JP2001359087A - Image storage method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image storage method by which the genuineness of a photographing time and a photographing place of images whose management is requested can be warranted and a huge amount of images can efficiently be retrieved. SOLUTION: A confirmation means 10 transmits time and a supervisory camera 4 imbeds the time to an image by electronic signature processing or the like. Confirming the imbedded information at its reception warrants the genuineness of the photographing time. Collating a background image having been acquired and stored in advance with a background image of received image data warrants the genuineness of a photographing place. Furthermore, the supervisory camera 4 automatically attaches event information obtained from other sensor to the image to facilitate the retrieval.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] The present invention relates to the storage and retrieval of captured images. Also, the present invention relates to a method and a system for storing and managing a large number of monitoring images on behalf of a monitoring supervisor, searching for necessary parts as required, and safely distributing them.

[0002]

2. Description of the Related Art Today, the amount of captured and stored images is:
It has increased. For example, due to the reduction in the price of video cameras, surveillance by moving image shooting has become common, and the number of shooting points is steadily increasing in the future. For example, security cameras are installed in retail stores such as supermarkets and convenience stores, financial institutions such as banks and post offices, and elevators in most cases. In addition, video cameras are installed everywhere in urban space, such as on expressways, at intersections, inside railway stations, and along rivers, and used for various surveillance operations.

In such a surveillance camera, not only the number of surveillance points, but also the amount of information per point has been increased due to colorization, high resolution, and further digitization.

As a method of accumulating and storing such a large amount of surveillance image data at low cost, a magnetic tape is generally used. For the search, a tag indicating the photographing period and the like is added to each tape cassette, and the necessary cassette is specified based on the tag. Regarding the specification of the necessary scene in the cassette, a continuous search is required due to the characteristics of the magnetic tape, and moreover, it is necessary to rely on humans. Magnetic disks and DVDs that can be accessed at random are becoming popular due to lower prices. However, except for those made for a specific purpose, tag information, which is a mark for identifying the necessary parts in the disk, is used. Most of the data we do not have, we have to search continuously.

[0005]

A large amount of moving image data generated by photographing for the purpose of monitoring as described above is:
Most of the data has no opportunity to be reused, and high costs cannot be spent on maintenance. However, when an accident or crime occurs, a small portion may become valuable data and must be kept for a certain period of time.

[0006] Holders of such infrequently used and large amounts of data seek to save storage space and costs. In this case, when searching for an image, there are the following problems.

First, it is necessary to provide an appropriate search means for a large amount of moving image data. With centralized management, the amount of data handled becomes enormous, and it becomes more difficult to extract necessary scenes. Although digitization and compression improve access speed and storage space, high-speed retrieval cannot be realized unless keys serving as markers for extracting necessary scenes are appropriately added. The problem is how to easily and accurately add the search key.

Second, the necessary scene is an important part related to, for example, an accident or a crime. Therefore, it must be able to guarantee authenticity that there is no falsification or forgery. While digitization has made it easier to process information, it has also increased the risk of unauthorized copying, tampering, or forgery. In particular, when information is transmitted and received electronically via a network, there is a great risk that unauthorized processing will be performed in the transfer process, and the authenticity assurance function becomes more important for a third-party management agent. . In particular, for moving image data, the problem is how to guarantee the authenticity of the shooting location and shooting time.

[0009]

In order to add an appropriate search key to image data, a video camera and other sensors are linked. For example, some event information such as opening and closing of a door obtained from a sensor of an automatic door at the entrance is added to a moving image. In the first place, the required scene is often related to the occurrence of some event, and this is considered to greatly improve the search function. Further, a search key adding mechanism can be realized relatively easily and can be automatically processed. The event information is
This indicates an event that has occurred in an area related to image data.

If the data at the time of reception is correct, it is possible to guarantee the authenticity of the data shooting location and data shooting time by performing falsification prevention processing by electronically watermarking or the like. However, when data is received via a network, it is possible that the data has already been invalidated at the time of reception. Therefore, for the shooting location, the background image is separately stored, and a guarantee is obtained by collation with the background image. The photographing time is guaranteed by managing time stamp processing by transmitting time data from a server or the like owned by the management agent to the photographing device.

Further, the present invention also includes the following configuration. Store and store data from multiple data holders,
Perform maintenance. Image data captured by a plurality of video cameras is received via a network and stored in a predetermined computer system.

[0012]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention relates to a method and system for managing a large amount of surveillance image data on behalf of a supervisor and presenting it to an image utilization organization. FIG. 1 shows the whole of the embodiment. The monitoring center 2 performs a monitoring operation using the monitoring camera 4. A typical example is
Security cameras installed in building managers, security companies, elevators, convenience stores and bank stores. Usually, images taken by such security cameras are stored on a video cassette tape or the like and stored for a certain period of time. For example, when some kind of crime occurs, these image data are used by the image utilizing organization 20 as a criminal investigation or evidence. Typical examples of the image utilizing organization 20 include a police and a court. In such a case, the supervisor needs to store a large amount of surveillance image data, and the image use organization requires a great deal of effort to extract necessary scenes from the huge amount of data. .

The video surveillance center 8 stores and manages a large number of surveillance images on behalf of the person in charge of surveillance, and provides a search service to an image utilizing organization, and is an implementation example of the present invention. The information processing according to the present invention is executed by a computer located or available at the monitoring center 2, the certification organization 6, the image user 20, the video surveillance center 20, and the like.

The video surveillance center 8 receives surveillance image data from the surveillance center 2 via the network 7, receives access from the image utilization organization 20, provides a search service, cuts out necessary scenes, and sends them to the user. . By providing such a center, the person in charge of monitoring is released from the storage location and maintenance of a large amount of image data, and the convenience for the image using organization is improved. However, exchanging information on a network in this way carries the risk of unauthorized access to data, and unauthorized tampering or forgery.

First, the video surveillance center 8 checks whether or not the photographer, the photographing place, and the photographing time of the received image data are correct by the confirming means 10. Regarding the confirmation of whether the sender of the received data is an unauthorized person,
Confirmation is made by making an inquiry to the certification organization 6 which is a certification organization.

The authenticity of the photographer, photographing time, and photographing place of the image data performed by the confirmation means 10 may be performed by the notary institution 9 as a third party organization. The monitoring center 2 sends the image data to the notary institution 9 and, after adding the notarization information, to the video surveillance center 8.
The confirmation means 10 authenticates the notary institution 9 that is the transmission source and confirms the added notarization information. In this example, the video surveillance center provides image data storage and retrieval services.

The responsibility for ensuring the authenticity of the photographer, the photographing time, and the photographing place is borne by the monitoring center.
It may be a simple form of sending directly without going through the notary authority 9. In this case, the confirmation unit authenticates the monitoring center 2. For example, if the monitoring target is a river or plant,
The image use organization is the same as or related to the person in charge of photography, does not require a strict notarization function, and is suitable for a storage and retrieval service. On the other hand, when an image relating to a crime with a security camera as in the above-described example or an image of a traffic violation or a traffic accident during road surveillance is used as evidence, a notarization function is required. The confirmation method by the confirmation means 10 may be changed according to the image data.

The image data whose authenticity has been confirmed is subjected to appropriate falsification prevention processing such as an electronic watermark or signature using the processing means 12, and then stored and managed in the storage means 14.

On the other hand, with respect to browsing and retrieval requests from the image utilizing organization 20, appropriate access control such as password processing is performed. The search means 16 provides efficient image extraction of necessary scenes from the image data of the storage means 14 to the authorized image utilizing organization. The problem is how to generate a search key for efficient search. For this purpose, a mechanism for efficiently generating a search key at the time of shooting is provided. This method will be described later.

The distribution means 20 sends the requested image to the image utilizing organization 20. At this time, the authentication organization 6 is used to prevent unauthorized tampering or duplication in the process of being distributed on the network.

As described above, at the time of proxy management of a large number of images, the authenticity of the data is confirmed at the time of data reception, falsification prevention processing is performed at the time of storage, and an efficient search service is provided for an access request from a user. By performing confidential communication in response to a data transmission request, it is possible to realize a management method capable of avoiding a risk occurring during information transmission on a network.

FIG. 2 is a diagram for explaining a method of confirming the photographer, photographing time and photographing place of the image in the confirming means 10. The confirmation means 10 transmits the correct time from the time server 102 to the management center, and the monitoring center 2 embeds this in image data at the time of shooting. At the time of embedding, a technique such as an electronic signature or a watermark is applied so that falsification is difficult. The management center applies these registrations to the certification organization 6 for each photographing device. The certification organization 6 sends a secret key corresponding to each photographing device to the management center 2 based on the public key method. Video surveillance center 8
Can obtain the corresponding public key from the certificate authority 6.

The image embedded with the time managed by the checking means 102 is encrypted by the surveillance center 2 using the secret key obtained from the certification organization 6 and transmitted to the video surveillance center 8. Upon receiving the image, the verification unit 106 obtains the public key corresponding to the image from the certification organization 6 and decrypts the public key. In this way, the device that has captured the image is authenticated. The photographing time is confirmed by confirming the embedded time information. Although it is possible to confirm that the photographing location is also correct by the authentication of the photographing device, the confirmation may be performed using a background image as follows.

The background image of the corresponding photographing device is stored in the background image DB 104 in advance, and the photographing place is confirmed by comparing this with the background of the received image. If the image is not the fixed point monitoring, information on the zoom and the rotation angle from the reference state is added to the image, and the image is converted to the reference state based on the information, and then collated. The image data for which the authenticity of the photographer, the photographing time, and the photographing place have been confirmed by such a procedure is subjected to appropriate falsification prevention processing such as an electronic watermark or signature using the processing means 12. Is stored in the storage means 14.

Next, a mechanism for providing appropriate search means will be described. The point is to generate an appropriate search key for efficient search. The process of adding a search key to data requires additional labor and cost if it is performed as a separate process after the generation and accumulation of data. Therefore, it is efficient if it can be automatically added at the time of generation. In many cases, a surveillance image targeted by the present invention becomes a necessary scene when a certain event (event) occurs. Therefore, as shown in FIG. 3, in the present invention, a signal from the sensor 5 for detecting this is synchronized with the surveillance camera 4 so that the tag generation means 24 generates a search key simultaneously with image capturing and adds it to the image data. . Taking a camera monitoring a bank cash dispenser as an example, an event of "people entering and leaving" can be considered, and the sensor for detecting this event is an automatic door sensor at the entrance. The control unit of the cash dispenser is used as a sensor to detect an event such as “payment of more than 300,000 cash”. Taking security cameras at convenience stores as an example,
By opening and closing the door of the refrigerator, interlocking with the POS terminal, and the like, it is possible to add detailed event information such as "the customer tried to buy drinking water" or "the woman in her twenties went shopping".

FIG. 4 shows examples of generated tags and transmitted image data. Among the tags generated, the sensor I
D is information for specifying the sensor that caught the event. The camera ID is information for identifying a camera related to the caught event, and a plurality of cameras may be associated with one sensor. For the caught event, the name of the event, the time at which the event occurred, or the time period during which the event occurred is information included in the tag. In addition, attribute information related to the event may be added. For example, in the above-described example of the cash dispenser, the event name is “payout”, the additional attribute is “payout amount”, and the attribute value is “300,000 or more”.

On the other hand, the image data is added to the ID of the camera that is shooting and information of the shooting period to form one record. The shooting period may be fixed length or variable length. The shorter and shorter the shooting period, the higher the efficiency of search and distribution.

In the search, when a place and a time are input, a necessary part is extracted from the camera ID and the photographing period added to the image data. When the event name or its attribute value is input, the camera ID and the event occurrence time are known from the matched tag, and the necessary part is extracted based on this.

FIG. 5 is a diagram showing timings at which tags and image data are generated and transmitted. The measurement data from the sensor 5 is input to the tag generation unit 24. Tag generation means analyzes the input data in real time,
If the event is recognized, a tag in the format illustrated in FIG. 4 is generated. On the other hand, information of the format illustrated in FIG. 4 is added to the image data captured by the camera 4 at certain time intervals. A tag generated during the shooting period is added to one record of this image data,
Is input to The encrypting means 26 immediately encrypts this using the secret key and transmits it to the video surveillance center.

The video surveillance center 8 may centrally manage the shooting time of the image. However, the surveillance center 2 has a local time server 22, and the video surveillance center 8 A method in which 102 are synchronized may be used. FIG.
This describes a camera for guaranteeing the authenticity of the shooting time. The time to be guaranteed is based on the time server 102 of the confirmation means 10 of the surveillance center or the notary. Time information from the time server 102 or a local time server 22 of the monitoring center synchronized with the time server 102 is input to the camera 4. The camera 4 embeds the time information in the image data generated by the photographing means 44 by using the built-in time information adding means 42. As a method of embedding this information, a technique of a digital watermark or a digital signature is applied, and the collation means 106 can be used to confirm whether there is any falsification or forgery. Thus, the authenticity of the photographed time can be confirmed after the photographing.

Next, FIG. 7 shows a method of distributing a necessary scene to an image utilizing organization. The image utilization organization 20 accesses the video surveillance center 8 via a network, and searches for image data stored in the storage unit 14 using the search unit 16. As described above, the search can be performed using not only the shooting time and the shooting location but also the event information and the like, and the necessary scene can be efficiently extracted. When a distribution request is issued for the extracted necessary scene, the distribution means obtains the public key of the image using organization 20 from the certification organization 6 and uses this to obtain the necessary scene and the notarization information such as the photographing time and the photographing place. To encrypt. The image utilizing organization 20 obtains image data of a required scene by performing decryption using a secret key obtained from the authentication organization 6 in advance.

According to the above embodiment, a large number of surveillance images can be managed intensively and efficiently on behalf of the supervisor. The agent can provide access to a large number of images and a search service through a network to a supervisor in charge or other image users. In transmitting and using information, unauthorized tampering and forgery can be prevented, and authenticity can be guaranteed.

[0033]

According to the present invention, it is possible to efficiently manage images. It becomes possible to accumulate images so that image retrieval can be easily performed, and to easily retrieve images.

[Brief description of the drawings]

FIG. 1 is a diagram showing an embodiment of a system implementing a method according to the present invention and the relationship between related organizations and users.

FIG. 2 is a diagram showing a method for confirming the authenticity of a photographer, a photographing time, and a photographing place of an image to be managed, and an example of realizing the method.

FIG. 3 is a diagram showing a method for automatically adding a tag required for a search to an image to be managed simultaneously with shooting, and an example of realizing the method.

FIG. 4 is a diagram illustrating an example of a data format of tag information and image information.

FIG. 5 is a diagram illustrating timings at which tag information and image information are generated and transmitted.

FIG. 6 is a diagram illustrating a configuration and a mechanism of a camera for guaranteeing a shooting time.

FIG. 7 is a diagram showing a method for efficiently searching image data and providing a secure distribution service of necessary scenes to an image using organization, and an example of implementation thereof.

[Explanation of symbols]

2 surveillance center, 4 surveillance camera, 6 certification body, 7
Network, 8 ... Video surveillance center, 20 ...
Image users

──────────────────────────────────────────────────の Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat ゛ (Reference) H04N 1/387 H04N 1/387 (72) Inventor Yasushi Fukumoto 1099 Ozenji, Aso-ku, Kawasaki-shi, Kanagawa Prefecture, Ltd. Inside Hitachi Systems Development Laboratory (72) Inventor Katsumi Kono 1099 Ozenji Temple, Aso-ku, Kawasaki City, Kanagawa Prefecture Incorporated Hitachi Systems Development Laboratory (72) Koichi Sano 1099 Ozenji Temple, Aso-ku, Kawasaki City, Kanagawa Prefecture, Ltd. Hitachi Systems Development Laboratory (72) Inventor Hiroshi Koumi 4-6, Kanda Surugadai, Chiyoda-ku, Tokyo F-term (Reference) 5B075 KK54 KK66 ND12 NK10 NK24 UU40 5C054 AA01 DA08 EA01 EA07 FC13 FE14 FE23 FE28 GB02 GB06 GB14 GD05 GD06 HA18 HA21 5C076 AA02 AA 14 BA06 5L096 BA02 CA02 EA35 EA45 HA02 HA07 JA03 JA18

Claims (12)

[Claims] 1. An image storage method in an image storage device for storing an image captured by an image capturing device, wherein a background image indicating a location captured by the image capturing device is
It is stored in advance, a captured image captured by the image capturing device is input, the captured image is compared with the background image, and it is determined whether the captured image has a predetermined relationship with the background image, An image storage method, wherein the captured image is stored when the captured image has a predetermined relationship with the background image. 2. The image storage method according to claim 1, wherein the determining process is performed when a predetermined relationship between the background image and an image indicating a background that has not fluctuated for a predetermined period or more in the captured image matches. An image storage method, characterized in that it is determined that the image has an image. 3. The image storage method according to claim 2, wherein said determining step determines whether or not they match in consideration of aging. 4. The image storage method according to claim 1, wherein the background image and the photographed image are sent to the image storage device through respective paths. 5. The image storage method according to claim 4, wherein the captured image is transmitted from the image capturing device to the image storage device using a first network, and the background image is stored in the image capturing device. An image storage method, wherein the image is sent from the device to the image storage device by using at least one of a second network and transportation on a recording medium capable of recording the image. 6. The image storage method according to claim 1, wherein said stored background image is replaced every predetermined period. 7. An image storage method in an image storage device for storing a region image in which a predetermined region has been photographed, wherein the step of receiving the region image photographed by the image photographing device comprises the steps of: Receiving an event information indicating an event that has occurred in relation to the area; and storing the received area image and the detected event information in association with each other for each time. . 8. The image storage method according to claim 7, wherein the image photographing device and the detection device are connected to the image storage device via a network, and wherein the area image and the event information are transmitted through the network. An image storage method, wherein the image is transmitted to the image storage device via an external storage device. 9. The image storage method according to claim 7, wherein the area is in a bank store, and the detecting device is a door sensor for detecting opening and closing of a door of the bank store and a transaction with a cash vending machine. An image storage method, comprising: at least one of an amount sensor for detecting that the amount is equal to or more than a predetermined amount. 10. The image storage method according to claim 7, wherein the area is in a retail store, and the detecting device is a door sensor for detecting opening / closing of a door of a storage and a POS terminal for detecting purchase information. An image storage method characterized by being at least one of them. 11. An image retrieval system for retrieving an image stored in advance, wherein an area image which is an image of a predetermined area photographed by a photographing device and event information indicating an event occurring in said predetermined area correspond to each time. A storage device connected to the storage device, receives a search request indicating an event transmitted from a terminal device connected to the storage device via a network, and searches for an area image corresponding to the search request. An image search system, comprising: a search device that outputs a searched area image. 12. The image retrieval system according to claim 11, wherein the terminal is connected to an authentication computer for authenticating a human authenticity through the network, and the terminal device requests the authenticity computer for authenticity of a user. The search device transmits, when receiving authenticity information indicating the authenticity of the user from the authentication computer, searching for an area image corresponding to the search request. Search system.