JP2019091437A - Target detection system, method for detecting target, and program - Google Patents

Abstract

To provide a target detection system that can conduct optimum detection processing while reducing the load on a system even if it is difficult to specify a detection target.SOLUTION: In the target detection system, a server 30 includes: identification information sending means 20 equipped to a detection target, the identification information sending means sending identification information of the detection target; detection means 40 for receiving a signal sent from the identification information sending means (tag 20) and detecting the detection target; imaging means 50; image analysis means 32 for performing image processing on an image taken by the imaging means 50 and estimating at least the position of the detection target; and control means for exercising control so that the image analysis means 32 chooses different processing to perform between when the detection target is detected by the detection means 40 in the imaging region of the imaging means 50 and when the detection target is not detected.SELECTED DRAWING: Figure 2

Description

  The present invention relates to an object detection system, an object detection method, and a program.

  As a wireless communication system, a technology using a BLE (Bluetooth Low Energy) device that transmits radio waves based on the Bluetooth (registered trademark) communication method and a device using a Wi-Fi (registered trademark) communication method is known. There is. In such a communication system, the receiving means detects an object by receiving the radio wave with respect to a detection target (for example, a person or the like) who possesses an apparatus (hereinafter, also referred to as a "tag") that transmits the radio wave. There is known a technique for estimating the area from the radio wave intensity etc. In addition, the technology for estimating the amount of activity, action content, posture, etc. of the object to be detected based on the values (for example, values of acceleration, angular velocity, barometric pressure, geomagnetism, etc.) of various sensors provided in the tag of the object to be detected. Proposed.

  However, in analysis using wireless communication, there is a problem that an error of 1 to 3 m may occur to the actual position in principle. In addition, there is also a problem that the information to be obtained is limited depending on conditions such as the attachment site of the tag to be detected.

  On the other hand, conventionally, an analysis technique using an image of a surveillance camera or the like installed at a fixed point has been known as such a detection method of the detection target and an estimation method of the position and posture.

  However, in the analysis method using an image captured by the imaging unit, the analysis can not be performed when the object to be detected is outside the angle of view of the imaging unit or when a blind spot occurs due to the presence of an obstacle or the like. There is a problem called.

In order to solve such problems, a technique has been proposed for estimating the action of a target in an unphotographed area (see, for example, Patent Document 1).
Patent Document 1 discloses a configuration that can recognize an action of a target person even in a blind spot by using an action estimation model.

  In addition, when the detection target is a person, misidentification of an individual may occur, for example, when identification from an image is difficult. On the other hand, a technology using identification information for identifying a monitoring target has also been proposed (see, for example, Patent Document 2).

  In a system for detecting an object to be detected, the detection of the object to be detected or the position estimation that is outside the imaging area or the blind spot of the imaging means must be reliably complemented by other detection means, and erroneous recognition of the object to be detected In order to prevent this, it is required to reliably associate the information for specifying the detection target with the position information.

In order to reliably complement the detection of the detection target and the estimation of the position, and to reliably associate the information for specifying the detection target with the position information, it is preferable to perform image processing using an image analysis unit.
However, when performing processing to recognize a person to be detected and its action using image analysis means, if it is difficult to specify the object to be detected, the load on the system including processing time and processing power is There is a problem of becoming larger.

  Therefore, in view of the above problems, it is an object of the present invention to provide an object detection system capable of performing an optimal detection process while reducing the load on the system even when it is difficult to specify an object to be detected. .

  In order to achieve the above object, an object detection system according to the present invention comprises: identification information transmitting means mounted on an object to be detected and transmitting identification information of the object to be detected; and a signal transmitted from the identification information transmitting means Image detecting means for receiving and detecting the object to be detected, imaging means, image processing performed on the image taken by the imaging means, and estimating at least the position of the object to be detected; Providing a control unit that controls the image analysis unit to execute different processing depending on whether the detection target is detected by the detection unit or not detected in the imaging area of the imaging unit; It features.

  According to the present invention, it is possible to provide an object detection system capable of performing an optimum detection process while reducing the load on the system even when it is difficult to specify an object to be detected.

It is a figure which shows an example of schematic structure of a target detection system. 1 is a block diagram showing an example of a schematic configuration of a target detection system. It is a block diagram which shows an example of a structure of an identification information transmission means (tag). It is a block diagram which shows an example of a structure of the detection means which is a receiver. It is a block diagram showing an example of composition of a server. FIG. 6 (a) is a correspondence table, and FIG. 6 (b) is a diagram schematically showing a region. It is a flowchart which shows an example of the process performed in a server. It is explanatory drawing which shows an example of the aspect which estimates the position of to-be-detected object. It is explanatory drawing which shows an example of the aspect which estimates the position of to-be-detected object. It is explanatory drawing which shows an example of the aspect which estimates the position of to-be-detected object. It is explanatory drawing which shows an example of the aspect which estimates the position of to-be-detected object. It is explanatory drawing which shows an example of the aspect which estimates the attitude | position of a to-be-detected object. It is explanatory drawing which shows an example of the aspect which estimates the position of to-be-detected object.

  Hereinafter, an object detection system, an object detection method, and a program according to the present invention will be described with reference to the drawings. In addition, this invention is not limited to embodiment shown below, It can change in the range which those skilled in the art could concidentize, such as other embodiment, addition, correction, deletion, and any aspect Even within the scope of the present invention, as long as the functions and effects of the present invention can be achieved.

FIG. 1 is a diagram showing an example of a schematic configuration of a target detection system.
As shown in FIG. 1, the object detection system 1 is attached to the object to be detected 10, and transmits identification information transmitting means (hereinafter also referred to as "tag") 20 for transmitting identification information of the object to be detected 10; It comprises a detection means 40 which is a receiving device for receiving the signal of the identification information transmitted from the means 20 and detecting the detection target 10, a photographing means (camera) 50, and a server 30.
The server 30 is communicably connected to the detection unit 40 and the imaging unit 50 via a network.
It is preferable that a plurality of detection means 40 be provided, and the detection means 40 be disposed along with the imaging means 50 at a desired place.

FIG. 2 is a block diagram showing an example of the configuration of a target detection system.
The server 30 analyzes an identification information analysis unit 31 that analyzes various data included in identification information received from the detection unit 40, and an image analysis unit 32 that is an image analysis unit that analyzes image data received from the imaging unit 50. An integration unit 33 that associates and integrates information based on results, a database (DB) 34 in which the integrated information is registered and stored, and a UI display unit that extracts and displays data from the database 34 in response to a request from a user 35 is provided.

FIG. 3 is a block diagram showing an example of the configuration of identification information transmitting means (tag) 20. As shown in FIG.
An acceleration sensor 22, an angular velocity sensor 23, an air pressure sensor 24, and a geomagnetic sensor 25 are mounted on the tag 20. The output value from each sensor is transmitted from the wireless transmission unit 21 using data such as BLE (Bluetooth Low Energy) as data of each sensor.

FIG. 4 is a block diagram showing an example of the configuration of the detection means 40 which is a receiving device.
The detection unit 40 receives an identification information (may include data of each sensor) transmitted from the tag 20, an information reception unit 41, an information processing unit 42 that processes the received identification information and data of each sensor, The information transmission unit 43 transmits the processed data to the server 30.

FIG. 5 is a block diagram showing an example of the configuration of the server 30. As shown in FIG.
The server 30 receives from the receiving unit 36 that receives data from the detection unit 40, the identification information analysis unit 31 that analyzes data received from the detection unit 40, the reception unit 37 that receives image data from the imaging unit 50, and the imaging unit 50 Image analysis unit 32 that analyzes the processed image data, an integration unit 33 that associates and integrates information based on each analysis result, a database (DB) 34 in which the integrated information is registered and stored, and a user And a UI display unit 35 for extracting and displaying data from the database 34 in response to a request from the user.

The information integrated by the integration unit 33 includes, for example, information of the detection target 10 estimated by analysis of the image analysis unit 32, and information of the detection target 10 transmitted from the tag 20 and analyzed by the identification information analysis unit 31. It is information.
That is, in the integration unit 33, position information and the like of the detection target 10 estimated by the image analysis unit 32, which is an image analysis unit, and identification information (tag ID and position information and the like) of the detection target 10 are associated. , Is registered in the database 34 as integrated information.

The server 30 is provided with a control means, and the control means performs processing in which the image analysis unit 32 differs depending on whether the detection target 10 is detected by the detection means 40 within the imaging area of the imaging means 50 or not. Control to run.
When another detection unit 40 outside the imaging area of the imaging unit 50 detects the detection target 10, the integration unit 33 changes the state, and the other detection unit 40 is included in the imaging area. A process of analyzing the image of the photographing means 50 is performed.

In the processing for estimating the position of the detection target 10, the processing by the image analysis unit 32 for the image data is overwhelmingly higher than the processing by the identification information analysis unit 31 for the data received by the detection unit 40 Requires ability. That is, analysis that does not involve image processing is light in calculation processing because the amount of data to be processed and the amount of calculation is small, but analysis that involves image processing has a large amount of data to be processed and the amount of calculation, and calculation processing is heavy Become.
Therefore, instead of always performing the image processing by the image analysis unit 32, control is performed so that the image processing by the image analysis unit 32 is executed only when the detection target 10 is detected within the imaging area of the imaging unit 50. Is preferable because the processing speed can be increased and the calculation cost can be reduced, and the preparation of computer resources can be reduced.
When the detection target 10 is detected by the detection unit 40 in the imaging area of the imaging unit 50, the image analysis unit 32 executes the image processing of the detection target 10, and the detection target 10 is in the imaging area of the imaging unit 50. In the case where the detection unit 40 does not detect the image, the image analysis unit 32 is controlled not to execute the image processing of the detection target 10.
When the detection target 10 is detected outside the imaging region of the imaging unit 50, the detection unit 40 estimates the position.

In the target detection method according to the target detection system of the present embodiment having the above-described configuration, the identification information transmission unit (tag) 20 mounted on the detection target 10 transmits identification information of the detection target 10, and detection The means 40 receives the signal transmitted from the identification information transmitting means 20, detects the object 10 to be detected, the step of imaging the object 10 to be detected by the imaging means 50, and the object 10 to be detected is an imaging device The image analysis unit 32, which is an image analysis unit, executes different processing depending on whether the detection unit 40 detects the detection region 40 or not within the 50 imaging regions, and the detection target 10 is in the imaging region of the imaging unit 50. If the image analysis unit 32 performs image processing of the detection target 10 when detection is performed by the detection unit 40 in Includes a flop, the.
That is, only when the object to be detected 10 is detected within the imaging area of the imaging means 50, the step of performing the image processing by the image analysis unit 32 to estimate the position of the object to be detected 10 is performed. When 10 is detected outside the imaging area of the imaging unit 50, the image processing step is not performed, and the position estimation step by the detection unit 40 is performed.

In the target detection system of the present embodiment, when the detection target 10 detected in the imaging area of the imaging unit 50 subsequently moves to a position within the imaging area and not detected by the detection unit 40, The estimation of the position by the image analysis means (image analysis unit 32) is continued.
This process will be specifically described in step S012 in the flowchart of FIG. 7 described later and in FIG.

In addition, when the detection target 10 detected in the imaging area of the imaging unit 50 subsequently moves to the outside of the imaging area or the non-imaging position, the position is estimated based on the information received by the detection unit 40.
The processing in the case where the detection target 10 is detected outside the imaging area of the imaging means 50 will be specifically described in step S 004 in the flowchart of FIG. 7 described later and FIG.

The estimation of the position of the detection target 10 based on the information received by the detection unit 40 is performed based on the position information of the detection unit 40 and the signal strength received from the identification information transmission unit 20.
The estimation of the position of the detection target 10 based on the position information of the detection means 40 will be specifically described in step S002 of the flowchart of FIG. 7 described later and FIG.
From the signal intensity from the identification information transmitting means 20 together with the position information of the detecting means 40, for example, the distance to the detecting means 40 can be determined.

The detection target 10 is, for example, a person.
The target detection system of the present embodiment can be configured such that the posture is estimated together with the estimation of the position of the detection target 10 from the information received by the detection unit 40 and / or the image captured by the imaging unit 50.

  In addition to the position information of the detection target 10 and the identification information, by associating the information of the posture, it is possible to use, for example, analysis of the action in picking of the worker of the physical distribution warehouse. The obtained data can be utilized for layout improvement etc. in the warehouse.

Further, similarly to the position estimation processing, the image analysis unit 32 performs image processing only when the detection target 10 is detected within the imaging area of the imaging unit 50, and estimates the estimation processing of the posture of the detection target 10 as well. The control to be performed contributes to the reduction of preparation of computer resources due to speeding up of processing and reduction of calculation cost.
Further, when the object to be detected 10 can not be recognized by the imaging unit 50 or when the recognition is difficult, it is possible to estimate the posture based on the information of the detection unit 40.

An example of a posture estimation method based on the information of the detection means 40 will be described below.
First, the inclination of the tag 20 is calculated based on the acceleration detected by the acceleration sensor 22 included in the tag 20 and the magnetic north detected by the geomagnetic sensor 25.
For example, when the person 10 to be detected bends, the inclination of the tag 20 takes a value in a predetermined range, and therefore, when the value is detected, it is estimated that the person 10 bends. In addition, when the person 10 takes a warped posture, it takes a value in a predetermined range in which the inclination in the direction opposite to the waist bending has a certain value. Therefore, when the value is detected, it is estimated that the person 10 has a warped posture.

  On the other hand, when the person 10 to be detected is crouched, the air pressure detected by the air pressure sensor 24 changes. The lower the air pressure is, the lower the air pressure. Therefore, when the air pressure is subtracted from the air pressure after the air pressure, the air pressure difference when squatting becomes a positive value. Conversely, when the person 10 stands up, the air pressure difference becomes a negative value. Therefore, a posture of squatting is estimated by detecting a pair of a barometric pressure peak (positive value) at squashing and a peak (negative value) at a rising action from squatting. Although squatting is recognized after the end of rising, it is estimated that the posture from squatting up to rising is the posture of squatting afterward.

This other attitude is estimated as follows.
Walking: The posture of walking is estimated from the acceleration detected by the acceleration sensor 22 and the angular velocity detected by the angular velocity sensor 23. For example, when acceleration in the vertical direction above a predetermined level is detected, and lateral swing (roll) that occurs in human walking is detected, it is estimated to be walking.
Stair climbing: When the person 10 climbs the stairs, the same acceleration and angular velocity as walking are detected. In the case of stair-climbing, the air pressure also changes, and vertical acceleration that does not occur in walking occurs. Therefore, when the same lateral swing as walking, acceleration in the vertical direction stronger than walking, and a change in atmospheric pressure are detected, the posture of going up and down the stairs is estimated. It should be noted that the stairs are climbed if the pressure is smaller, and the stairs are down if the pressure is larger.
Sitting: When the person 10 is seated, the acceleration is detected in the downward direction and then stabilized, and after seating, the tilt of the tag 20 is stabilized at a value different from the reference value. Therefore, after the acceleration is detected in the downward direction, it is estimated that seating is stable if the inclination is stabilized at a value different from the reference value.

FIG. 6 shows an example of the correspondence between the imaging area of the imaging means 50 and the detection range of the detection means 40. FIG. 6 (a) is a correspondence table, and FIG. 6 (b) is a state when each area is viewed from above Are schematically shown.
Although FIG. 6B shows an example in which two cameras are installed as the imaging unit 50, the number of installed units and the installation position are not limited to this.
The imaging area 60a of the imaging means 50a is represented as Cam1, and the imaging area 60b of the imaging means 50b as Cam2.

  As shown in FIGS. 6A and 6B, in the imaging area 60a of the camera 50a (the imaging available area represented by Cam1), the detection target 10 is disposed so as to be detectable in the imaging area. Two detection means A (40a) and detection means B (40b) are installed. A signal receivable area by the detecting means A (40a) is indicated by Ra, and a signal receivable area by the detecting means B (40b) is indicated by Rb. Further, outside the imaging area 60a (Cam 1), a detection means C (40c) is installed. A signal receivable area by the detection means C (40c) is indicated by Rc.

These correspondences are shown in the correspondence table of FIG. 6 (a).
The correspondence table is referred to in the integration processing of information in the server 30 in the target detection system of the present embodiment.
Further, in the present embodiment, the image analysis unit 32, which is an image analysis unit, is controlled to execute the image processing only when the detection target 10 is detected by the detection unit 40 in the imaging region of the imaging unit 50. Therefore, the image analysis unit 32 is controlled to execute the image processing only when the detection unit 40 arranged to detect the detection target 10 in the imaging area of the imaging unit 50 detects the detection target 10. Referenced to.

FIG. 7 is a flowchart showing an example of processing performed by the server 30. These processes are controlled by the control means.
First, the information received by the reception unit 36 from the detection unit 40 is output to the identification information analysis unit 31, and the identification information analysis unit 31 analyzes the tag information.
As a step of information input to the integration unit 33, an analysis result of tag information is acquired from the identification information analysis unit 31 (S001).
Information to be acquired includes information that data could not be acquired (impossible to acquire).

Next, based on the same correspondence table as that shown in FIG. 6A, the correspondence between the reception area of the detection means 40 and the imaging area of the imaging means 50 is referred to (S002). Specifically, the correspondence relationship and the coordinates of the reception area of the detection means 40 (area information estimated from the information of the tag 20) and the position information estimated from the photographed image are compared.
The position information can be estimated, for example, from an image of a frame (previous frame) at a previous time, or can be estimated from a camera installation position and a camera setting (such as a lens focal length).

The reception area of the detection unit 40 and the imaging area of the imaging unit 50 are compared based on the correspondence table (S002), and it is determined whether the reception area of the detection unit 40 is within the imaging area of the imaging unit 50 S003).
If the reception area of the detection means 40 is not within the imaging area and the tag 20 is present within the reception area of the detection means 40, it is determined that the tag 20 is outside the imaging area of the imaging means 50. Ru. In this case, as the position information, the reception area of the detection means 40 which is outside the imaging area and the tag information (hereinafter, also referred to as “tag ID”) are integrated (S 004).

If there is a reception area of the detection unit 40 within the imaging area of the imaging unit 50, the reception unit 37 receives data to analyze the photographed image, and outputs the received image data to the image analysis unit 32. And execute actual analysis processing (S005).
It should be noted that the analysis processing of the image is performed only when the object to be detected 10 is detected in the imaging area of the imaging unit 50 and the process proceeds to the ID integration processing because the load is large.
The analysis result of the photographed image is acquired from the image analysis unit 32 (S006), and it is determined whether the estimated position (coordinates) based on the image can be acquired (S007).

When the estimated position (coordinates) based on the image can be acquired, it is determined whether the estimated position (coordinates) based on the photographed image exists within the reception area of the detection means 40 estimated from the tag information. To do (S008).
If the estimated position (coordinates) based on the photographed image exists in the reception area of the detection means 40 estimated from the tag information, whether or not the estimated position (coordinates) is the same as that previously registered in the database It is determined (S009).
If it is different from the previously registered estimated position (coordinates), the estimated position (coordinates) based on the image and the tag ID are integrated (S010). On the other hand, if it is the same as the previously registered estimated position (coordinates), the previous estimated position and the tag ID are integrated (S011).

On the other hand, when the estimated position (coordinates) based on the photographed image does not exist in the reception area of the detection means 40 estimated from the tag information, it is judged whether or not there is past position information of the same tag ID (S012) .
If there is past position information of the same tag ID, the previous estimated position and the tag ID are integrated (S011). This may be the case, for example, when the detection target 10 detected in the imaging area of the imaging unit 50 is moved to a position within the imaging area and not detected by the detection unit 40. In this embodiment, even in such a case, it means that the estimation of the position by the image analysis means is continued.

On the other hand, when the past position information of the same tag ID can not be confirmed, the position information (coordinates) can not be associated with the tag ID. Further, even when the tag 20 is not recognized by the detection unit 40 having the reception area at the estimated position (coordinates) obtained from the image, the position information (coordinates) can not be associated with the tag ID. In the latter case, the possibility that the detection target 10 does not possess the tag 20 is also assumed.
In these cases, assuming that the detection target 10 is an unknown person, the estimated position (coordinates) and the unknown person information are integrated (S013).

  In step S 007, when it is determined that the estimated position (coordinates) based on the image is not obtainable, that is, although the reception area of the detection unit 40 is in the imaging area of the imaging unit 50, the estimated position (image As a case where the coordinates can not be acquired, for example, it is assumed that the detection target 10 carrying the tag 20 exists in the blind spot of the photographing means 50. In this case, the estimated position (area) based on the tag information is integrated with the tag ID.

It is determined whether the estimated position (area) based on the tag information is the same as the one previously registered in the database (S014).
If the estimated position (area) is different from the previously registered estimated position (area), the estimated position (area) based on the tag information and the tag ID are integrated (S015). On the other hand, if it is the same as the previously registered estimated position (area), the previous estimated position and the tag ID are integrated (S016).
In step S014, the estimated position (area) based on the tag information can be determined based on the position information of the detection unit 40 and the signal strength received from the identification information transmission unit 20.
For example, if the signal strength from the identification information transmitting means 20 is high, it can be determined that the position is close to the detection means 40 and the signal strength is low if the signal strength is low. It is also possible to compare the signal strength with a predetermined threshold value and make a decision to proceed to either step S016 or step S015 according to the result.

The information integrated in the above-mentioned steps S010, S011, S013, S015 and S016 is registered in the database 34 (S017).
It is determined whether the determination on all the acquired detection results is completed (S018), and if not completed, the process returns to the comparison based on the correspondence table in step S002 again to continue the process.

Then, when the judgment on the acquired detection result is completed, it is judged whether or not the detection is ended (S019), and if not ended, the analysis result of the tag information is acquired again (S001), and the next frame or The processing of detection and judgment of the detection target after the elapse of a predetermined time is repeated.
In addition, the above-mentioned processing is basically performed for a long time, and the detection is limited to the time of sleep of the function on the system and the execution of maintenance / inspection.

Hereinafter, the aspect which estimates the position of the to-be-detected object 10 is demonstrated based on FIGS. 8-13. 8 to 13 schematically show the respective regions viewed from the top, as in FIG. 6 (b).
Note that the schematic diagrams of FIGS. 8 to 13 schematically show the positional relationship between each area area and the person, and do not represent the size relationship. Also, the number of installed detection means 40 is not limited to the illustrated example.
Each process is performed by the server 30 as shown in FIG. 2, and the integration unit 33 of the server 30 stores a correspondence table as shown in FIG. 6 (a).

(Example 1)
In FIG. 8, a photographing unit 50, a photographing area 60 of the photographing unit 50, three detection units 40 (40a, 40b, 40c), a reception area R (Ra, Rb, Rc) of each detection unit 40, and a tag 20 are shown. One detected subject 10 (person) is shown.

In the present embodiment, the correspondence table indicates the correspondence between the imaging area 60 and the reception areas Ra, Rb, and Rc of the detection units 40.
In the present embodiment, the tag 20 is detected by the detection means 40a. In addition, a person 10 is detected in the imaging area 60.

According to the correspondence table, it is specified that the person 10 detected from the image photographed by the photographing means 50 is the person possessing the tag 20.
Based on the result, the tag information and the position information of the person 10 are integrated and registered in the database 34.

(Example 2)
9, an imaging unit 50, an imaging area 60 of the imaging unit 50, three detection units 40 (40a, 40b, 40c), a reception area R (Ra, Rb, Rc) of each detection unit 40, and a tag 20 are shown. One detected subject 10 (person) is shown.

  In the present embodiment, the tag 20 is detected by the detection means 40c. On the other hand, in the imaging area 60, the person 10 is not detected.

Since the person 10 is not detected from the image photographed by the photographing means 50 according to the correspondence table, the person 10 carrying the tag 20 is within the reception area Rc of the detection means 40c outside the photographing area 60. Is identified.
Based on the result, the tag information and the position information of the person 10 are integrated and registered in the database 34.

(Example 3)
10, an imaging unit 50, an imaging area 60 of the imaging unit 50, three detection units 40 (40a, 40b, 40c), a reception area R (Ra, Rb, Rc) of each detection unit 40, and a tag 20a Two persons, the to-be-detected object 10a (person) and the to-be-detected object 10b (person) having the tag 20b, are shown.

  In the present embodiment, the tag 20a is detected by the detection unit 40a, and the tag 20b is detected by the detection unit 40b. In the imaging area 60, two persons (10a, 10b) are detected.

Since the person 10a at a position closer to the photographing means 50 is detected by the sensor 40a according to the correspondence table, the person 10a is identified as the person carrying the tag 20a. Similarly, since the person 10b at a position far from the photographing means 50 is detected by the sensor 40b, the person 10b is identified as the person who holds the tag 20b.
Based on the result, the tag information and the position information of the person 10a and the person 10b are integrated and registered in the database 34.

(Example 4)
11, an imaging unit 50, an imaging area 60 of the imaging unit 50, three detection units 40 (40a, 40b, 40c), a reception area R (Ra, Rb, Rc) of each detection unit 40, and a tag 20a Two persons, the to-be-detected object 10a (person) and the to-be-detected object 10b (person) not having the tag, are shown.

In the present embodiment, the tag 20a is detected by the detection means 40a. Although the person 10b is in the reception area Rb of the detection means 40b, it is not detected by the detection means 40b because it does not have a tag.
On the other hand, the photographing unit 50 detects two persons 10a and 10b.

Since the person 10a at a position closer to the photographing means 50 is detected by the sensor 40a according to the correspondence table, the person 10a is identified as the person carrying the tag 20a. On the other hand, the person 10b at a position far from the photographing means 50 is not detected by the sensor 40b, and thus is regarded as an unknown person.
Based on the result, the tag information and the position information of the person 10a are integrated and registered in the database 34. On the other hand, although the person 10b determined as an unknown person is not registered in the database, the estimated position may be registered in combination with unknown person information.

(Example 5)
In FIG. 13, an imaging unit 50, an imaging area 60 of the imaging unit 50, three detection units 40 (40a, 40b, 40c), a reception area R (Ra, Rb, Rc) of each detection unit 40, and a tag 20 are shown. One detected subject 10 (person) is shown.
In FIG. 13, for example, in the mode shown in FIG. 8, the person 10 registered with the tag information and the position information integrated in the database 34 is then detected by the detecting means 40 (40a, 40b) within the photographing area 60. It shows that it has moved to a position where it can not be

In the present embodiment, the person 10 is detected in the imaging area 60, but the tag 20 is not detected by the detection means 40a or 40b.
In this case, the presence or absence of the past position information of the same tag is determined, and when the position information exists, the position information based on the previous image and the tag information are integrated and registered.

The object detection system according to the present embodiment performs image analysis when the person 10 detected in the imaging area 60 subsequently moves to a position in the imaging area 60 that is not detected by the detection unit 40 (40a, 40b). The estimation of the position by the image analysis unit 32, which is a means, is continued.
That is, even when the person 10 goes out of the area that can be received by the receiving device, tagging can be continued as long as detection is continued by the imaging unit 50.

(Example 6)
The target detection system of the present embodiment can estimate the posture as well as the position of the detection target 10 from the information received by the detection unit 40 and / or the image captured by the imaging unit 50.

In the estimation of the posture, for example, a walking state or a crouched posture can be detected from the tag 20 mounted at the position of the waist of a person, but the movement of a hand without the tag mounted can not be detected.
On the other hand, when estimation from an image is performed by the photographing unit 50, it is possible to recognize various postures and actions by learning.

FIG. 12 has the same configuration as that of FIG. 8, and the imaging means 50, the imaging area 60 of the imaging means 50, the three detection means 40 (40a, 40b, 40c) and the reception area R (Ra, each detection means 40). One person to be detected 10 (person) having Rb, Rc) and a tag 20 is shown.
As in FIG. 8, since the person 10 can be specified by both the image and the tag information, it is considered that the posture and the motion can be estimated with high accuracy.

  Based on the result, tag information, position information and posture information of the person 10 are integrated and registered in the database 34.

  In the example of the configuration shown in FIG. 9, since the photographing unit 50 can not estimate the posture from the image, the posture is estimated based on the information of the detection unit 40. Based on the obtained result, the tag information, the position information and the posture information are integrated and registered in the database 34.

As described above, according to the target detection system of the present embodiment, the information obtained by the tag 20 and the information obtained from the image photographed by the photographing unit 50 are complementarily integrated and processed, and the position of the detection target 10 It is possible to estimate information and the like with high accuracy without loss.
Then, even when it is difficult to specify the detection target, it is possible to perform the optimum detection processing while reducing the load on the system.

<Program>
The programs executed in the target detection system of the present embodiment are as follows.
Identification information transmitting means 20 mounted on the detection subject 10 and transmitting identification information of the detection subject 10, and detection means 40 for receiving the signal transmitted from the identification information transmission means 20 and detecting the detection subject 10 An imaging unit 50, an image analysis unit 32 that is an image analysis unit that performs image processing on the image captured by the imaging unit 50, and estimates at least the position of the detection target 10; In the target detection system, the image analysis unit 32 executes different processing depending on whether the detection target 10 is detected by the detection unit 40 in the imaging area of the imaging unit 50 or not.
That is, according to the program of the present embodiment, when the detection target 10 is detected by the detection unit 40 in the imaging area of the imaging unit 50, the image analysis unit 32 performs the image processing of the detection target 10 and the detection target The position of 10 is estimated, and when it is not detected, the image analysis unit 32 does not perform image processing of the detection target 10, and the position estimation by the detection unit 40 is performed.

  The program executed in the target detection system of this embodiment is a file of an installable format or an executable format, and is a CD-ROM, a flexible disk (FD), a CD-R, a DVD (Digital Versatile Disk), a USB (Universal) It may be configured to be recorded and provided in a computer readable recording medium such as Serial Bus), or may be configured to be provided or distributed via a network such as the Internet. In addition, various programs may be configured to be provided by being incorporated in a ROM or the like in advance.

10 Detection target (person)
20 Identification Information Transmission Means (Tag)
30 server 31 identification information analysis unit 32 image analysis means (image analysis unit)
33 integrated part 34 database 35 UI display part 40 detection means 50 imaging means 60 imaging area

JP, 2011-90408, A Patent No. 5370927 gazette

Claims (9)

Identification information transmitting means which is attached to the detection target and transmits identification information of the detection target;
A detection unit that receives the signal transmitted from the identification information transmission unit and detects the detection target;
Shooting means,
An image analysis unit that performs image processing on the image captured by the imaging unit and estimates at least the position of the detection target;
Control means for controlling the image analysis means to execute different processes depending on whether the detection object is detected by the detection means or not in the imaging area of the imaging means; An object detection system characterized by   When the detection target is detected by the detection unit in the imaging area of the imaging unit, the image analysis unit executes image processing of the detection target, and when not detected, the image analysis unit detects the detection target. The object detection system according to claim 1, wherein the image processing of the object is not performed.   The object detection system according to claim 1 or 2, wherein the information of the position of the detection object estimated by the image analysis means is associated with identification information of the detection object.   When the object to be detected detected in the imaging area of the imaging means moves to a position within the imaging area that is not detected by the detection means, estimation of the position by the image analysis means is continued. The object detection system according to any one of claims 1 to 3, characterized in that   When the object to be detected detected in the imaging area of the imaging means has moved to the outside of the imaging area or the non-imaging position, position estimation is performed based on the information received by the detection means. The object detection system according to any one of claims 1 to 4.   The estimation of the position of the detection target based on the information received by the detection means is performed based on the position of the detection means and the signal strength received from the identification information transmission means. The object detection system according to any one of the above. The detection target is a person,
The posture is estimated together with the estimation of the position of the object to be detected from the information received by the detection means and / or the image photographed by the photographing means. Object detection system. The identification information transmitting means mounted on the detection target transmitting the identification information of the detection target;
The detection means receives the signal transmitted from the identification information transmission means, and detects the object to be detected, and the object is detected by the photographing device.
The image analysis unit executes different processing depending on whether the detection target is detected by the detection unit or not in the imaging area of the imaging unit.
The image analysis unit executes image processing of the detection target when the detection target is detected by the detection unit in a shooting area of the shooting unit;
Estimating the position of the object to be detected. Identification information transmitting means which is attached to the detection target and transmits identification information of the detection target;
A detection unit that receives the signal transmitted from the identification information transmission unit and detects the detection target;
Shooting means,
An image analysis unit that performs image processing on the image captured by the imaging unit and estimates at least the position of the detection target;
An object detection system including control means;
A program characterized in that the image analysis unit executes different processing depending on whether the detection target is detected by the detection unit or not in the imaging area of the imaging unit.