unit 41 acquires weather information from the weather information acquisition device 80, for example, via a network. The said weather information shows the present weather data etc. of the vicinity (for example, municipality) where the imaging device 10 is installed.

  In the weather parameter setting step S305, the environment analysis unit 41 determines the weather parameter with reference to the weather table 43 in FIG. 4 described above based on the weather information acquired in step S304.

  In the weather table 43 of FIG. 4, the weather parameter 43b is a uniform parameter according to the weather, but parameters are prepared in advance for each item that can be stressed, such as an umbrella, and so on. If the angle of view is clear enough to grasp the details, the weather parameters may be changed for each person carrying these objects, and the calculation may be performed according to the stress level of each person. However, for example, if the carrying item is an “umbrella”, it is assumed that the umbrella is wet when it is raining or snowing, and it is assumed that the umbrella is dry during other weather conditions. There is also a need for parameters that take into account the relevance to the weather.

In addition, the weather table 43 described above shows the values of the weather parameters when indoor congestion judgment is assumed, but in the case of the outdoors, people wear umbrellas during rainy weather or snowfall, so it is normal However, since a personal space is automatically secured by inserting an umbrella, it is necessary to create a weather table 43 in consideration of reducing stress and the like.
Further, in the environment analysis unit 41, if a rain sensor or the like can be installed in a place where there is no difference between the installation position of the imaging device 10 and the weather, the rainfall information output from the rain sensor instead of the weather information collecting step S304. May be obtained.

In environment analysis step S306, the environment analysis unit 41 calculates an environment parameter from the stress parameter set in step S303 and the weather parameter set in step S305. The calculation method is shown in (Formula 2) below.
Environmental parameter = Stress parameter × Weather parameter (Equation 2)
In this example, two parameters, stress parameter and weather parameter, are used to calculate the environmental parameters. However, the number of items can be increased according to the weather sensor used and weather information. This makes it possible to calculate more optimal environmental parameters. In particular, it is difficult to extract and discriminate from images.For example, folded umbrellas have a large impact on the degree of stress, but for relatively small items to be carried, environmental parameters can be used instead of measuring the number of people based on the occupied area. Effective utilization enables more appropriate congestion analysis.

  In the environmental parameter output step S307, the environmental analysis unit 41 outputs the environmental parameter calculated in step S306 to the congestion determination unit 34, and proceeds to the environmental number calculation step S207 in FIG.

In the environmental number calculation step S207, the congestion determination unit 34 calculates the number of persons considering the degree of human stress (hereinafter, environmental number) from the estimated number of persons calculated in step S204 and the environmental parameters acquired in step S206. The calculation method of the number of people in the environment is shown below (Formula 3).
Number of people in the environment = Estimated number of people x Environment parameters (Equation 3)
As a result, even if the estimated number of people is the same, the number of people in the environment changes greatly if the environment is different. For example, if the estimated number of people is 15, the number of people in the environment with a cloudy / discomfort index of 70 remains as it is, but the number of people in the environment with a rainy weather (less than 1 mm of precipitation) and a discomfort index of 80 is 23 ( 15 x 1.2 x 1.3), an increase of 8 people.

In this implementation, an explanation was given of calculating the number of people in the environment based on the estimated number of people and environmental parameters.However, even if there are a large number of people, the attendance time zone may vary depending on the time zone. Since the average speed of the people flow also changes, the number of people in the environment that reflects the speed of movement with time can also be calculated. Furthermore, since there is a difference in movement speed between weekdays and holidays even in the same time zone, changes in the estimated number of people according to the day of the week may be incorporated.
In addition, in an environment in which a person cannot move in a straight line direction unless he / she moves meandering, the number of people in the environment reflecting the degree of straight advance according to the person's trajectory may be calculated in consideration of the high degree of stress.

  In the congestion degree determination step S208, the congestion determination unit 34 determines the degree of congestion using the number of people in the environment calculated in step S207. The congestion determination unit 34 presets a threshold value for the number of people to be congested (hereinafter referred to as a congestion threshold value), and determines that it is “congested” if the number of environmental people is greater than or equal to the congestion threshold value. If it is less than the congestion threshold, it is determined that the state is “normal”. Alternatively, the degree of congestion is determined in three stages, and the threshold value in the case of providing in three stages, for example, the threshold value of “quiet state” is 5 or less and the threshold value of “normal state” is 20 Depending on the number of people in the environment, it may be possible to make judgments in three stages: quiet, normal, and crowded. Alternatively, the maximum number of people existing in the image may be set in advance, the congestion rate may be calculated by setting the maximum number of people to 100%, and this may be used as the determination result.

  In determination result output step S209, the congestion determination unit 34 outputs a determination result corresponding to the degree of congestion determined in step S208. The congestion determination unit 34, for example, draws the determination result of the congestion degree determination step S208 on the image of the imaging device 10 displayed on the display device 60, or outputs an alarm when it is determined that the congestion state is present. To inform the user. Further, the input video and the congestion determination result acquired from the imaging device 10 are stored in the recording device 50 as necessary.

  In the information update step S210, the congestion determination unit 34 holds the various parameters, environmental number of persons, and determination results calculated up to step S208 for a preset time. For example, in the case where the number of people in the environment and the determination result in the congestion degree determination step S208 change suddenly within a set fixed time, the information on these abnormal states is also output separately from the congestion determination result. To let you know. In the information update step S210, the congestion determination unit 34 averages and updates various parameters and the number of environmental persons calculated for each time zone and day of the week. If there is a certain difference between the value and each value acquired in the congestion degree determination step S208, the result may be output to notify the user.

Also, in order to make it easy to separate individual persons on the image even in a crowded state, overhead imaging may be performed by installing the imaging device 10 on the ceiling or the like. A distance image may be obtained from the parallax of two camera images of the objects, and the position of each object and the total number of people may be accurately acquired.
In this embodiment, the congestion degree is determined by using the number of people. However, the occupation area by each type and the total area are calculated, and the congestion degree is determined by the ratio to the area of the entire analysis target area. Also good.
It can also be used to improve the environment, such as by adjusting the temperature and humidity appropriately to reduce the stress level of the person based on the acquired number of environmental people and the environmental parameters used. .

  As described above, according to the digital congestion situation analysis system according to an embodiment of the present invention, not only the number and movement of objects obtained from video information of the imaging apparatus, but also environmental conditions and objects carried or used by people. It is possible to analyze the congestion situation taking into consideration human senses such as cramping and stress that vary depending on the situation.

  Note that the configuration and operation of the above-described embodiment are merely examples, and it is needless to say that the configuration and operation can be appropriately changed and executed without departing from the spirit of the present invention.

  The present invention relates to a congestion situation analysis system using video, and is particularly applied to a congestion situation analysis system that analyzes a congestion situation in consideration of a human stress level.

  10: imaging device, 20: congestion analysis device, 30: object information analysis unit, 31: image input unit, 32: motion information extraction unit, 33: person / object analysis unit, 34: congestion determination unit, 35: person / Object information table, 35a: type, 35b: number of people, 35c: occupied area (width × depth), 40: environment parameter generation unit, 41: environment analysis unit, 42: environment table, 42a: discomfort index, 42b: Stress parameter, 42c: General sense, 43: Weather table, 43a: Weather, 43b: Weather parameter, 43c: General sense, 50: Recording device, 60: Display device, 70: Weather sensor, 80: Weather information acquisition device , 401: input image, 402: background image, 403: difference image, 404: binarized image, 405: differencer, 406: binarizer, 407: humanoid object, 408: luminance change area Area, 409: Image of pixel value “255”, 901, 902, 903: Passerby, 1001, 1002, 1003: Occupied area.

Claims (4)

In a congestion situation analysis system comprising: an imaging device that captures an image of a region to be analyzed; and a congestion analysis device that processes the image captured by the imaging device and collects information in the analysis target region.
The congestion analysis device has an object information analysis unit,
The object information analysis unit
Person / object information extracting means for extracting information about a person and an object included in an image captured by the imaging device;
Type determination means for determining the type of person and object extracted by the person / object extraction means;
Estimated number calculation that calculates the estimated number of people including the cramps felt by humans using the personal space of the person extracted by the person / object extraction means and the occupied area for each type of object determined by the type determination means Means,
Congestion degree determination means for determining a congestion degree close to a human sense based on the estimated number of persons calculated by the estimated number of persons calculation means;
Congestion situation analysis system characterized by comprising. In the congestion situation analysis system according to claim 1,
The congestion analysis apparatus has an environment parameter generation unit,
The environmental parameter generation unit includes environmental parameter calculation means for calculating environmental parameters based on weather information and weather information from at least one weather sensor,
The object information analysis unit uses the environmental parameters calculated by the environmental parameter calculation means, and reflects the environmental number in the analysis target area with respect to the estimated number of persons calculated by the estimated number calculation means. Environmental number calculation means for calculating
The congestion degree determination means determines a congestion degree including a human stress level according to an environmental state in the analysis target area based on the environmental number of persons calculated by the environmental number of persons calculation means. Congestion situation analysis system. In the congestion situation analysis system according to claim 1,
The environmental parameter calculation means has an environment table and a weather table, and uses the stress parameter acquired from the environment table and the weather parameter acquired from the weather table to indicate the discomfort level of the person in the analysis target area A congestion situation analysis system characterized by calculating environmental parameters. In the congestion situation analysis system according to claim 2,
The environmental parameter calculation means has an environment table and a weather table, and uses the stress parameter acquired from the environment table and the weather parameter acquired from the weather table to indicate the discomfort level of the person in the analysis target area A congestion situation analysis system characterized by calculating environmental parameters.

Images