JP2016025522A - Image processing system and image processing system control method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To perform object detection in consideration of the difference of incident light within a photographing area by obtaining a spectral reflectance characteristic to a target.SOLUTION: In the case of obtaining the spectral distribution of a solid body existing within image data captured by plural monitor cameras, this system previously holds the spectral distribution of a background matter solid body that becomes a subject, obtains the spectral distribution of incident light from a background matter near a target, and obtains a spectral reflectance characteristic to the target by using the spectral distribution of the incident light, thereby making it possible to perform the object detection in consideration of the difference of incident light within a photographing area.SELECTED DRAWING: Figure 6

Description

  The present invention relates to an image processing apparatus and a method for controlling the image processing apparatus, and more particularly to a technique for tracking an object present in images taken by a plurality of surveillance cameras and specifying the object by color.

  2. Description of the Related Art Conventionally, there has been a technology that uses a person or a vehicle as a monitoring target and performs verification of the target among a plurality of cameras. For example, there are technologies that analyze face recognition and walking conditions when a person is a target, and technologies that analyze the shape and vehicle number of a vehicle when a vehicle is an object. With these techniques, an image suitable for recognition is not always obtained depending on the orientation of the photographed object and the photographing position.

  In the conventional technique (Non-Patent Document 1) for extracting color feature amounts, an RGB three-dimensional color histogram is used as the color feature amount. Persons are associated based on the difference in similarity between the obtained histograms. In such a color space of RGB values, when the influence of the light source is removed, there is a possibility that the object becomes a different object. In particular, in a surveillance camera installed in a wide area including the outdoors, it is rare that the incident light to the object is the same. Therefore, the determination accuracy may be reduced.

  As described above, there is a technique that uses a color feature amount obtained by removing an illumination light source by image processing using spectroscopy instead of a color space in which appearance changes depending on a light source (Patent Document 1). In Patent Document 1, a light source characteristic that causes a color change according to an image and a spectral reflectance of the subject are estimated, thereby estimating a subject-specific color from which the influence of light source fluctuation is removed. As a result, the objects are compared using colors that do not depend on the light source environment.

JP 2009-231921 A

Takuya Honda and four others, “Study on person matching using color features for room entry / exit analysis” Proceedings of the IEICE General Conference 2009_Information & Systems (2), 225, 2009-03-04

The color of an object varies with light. However, the light does not always shine uniformly within the imaging area of the surveillance camera. In particular, in a shooting area installed outdoors or the like, a shadowed part such as a building or a tree becomes dark. In the above-mentioned Patent Document 1, the spectral reflectance of an object is not calculated in consideration of the difference in incident light within the imaging region, and the color feature amount is calculated using different incident light, so that the accuracy of object detection decreases. There was a problem to do.
In view of the above-described problems, an object of the present invention is to obtain a spectral reflectance specific to an object and perform object detection in consideration of a difference in incident light in an imaging region.

  An image processing apparatus according to an embodiment of the present invention includes a monitoring control device and a plurality of monitoring devices, and an image region captured in an image processing device that analyzes image data captured by the plurality of monitoring devices and identifies an object. A holding means for holding background object information having position information of an object existing inside and spectral distribution information of the object solid, and a background object close to the tracking object solid is extracted, and the background object information of the extracted background object is Extracting means for obtaining, calculating means for obtaining a spectral distribution of incident light on the background object from the extracted background object information and spectral reflectance of the captured image data, and a spectral distribution of the incident light on the background object and photographing Calculating means for obtaining the spectral reflectance of the tracking object from the obtained image data and obtaining the spectral distribution of the tracking object solid, and the tracking object in different photographed image data as in the tracking object An object to be a tracking target candidate based on the calculating means for obtaining the spectral distribution of the incident light on the candidate object, and the spectral reflectance of the object to be tracked and the calculated spectral distribution of the incident light. The calculation means for obtaining the spectral distribution of the solid of the object, the spectral distribution of the obtained tracking object solid is compared with the solid spectral distribution of the object as the candidate of the obtained tracking object, and the tracking object is determined based on the similarity. And determining means for determining information of an object that is a candidate for an object.

  According to the present invention, a background object having the same incident light as that of the tracking object is acquired, the spectral distribution of the incident light of the tracking object is estimated from the spectral distribution of the solid of the background object held in advance, and the tracking object solid is obtained. By acquiring the spectral distribution, it is possible to improve the accuracy of object detection.

It is a block diagram which shows an example of a structure of the image processing apparatus which concerns on this invention. It is a block diagram which shows an example of a structure of the monitoring apparatus which concerns on the 1st Embodiment of this invention. It is a block diagram which shows an example of a structure of the computer system used with the monitoring control apparatus and target object detection apparatus which are one component of the image processing apparatus which concerns on the 1st Embodiment of this invention. It is a block diagram which shows the structural example of the imaging device which is one component of the monitoring apparatus which concerns on the 1st Embodiment of this invention. It is a flowchart which shows an example of the process sequence of the monitoring control apparatus which concerns on the 1st Embodiment of this invention. It is a flowchart which shows an example of the acquisition process of the tracking target object information which concerns on the 1st Embodiment of this invention. It is a figure which shows an example of the tracking target object information table for storing the information which concerns on the tracking target object which concerns on the 1st Embodiment of this invention, and the candidate of a target object. It is a figure which shows an example of the tracking target object picked-up image which concerns on the 1st Embodiment of this invention. It is a flowchart which shows an example of the detection process of the tracking target object information which concerns on the 1st Embodiment of this invention. It is a flowchart which shows an example of the acquisition process of the background object information table which concerns on the 1st Embodiment of this invention. It is a figure which shows an example which divided | segmented the imaging | photography area | region which concerns on the 1st Embodiment of this invention. It is a figure which shows an example of the background object information table which concerns on the 1st Embodiment of this invention. It is a figure which shows an example of the comparison in the spectral reflectance which concerns on the 1st Embodiment of this invention. It is a figure which shows an example of the component of the image processing apparatus which concerns on the 2nd Embodiment of this invention.

Hereinafter, refer to the attached drawings. The present invention will be described in detail based on preferred embodiments thereof. The configurations shown in the following embodiments are merely examples, and the present invention is not limited to the illustrated configurations.
[First Embodiment]
FIG. 1 is a block diagram illustrating an example of an image processing apparatus according to an embodiment of the present invention. The image processing apparatus according to the present embodiment includes a monitoring control apparatus 101 and two or more monitoring apparatuses (102 to 107), analyzes image data captured by the plurality of monitoring apparatuses, and specifies an object. Do. These monitoring control device 101 and monitoring devices (102 to 107) are connected by a network.

  When the tracking target object is detected by any of the monitoring apparatuses (102 to 107), the monitoring control apparatus 101 transfers information on the tracking target object to another monitoring apparatus via the network. Then, it is detected whether or not an object having a color feature amount similar to the tracking object is photographed.

  Further, when the detection result is sent from each of the monitoring devices (102 to 107), the monitoring control device 101 acquires information on the candidate object, and information on the detected monitoring device and the candidate object. And the image data are displayed on the output device 308 described later. The details of the other monitoring control device 101 and each information will be described in “Monitoring control device” described later.

As shown in FIG. 2, the monitoring device 201 that is one of the monitoring devices (102 to 107) includes an imaging device 202 and an object detection device 203.
The imaging device 202 is configured using an imaging device capable of acquiring spectral characteristics, for example, an imaging device such as a multiband camera, and generates spectral image data.

The object detection device 203 detects a tracking object from the spectral image data generated by the imaging device 202.
A target similar to the tracking target is detected in the image data captured by the imaging device 202 from the tracking target information sent from the monitoring control apparatus 101. If there is a candidate target, the candidate target information is monitored and controlled. Transfer to device 101. Details of the other monitoring devices and each information will be described in “Imaging device in monitoring device” and “Object detection device in monitoring device”.

The configuration of each device will be described below.
<Monitoring control device>
The monitoring control apparatus 101, which is a component of the image processing apparatus in the present embodiment, is configured by a computer system as shown in FIG.
Specifically, a CPU 301, a ROM 302, a RAM 303, an external storage device 304, an input I / F 305, an input device 306, an output I / F 307, an output device 308, and a network I / F 309 are provided.

The CPU 301 performs overall control processing of the computer system.
A ROM (Read Only Memory) 302 is a read only memory. The ROM 302 includes a PROM (Programmable ROM) in which a user can electrically write a program and a mask ROM into which contents are written at the time of manufacture. In this embodiment, any ROM may be used.

  A RAM (Random Access Memory) 303 is a memory that can be freely written and read. The RAM 303 has a function of temporarily storing data when performing the processing of the present embodiment.

  The external storage device 304 is a nonvolatile storage device. Examples of the external storage device 304 include an HD drive, an FD drive, an MO drive, a CD-RW drive, a DVD-RW drive, and a Blu-ray (registered trademark) drive. The external storage device 304 can store program data and data necessary for the operation of the program data.

  The Input I / F 305 processes an input signal from the input device 306 from a keyboard, a mouse, or the like included in the computer system, and is used for operation of the computer system and data input.

The input device 306 includes a keyboard and a mouse for sending input signals to the computer system via the Input I / F 305.
The Output I / F 307 processes output signals from a monitor or printer connected to the computer system, and is used for operation of the computer system and data output.

  An output device 308 indicates an output device that receives and processes an output signal from the computer system via the Output I / F 307. The output device 308 is realized using a CRT, a liquid crystal display, a plasma display, a rear projection television or a front projector, and a control circuit thereof.

  The system bus 310 exchanges various data between blocks in the computer system and supplies power. The system bus 310 includes an address line, a data line, a control line, a power supply / ground line, and the like.

  The network I / F 309 processes communication control for connecting to a network such as an intranet, and various communication interfaces are applied depending on the user environment. Connected to a wireless LAN or Ethernet interface.

<Imaging device in monitoring device>
The imaging device 202 that is one component of the monitoring device 201 in this embodiment uses an imaging device that can acquire spectral characteristics. An example of the imaging apparatus will be described with reference to FIG.
In FIG. 4, reference numeral 401 denotes a light incident portion composed of a lens and a shutter, 402 denotes four or more types of color filters having different spectral transmittances, and 403 denotes an image sensor. Reference numeral 404 denotes a gain adjustment unit (AGC), 405 denotes an A / D conversion unit, 406 denotes a signal processing unit, 407 denotes an external storage device, 408 denotes a D / A conversion unit, and 409 denotes a display unit. Reference numeral 410 denotes a CPU, which is related to all processing of each configuration. Reference numeral 411 denotes a ROM, and 412 denotes a RAM, which stores programs, data, work areas, and the like necessary for the processing of the CPU 410. Reference numeral 413 denotes an operation unit for the user to instruct the operation of the camera.

An example of the processing of the imaging device 202 described above will be described.
Light from the imaged object reaches the image sensor 403 via the light incident portion 401 and the color filter 402, and is converted into an electric signal by photoelectric conversion. The electric signal converted by the image sensor 403 is subjected to gain adjustment by a gain value set in advance by the gain adjustment unit 404.

The gain-adjusted electrical signal is converted into a digital signal by the A / D converter 405. Thereafter, the converted digital signal is subjected to image processing such as demosaicing and noise reduction in the signal processing unit 406, and spectral image data is generated. The spectral image data generated by the signal processing unit 406 is stored in the storage device 407.
The spectral image data generated by the signal processing unit 406 is converted into RGB values, converted into a voltage by the D / A conversion unit 408, and displayed on the display unit 409.

<Object detection device in monitoring device>
Next, the object detection device 203 that is a component of the monitoring device 201 in the present embodiment will be described.
The object detection apparatus in the present embodiment is configured by a computer system as shown in FIG. Since the computer system shown in FIG. 3 has already been described in “Monitoring and Control Device”, a description thereof will be omitted.

Up to this point, the configuration of the image processing apparatus according to the present embodiment has been described. Hereinafter, the monitoring control device and the processing of the monitoring device in the present embodiment will be described.
<Processing of monitoring control device>
The process of the monitoring control apparatus 101 in this embodiment is demonstrated using FIG.
In any monitoring device connected to the network 108, when information related to an object to be tracked (hereinafter referred to as “tracking object information”) is transferred, the tracking object information and the transferred monitoring device are transmitted in S501. Is acquired and stored in the RAM 303.

  In S502, the acquired tracking object information is transferred to a monitoring apparatus other than the monitoring apparatus that transferred the tracking object information, and the tracking object detection process is executed. The tracking object detection process will be described in detail in the following “Tracking Object Detection Process in Monitoring Device”. When there is an object similar to the tracking object in each monitoring device, the image data or the like (hereinafter referred to as “tracking object candidate information”) is transferred as the tracking object candidate.

In step S <b> 503, the acquired tracking target candidate information and the transferred monitoring device identification information are stored in the RAM 303 via the network 108.
In step S504, the sent tracking object candidate information is displayed on the output device 308 together with the monitoring device identification information.
Details of the tracking object information and the tracking object candidate information will be described in “Trace Object Information Acquisition Process in Monitoring Device” and “Tracking Object Information Detection Process in Monitoring Device”.

<Acquisition processing of tracking object information in monitoring device>
FIG. 6 and FIG. 7 show the processing (tracking object acquisition processing) until the monitoring device connected to the network 108 (explained as an example of 102) transfers the tracking object information to the monitoring control device 101. This will be described with reference to FIG. FIG. 6 shows a flowchart of the tracking target acquisition process. FIG. 7 shows a tracking object information table for storing information related to the tracking object. FIG. 8 shows an example of a tracking object photographed image.

  In step S601, the background object information table related to the imaging apparatus 202 is acquired. This background object information table has information on a subject that is stationary in an image area photographed by the imaging apparatus 202, for example, position information of an object existing in the image area to be photographed and spectral distribution information of an object solid. It is the table which stored the information regarding background object information. The background object information table and the method of creating the table will be described in detail in “Background object information table acquisition process of monitoring device”.

  In step S <b> 602, image data (hereinafter, referred to as “tracked object captured image data”) in which a subject to be tracked is captured from an image captured by the imaging device 202 of the monitoring device 102 is acquired. The tracking target photographed image data is acquired by displaying an image generated by the imaging device 202 stored in the storage device 407 on the output device 308 and selecting image data by the input device 306 by the user.

  As the selected image data, an RGB image and spectral image data corresponding to the RGB image are acquired. Further, the spectral reflectance of the tracking object is acquired from the spectral image data. For example, FIG. 8A shows an example in which the tracking object photographed image data is displayed on the output device 308 such as a display of the monitoring device 102. The acquired tracked object captured image data is stored in the object captured image data 702 of the tracked object information table and held in the RAM 303.

  In step S <b> 603, the user uses the input device 306 to specify a tracking target existing in the target captured image data obtained in step S <b> 602. Hereinafter, the designated image data is referred to as “tracking object image data”. The designated tracking object image data is held in the RAM 303 as the object image data 703 in the tracking object information table. In the present embodiment, the tracking target object image data is stored as image data. However, the present invention is not limited to this, and a coordinate area in the target captured image data 702 can be held.

In addition, the method for specifying the tracking object by the user may be performed by automatically extracting an object existing in the image data and selecting the image data of each extracted object.
For example, FIG. 8B shows an example in which the tracking target is selected in the tracking target photographed image data shown in FIG. The portion selected by the bold line becomes the object image data 703 in the tracking object information table. Object image data 703 in the tracking object information table is stored in the RAM 303.

  In step S <b> 604, information about a subject (hereinafter referred to as “background object of the tracking target”) that is close to the tracking target is obtained. This process will be described in detail in “Background Object Information Table Acquisition Process of Monitoring Device”. The acquired background object information of the tracking object is stored in the RAM 303. That is, a background object close to the tracking object solid is extracted, and background object information is acquired. In addition, the subject information is acquired from a monitoring device that captures the tracking target. That is, the background object corresponding to the tracking object is specified from the background object information held in the monitoring device that captured the tracking object, and the spectral reflectance of the background object is acquired from the spectral image data. This spectral image data is the spectral image data acquired in S602.

In S605, from the background object information of the tracking object acquired in S604, the spectral distribution (Rx (λ)) of the solid of the background object is obtained from the background object information table acquired in S601. The spectral distribution of the solid of the background object is acquired from the background object information held in the monitoring device that captured the tracking object. The spectral distribution of the solid of the background object is spectral distribution information of the object solid of the background object included in the background information held in the monitoring device. For example, when the background object of the tracking target acquired in S604 is a wall, the spectral distribution of the wall is obtained in S605.
In S606, the spectral reflectance (Cx (λ)) of the background object acquired in S605 and the spectral reflectance (Cx (λ)) of the background object are acquired from the object captured image data 702 obtained in S602. The spectral distribution (Ex (λ)) of the incident light is calculated from Equation (1).
Ex (λ) = Cx (λ) / Rx (λ) Equation 1
The calculated spectral distribution (Ex (λ)) of the incident light is stored in the RAM 303 as the photographing incident light information 705 in the tracking object information table. As described above, the spectral distribution of the incident light on the background object is obtained from the background object information obtained in S605 and the spectral reflectance of the captured image data obtained in S602. Then, from the spectral distribution of the solid of the background object acquired in S605 and the spectral reflectance of the background object acquired in S602, the spectral distribution of the incident light on the tracking object is obtained.

In S607, the spectral reflectance (Cy (λ)) of the tracking target image is acquired from the spectral distribution (Ex (λ)) of the incident light acquired in S606 and the tracking target image data acquired in S603. The spectral distribution (Ry (λ)) of the solid is calculated by Equation 2.
Ry (λ) = Cy (λ) × Ex (λ) Equation 2
The obtained spectral distribution (Ry (λ)) of the tracking object solid is stored in the spectral reflectance 704 of the object solid in the tracking object information table. In this way, the spectral reflectance of the tracking object is obtained from the spectral distribution of the incident light on the background object and the captured image data, and the spectral distribution of the tracking object solid is obtained. That is, the spectral distribution of the tracking object solid is obtained from the spectral distribution of the incident light obtained in S606 and the spectral reflectance of the tracking object acquired in S603.

  In S608, the installation location of the imaging device is stored as information for identifying the imaging device in the installation location information 701 of the imaging device in the tracking target object information table. Further, the captured time information stored in the storage device 407 of the imaging apparatus is acquired and stored in the shooting time information 706 of the tracking target object information table. The tracking object information table stored in the RAM 303 is transferred to the monitoring control apparatus 101 as tracking object information.

  As described above, in this embodiment, a background object having the same incident light as that of the tracking object is acquired, and the spectral distribution of the incident light of the tracking object is estimated from the solid spectral distribution of the background object held in advance. Obtain the spectral distribution of the tracking object solid.

<Background object information table acquisition process of monitoring device>
An example of background object information table acquisition processing in the present embodiment will be described with reference to FIGS. 10, 11, and 12. FIG. 10 is a flowchart for explaining an example of the background object information table acquisition process. FIG. 11 shows an example in which the imaging region is divided. FIG. 12 shows an example of the background object information table. The background object information table acquired by the following processing is stored in the ROM 302 or RAM 303 of the object detection device 203 and the external storage device 304.

  In step S1001, a shooting area in which a person can operate is divided into equal intervals with the location of the shooting apparatus as the origin. An example in which the imaging region is divided is shown in FIGS. For example, FIG. 11B is a diagram in which the shooting area is divided from the sky at equal distance intervals from the monitoring device. FIG. 11A is a diagram in which a region photographed by the monitoring device is divided at equal distance intervals in parallel with the ground, and the photographing region is divided from the monitoring device at equal distance intervals from the sky as in FIG. 11B. It is.

  The divided shooting areas are managed by numbers, and the numbers are stored in shooting area division numbers 1201 of the background object information table. The divided areas are stored in the shooting area information 1202 of the background object information table.

In step S <b> 1002, the user determines a background object to be captured in the captured image for each divided area. For example, as shown in FIG. 11A, it is determined what a background object is a subject that is captured for each divided area.
More specifically, the wall is determined to be a background object in the region 14 of FIG. The position information of the extracted background object is acquired by the sensor of the imaging device. The acquired position information is stored in the background object position information 1203.

  In step S1003, in step S1002, the user acquires the spectral distribution (Rx (λ)) of the background object solid using the spectrocolorimeter with respect to the acquired background object. The obtained spectral distribution (Rx (λ)) of the background object solid is input together with the identification number of the divided area by the input device 306 of the object detection device 203 and stored in the background object solid spectral distribution 1204 of the background object information table. To do. In this embodiment, the spectral distribution of the solid of the background object is acquired for each region. However, if the background object is the same, the color may be measured once. In this embodiment, the spectral distribution of the background object solid is stored for each region. However, the background object information table and the spectral distribution of the background object solid may be managed separately as separate tables.

<Detection processing of tracking object in monitoring device>
The tracking target object detection process in this embodiment will be described with reference to the flowchart of FIG.
In step S <b> 901, the tracking target object information transferred from the monitoring control apparatus 101 to each monitoring apparatus (explained by the monitoring apparatus 103 as an example) is acquired via the network 108. The acquired additional object information is held in the RAM 303 of the monitoring device 103.

  In step S902, the background object information table of the monitoring apparatus 103 is acquired. The background object information table acquisition process is the same as the “background object information table acquisition process” in S601 of “tracking target object acquisition process”. Here, the background object close to the tracking object solid is extracted, and the background object information of the extracted background object is acquired.

In step S <b> 903, attention image data is acquired from the storage device 407 in order to acquire a tracking target candidate from the video data captured by the monitoring device 103. That is, spectral image data captured by a monitoring device other than the monitoring device that identifies the tracking object is acquired from the monitoring device. This spectral image data is attention image data.
In step S904, objects existing in the target image data are automatically extracted. Then, in the extracted object, the target object image data is set. Then, the spectral reflectance of the object that is a candidate for the tracking target is acquired from the spectral image data acquired in S903. Here, instead of performing person detection based on color feature amounts for all the extracted objects, the color feature amounts are determined in advance using conventional recognition technology (detection by detecting the state of face recognition or walking). The object to be obtained may be selected and determined.

In step S905, the background object information of the candidate object image is acquired. This processing is performed in the same manner as “acquisition of background information of tracking target” in S604 of “tracking target acquisition processing”. In other words, the background object corresponding to the object to be tracked object candidate is identified from the background object information held in the monitoring device that captured the object to be tracked object object, and the spectral reflectance of the background object is determined. , Acquired from the spectral image data acquired in S903.
In S906, the spectral distribution (Rm (λ)) of the background object solid is acquired from the background object information obtained in S905 and the background object information table obtained in S902. That is, the spectral distribution of the solid of the background object is acquired from the background object information held in the monitoring device that captured the object that is a candidate for the tracking target.

In S907, the spectral reflectance (Cm (λ)) of the background object obtained in S906 and the spectral reflectance (Cm (λ)) of the background object are obtained from the target image data obtained in S903. A spectral distribution (Em (λ)) of incident light is calculated. That is, the spectral distribution of the incident light on the object to be a tracking target object is obtained from the solid spectral distribution of the background object acquired in S906 and the spectral reflectance of the background object acquired in S905. That is, similarly to the tracking object, the spectral distribution of the incident light on the object to be a tracking object candidate in different captured image data is obtained.
Em (λ) = Cm (λ) / Rm (λ) Equation 3
The calculated spectral distribution (Em (λ)) of incident light is stored in the RAM 303 as photographing incident light information 705 in the object candidate information table.

In S908, the spectral reflectance (Cn (λ)) of the object candidate is acquired from the spectral distribution (Em (λ)) of the incident light acquired in S907 and the target image data acquired in S904. The spectral distribution (Rn (λ)) is calculated by Equation 4. As described above, the solid spectral distribution of the object as the tracking target candidate is obtained from the spectral distribution of the incident light obtained in S907 and the spectral reflectance of the object as the tracking target candidate acquired in S904. That is, the solid spectral distribution of the object as the tracking target candidate is obtained from the spectral distribution of the incident light calculated at S606 and the spectral reflectance of the object as the tracking target candidate calculated at S908.
Rn (λ) = Cn (λ) × Em (λ) Equation 4
The obtained spectral distribution (Rn (λ)) of the target object solid is stored in the spectral reflectance 704 of the target object solid image in the target object information table.

In S909, the spectral distribution (Ry (λ)) of the tracking target object solid is acquired from the RAM 303 from the tracking target information acquired in S901.
In S910, the spectral distribution (Rn (λ)) of the tracking object candidate solid obtained in S908 and the spectral distribution (Ry (λ)) of the tracking target solid acquired in S909 are acquired and compared to obtain the similarity. . Note that the size of the tracking target object and the tracking target candidate image does not necessarily match. Therefore, for example, the maximum and minimum ranges of each wavelength are obtained and compared in the spectral reflectance of the object.

Taking FIG. 13 as an example, it is assumed that the spectral reflectance of the tracking object solid is in the range as shown in FIG. Further, when the spectral reflectance of the solid in the tracking target candidate is in the range as shown in FIG. 13B, a difference as shown in FIG. 13C occurs when the two are compared.
If the spectral reflectance is included within the threshold value determined for each wavelength, the object candidate information table is transferred as tracking object candidate information to the monitoring control apparatus 101 as a tracking object candidate in S912. In addition, the acquired object candidate is transferred from a thing with a high similarity. In this way, the spectral distribution of the tracking object solid obtained in S607 is compared with the spectral distribution of the solid object of the object that is the tracking object candidate obtained in S908, and the tracking object candidate is obtained based on the similarity. Determine object information. As described above, in the present embodiment, the solid spectral distribution of the tracking target object is compared with the solid spectral distribution of the tracking target candidate object. The spectral distribution of each solid is determined as follows. That is, the background object is extracted from the captured image data, the spectral distribution of the incident light of the background object is obtained, and the spectral distribution of the incident light is used to determine the spectral of the tracking object or the candidate object of the tracking object. The spectral distribution of the solid is obtained from the reflectance.

[Second Embodiment]
The monitoring apparatus 201 according to the first embodiment of the present invention includes an object detection apparatus 203 in each monitoring apparatus 201 in order to reduce the data to be transferred to the monitoring control apparatus and distribute the object detection processing. The apparatus 202 and the object detection apparatus 203 were comprised. However, in the second embodiment, considering the installation burden of the object detection device 203 in each monitoring device 201, the object detection processing is performed by the monitoring control device.

An example of components of the image processing apparatus according to the second embodiment will be described with reference to FIG.
As shown in FIG. 14, in the image processing apparatus of this embodiment, the monitoring control device 1401 includes a monitoring control unit 1402 and an object detection unit 1403. In addition, the monitoring device includes an imaging device 1410.
In this case, video data captured by the imaging device 1410 is sequentially transferred to the monitoring control device 1401 via the network 108.

  The video data is managed in association with the transferred imaging device. The background information table held in each monitoring device is managed by the object detection unit 1403 in association with the monitoring device and the background information table. As a result, the video data is captured on the distributed imaging devices 1410 (monitoring cameras), and data management and detection processing of image data and the like are performed in the monitoring control device 1401.

  In the acquisition processing of the background object information table of the monitoring apparatus of the image processing apparatus according to the first embodiment, a person obtains the spectral distribution of the solid of the background object using a spectrocolorimeter with respect to the background object. However, in the image processing apparatus according to the third embodiment, the present invention is not limited to this, and if there is a database that holds the spectral distribution of the solid of the existing background object, it may be obtained from that.

(Other embodiments)
The present invention can also be realized by executing the following processing. That is, software (computer program) that implements the functions of the above-described embodiments is supplied to a system or apparatus via a network or various computer-readable storage media. Then, the computer (or CPU, MPU, etc.) of the system or apparatus reads out and executes the program.

101 monitoring control devices 102 to 107 monitoring device 108 network

Claims (9)

In an image processing apparatus configured by a monitoring control device and a plurality of monitoring devices, analyzing image data taken by the plurality of monitoring devices and specifying an object,
Holding means for holding background object information having position information of an object existing in a captured image region and spectral distribution information of an object solid;
Extracting means for extracting a background object close to the tracking object solid, and acquiring background object information of the extracted background object;
Calculating means for obtaining a spectral distribution of incident light on the background object from the extracted background object information and the spectral reflectance of the captured image data;
A calculation means for obtaining a spectral distribution of the tracking object solid by obtaining a spectral reflectance of the tracking object from a spectral distribution of incident light on the background object and the captured image data;
Similar to the tracking object, calculation means for obtaining a spectral distribution of incident light on an object to be a tracking object candidate in different captured image data;
From the calculated spectral distribution of the incident light and the spectral reflectance of the object to be tracked object candidate, a calculation means for obtaining the solid spectral distribution of the object to be tracked object candidate;
The obtained spectral distribution of the tracking object solid is compared with the obtained spectral distribution of the solid object of the object to be tracked, and information on the object to be the tracking object candidate is determined based on the similarity. A determination means;
An image processing apparatus comprising: In an image processing apparatus configured by a monitoring control device and a plurality of monitoring devices, analyzing image data taken by the plurality of monitoring devices and specifying an object,
First acquisition means for acquiring spectral image data obtained by photographing the tracking object from the monitoring device;
Second acquisition means for acquiring the spectral reflectance of the tracking object from the spectral image data acquired by the first acquisition means;
A third acquisition means for identifying a background object corresponding to the tracking object from the background object information held in the monitoring device that captured the tracking object, and acquiring the spectral reflectance of the background object from the spectral image data; ,
A fourth acquisition means for acquiring the spectral distribution of the solid of the background object from the background object information held in the monitoring device that captured the tracking object;
A first calculation for obtaining a spectral distribution of incident light on the tracking object from the spectral distribution of the solid of the background object acquired by the fourth acquisition means and the spectral reflectance of the background object acquired by the third acquisition means. Means,
Second calculation means for obtaining the spectral distribution of the tracking object solid from the spectral distribution of the incident light obtained by the first calculation means and the spectral reflectance of the tracking object obtained by the second acquisition means;
Fifth acquisition means for acquiring from the monitoring device spectral image data captured by a monitoring device other than the monitoring device that identifies the tracking object;
Sixth acquisition means for acquiring a spectral reflectance of an object to be a tracking target candidate from the spectral image data acquired by the fifth acquisition means;
The background object corresponding to the object to be tracked object candidate is identified from the background object information held in the monitoring device that has photographed the object to be tracked object candidate, and the spectral reflectance of the background object is determined as the spectral object. Seventh acquisition means for acquiring from image data;
An eighth acquisition means for acquiring the spectral distribution of the solid of the background object from the background object information held in the monitoring device that captured the object that is the candidate for the tracking object;
Based on the spectral distribution of the solid of the background object acquired by the eighth acquisition unit and the spectral reflectance of the background object acquired by the seventh acquisition unit, the spectral distribution of the incident light on the object to be tracked object is obtained. A third calculating means to obtain;
Based on the spectral distribution of the incident light obtained by the third calculating means and the spectral reflectance of the object as the tracking target candidate acquired by the sixth acquiring means, the spectral distribution of the solid of the object as the tracking target candidate is obtained. A fourth calculating means for obtaining
The spectral distribution of the tracking target object solid obtained by the second calculation means is compared with the spectral distribution of the solid object of the object that is the tracking target candidate obtained by the fourth calculation means, and tracking is performed based on the similarity. A determination means for determining information on an object that is a candidate for a target;
An image processing apparatus comprising: The background object information is held in advance by each monitoring device,
The information of the divided area, the position information of the background object existing in the divided imaging area, and the spectral distribution of the solid of the background object are held for each area obtained by dividing the imaging area. Image processing apparatus. The background object information is held in advance by the monitoring control device,
The information of the divided area, the position information of the background object existing in the divided imaging area, and the spectral distribution of the solid of the background object are held for each area obtained by dividing the imaging area. Image processing apparatus. In a control method of an image processing apparatus configured by a monitoring control device and a plurality of monitoring devices, analyzing image data taken by the plurality of monitoring devices and specifying an object,
A holding step for holding background object information having position information of an object existing in a captured image region and spectral distribution information of an object solid;
An extraction step of extracting a background object close to the tracking object solid and acquiring background object information of the extracted background object;
A calculation step for obtaining a spectral distribution of incident light on the background object from the extracted background object information and the spectral reflectance of the captured image data;
Obtaining the spectral reflectance of the tracking object from the spectral distribution of the incident light on the background object and the captured image data, calculating the spectral distribution of the tracking object solid,
Similar to the tracking object, a calculation step for obtaining a spectral distribution of incident light on an object to be a tracking object candidate in different captured image data;
From the calculated spectral distribution of the incident light and the spectral reflectance of the object as the tracking target object, a calculation step for obtaining the solid spectral distribution of the object as the tracking target candidate;
The obtained spectral distribution of the tracking object solid is compared with the obtained spectral distribution of the solid object of the object to be tracked, and information on the object to be the tracking object candidate is determined based on the similarity. A decision process;
A control method for an image processing apparatus, comprising: In a control method of an image processing apparatus configured by a monitoring control device and a plurality of monitoring devices, analyzing image data taken by the plurality of monitoring devices and specifying an object,
A first acquisition step of acquiring, from the monitoring device, spectral image data obtained by imaging the tracking object;
A second acquisition step of acquiring the spectral reflectance of the tracking object from the spectral image data acquired in the first acquisition step;
A third acquisition step of identifying a background object corresponding to the tracking object from the background object information held in the monitoring device that captured the tracking object, and acquiring the spectral reflectance of the background object from the spectral image data; ,
A fourth acquisition step for acquiring a spectral distribution of the solid of the background object from the background object information held in the monitoring device that images the tracking object;
The first calculation for obtaining the spectral distribution of the incident light on the tracking object from the spectral distribution of the solid of the background object acquired in the fourth acquisition step and the spectral reflectance of the background object acquired in the third acquisition step. Process,
A second calculation step of obtaining a spectral distribution of the tracking object solid from the spectral distribution of the incident light obtained in the first calculation step and the spectral reflectance of the tracking object obtained in the second acquisition step;
A fifth acquisition step of acquiring from the monitoring device spectral image data captured by a monitoring device other than the monitoring device that identifies the tracking object;
A sixth acquisition step of acquiring the spectral reflectance of the object that is a candidate for the tracking target from the spectral image data acquired in the fifth acquisition step;
The background object corresponding to the object to be tracked object candidate is identified from the background object information held in the monitoring device that has photographed the object to be tracked object candidate, and the spectral reflectance of the background object is determined as the spectral object. A seventh acquisition step of acquiring from image data;
An eighth acquisition step of acquiring the spectral distribution of the solid of the background object from the background object information held in the monitoring device that captured the object that is a candidate for the tracking target;
Based on the spectral distribution of the solid of the background object acquired in the eighth acquisition step and the spectral reflectance of the background object acquired in the seventh acquisition step, the spectral distribution of incident light on the object that is a candidate for the tracking target is calculated. A third calculation step to be obtained;
Based on the spectral distribution of the incident light obtained in the third calculation step and the spectral reflectance of the object as the tracking target candidate acquired in the sixth acquisition step, the spectral distribution of the solid of the object as the tracking target candidate A fourth calculation step for obtaining
The spectral distribution of the tracking target object solid obtained in the second calculation step is compared with the spectral distribution of the solid of the object that is the tracking target candidate obtained in the fourth calculation step, and tracking is performed based on the similarity. A determination step for determining information on objects that are candidates for the object;
A control method for an image processing apparatus, comprising: In a program comprising a monitoring control device and a plurality of monitoring devices, causing a computer to execute a control method of an image processing device that analyzes image data taken by the plurality of monitoring devices and identifies an object,
A holding procedure for holding background object information having position information of an object existing in a captured image region and spectral distribution information of an object solid,
An extraction procedure for extracting a background object close to the tracking object solid and acquiring background object information of the extracted background object;
A calculation procedure for obtaining a spectral distribution of incident light on the background object from the extracted background object information and the spectral reflectance of the captured image data,
Obtaining the spectral reflectance of the tracking object from the spectral distribution of the incident light on the background object and the captured image data, and calculating the spectral distribution of the tracking object solid,
Similar to the tracking object, a calculation procedure for obtaining a spectral distribution of incident light on an object to be a tracking object candidate in different captured image data;
From the calculated spectral distribution of the incident light and the spectral reflectance of the object as the tracking target object, a calculation procedure for obtaining the solid spectral distribution of the object as the tracking target candidate;
The obtained spectral distribution of the tracking object solid is compared with the obtained spectral distribution of the solid object of the object to be tracked, and information on the object to be the tracking object candidate is determined based on the similarity. Decision procedure;
A program that causes a computer to execute. In a program comprising a monitoring control device and a plurality of monitoring devices, causing a computer to execute a control method of an image processing device that analyzes image data taken by the plurality of monitoring devices and identifies an object,
A first acquisition procedure for acquiring spectral image data obtained by photographing a tracking object from a monitoring device;
A second acquisition procedure for acquiring the spectral reflectance of the tracking object from the spectral image data acquired in the first acquisition procedure;
A third acquisition procedure for identifying a background object corresponding to the tracking object from the background object information held in the monitoring device that captured the tracking object and acquiring the spectral reflectance of the background object from the spectral image data; ,
A fourth acquisition procedure for acquiring the spectral distribution of the solid of the background object from the background object information held in the monitoring device that captured the tracking object;
First calculation for obtaining the spectral distribution of the incident light on the tracking object from the spectral distribution of the solid of the background object acquired in the fourth acquisition procedure and the spectral reflectance of the background object acquired in the third acquisition procedure. Procedure and
A second calculation procedure for obtaining a spectral distribution of the tracking object solid from the spectral distribution of the incident light obtained in the first calculation procedure and the spectral reflectance of the tracking object obtained in the second acquisition procedure;
A fifth acquisition procedure for acquiring, from the monitoring device, spectral image data captured by a monitoring device other than the monitoring device that identifies the tracking object;
A sixth acquisition procedure for acquiring the spectral reflectance of an object as a tracking target candidate from the spectral image data acquired in the fifth acquisition procedure;
The background object corresponding to the object to be tracked object candidate is identified from the background object information held in the monitoring device that has photographed the object to be tracked object candidate, and the spectral reflectance of the background object is determined as the spectral object. A seventh acquisition procedure for acquiring from image data;
An eighth acquisition procedure for acquiring the spectral distribution of the solid of the background object from the background object information held in the monitoring device that captured the object to be tracked object candidate;
From the spectral distribution of the solid of the background object acquired in the eighth acquisition procedure and the spectral reflectance of the background object acquired in the seventh acquisition procedure, the spectral distribution of the incident light on the object that is a candidate for the tracking target is calculated. A third calculation procedure to be obtained;
The spectral distribution of the solid of the object that is a candidate for the tracking target based on the spectral distribution of the incident light obtained in the third calculation procedure and the spectral reflectance of the object that is the candidate for the tracking target acquired in the sixth acquisition procedure. A fourth calculation procedure for obtaining
The spectral distribution of the tracking object solid obtained in the second calculation procedure is compared with the spectral distribution of the solid of the object that is a candidate for the tracking object obtained in the fourth calculation procedure, and tracking is performed based on the similarity. A decision procedure for determining information on objects that are candidates for the object;
A program that causes a computer to execute.   A computer-readable storage medium storing the program according to claim 7 or 8.