JP5069091B2 - Surveillance camera and surveillance camera system - Google Patents

Description

  The present invention relates to a surveillance camera and a surveillance camera system.

  As a conventional surveillance camera system that detects an abnormal state of an intruder or a facility by detecting a change in the state of an image capturing a monitoring area or records an image, the one described in Patent Document 1 below is known. Yes.

  In the monitoring camera system, it is desired that the moving body (suspicious person) is a clear image that can be identified by a large number of unspecified persons while monitoring and monitoring the moving body (suspicious person). The surveillance camera system of Patent Document 1 is a surveillance camera system that detects a moving object with a wide-angle camera and tracks the moving object with a pan / tilt compatible telephoto camera, so that an intruder can be accurately monitored even in a wide monitoring range. It has become.

JP-A-11-103457

  However, the method disclosed in Patent Document 1 requires two cameras, a wide-angle camera and a telephoto camera, and is adapted to follow the pan / tilt of the telephoto camera based on the position and moving speed information of the moving body captured by the wide-angle camera. Since there is a need to synchronize with two cameras, it is difficult to install the camera and to control the tracking of the telephoto camera, and there is a problem that can only be handled by a high-cost system. In addition, when a plurality of moving bodies move in different directions, there is a problem that one of them cannot be tracked by the telephoto camera. On the other hand, a method of obtaining a high-precision image for a plurality of moving objects by capturing a high-definition image even with only one camera is conceivable, but the amount of information increases and information transmission There was a problem that the load of the.

  The present invention has been made in order to solve the above-described problems. The object of the present invention is to use only one camera even in a wide monitoring range, and not to have a complicated mechanism but to reduce the cost. It is an object of the present invention to provide a surveillance camera system capable of accurately tracking and monitoring a moving body and reducing the amount of data transmission.

The surveillance camera according to the present invention is an image data generation unit that captures an imaging target and generates first full-screen video data, and the first full-screen video data is converted into second full-screen video data having a predetermined pixel size. Image data conversion means for conversion, moving body detection means for detecting a moving body from the first full-screen video data, template block setting means for setting a template block in the moving body, and template blocks from the first full-screen video data. image data transmitting means for transmitting template block extracting means, the image data of the template block to a second time-division multiplexing with full-screen video data or external selectively switching, withdrawn as image data size, tracking a moving object Means, the center of gravity of the template block and the position of the image data, and within the template block Template block related information including luminance values, is configured to include a memory to be stored is assigned a label, means for tracking a moving body, the moving body by the moving body detection means is detected, within a predetermined area If there is no template block stored in the memory in the specified area, the mobile block detected by the mobile unit detection means is set. The template block related information of the created template block is stored in the memory with a label attached, and when there is a template block stored in the memory in a predetermined area, the moving object of the template block stored in the memory The brightness value of the object and the brightness value of the moving object detected by the moving object detection means are compared. If there is similarity within a predetermined range, the template block related information of the template block having the highest similarity is updated to the template block related information of the template block set for the moving body detected by the moving body detecting means. you.

  The surveillance camera according to the present invention reduces the amount of data transmission by thinning the captured first full-screen video data and converting it to second full-screen video data having a predetermined pixel size. . Further, when a moving body is detected in the monitoring area, only the image data of the specific area including the moving body, that is, the template block is extracted from the first full-screen video data without converting the pixel size, so that the transmission data is only slightly increased. Thus, a multiple monitor in which a high-definition moving body is picture-in-picture to the second full-screen video data becomes possible. As a result, it is possible to obtain highly accurate images for a plurality of moving objects with one fixed monitoring camera, and to reduce the amount of data transmission.

  Hereinafter, the present invention will be specifically described with reference to the drawings showing embodiments thereof.

<Embodiment 1>
(Constitution)
The configuration of the surveillance camera system according to the present embodiment will be described below with reference to the drawings. FIG. 1 is a diagram showing a surveillance camera system according to the present embodiment. The surveillance camera system 1 records and displays the input image data by inputting at least a surveillance camera 2 capable of photographing a surveillance target and distributing the network and image data captured by one or a plurality of the surveillance cameras 2. The monitoring terminal device 3 is provided with a function for performing processing and the like, and a display device 4 that receives and displays image data processed by the monitoring terminal device 3.

  The monitoring camera 2 includes a lens unit 21 that collects visible light, and an image sensor (image data generating means) 22 that is configured by a CCD, a CMOS, or the like that converts the collected visible light into image data. Further, the image data input from the image sensor 22 is divided into areas and the resolution setting process is performed by an image data processing unit (image data converting means) 23 and the video data (23i) input from the image data processing unit 23. And a moving body tracking processing unit (moving body detection means) 24 for detecting and tracking the moving body. Further, a LAN that transmits and receives digital data between the image data compression unit (image data compression means) 25 that controls the amount of information of the image data input from the image data processing unit 23 using, for example, the JPEG method, and the monitoring terminal device 3. A connection unit (image data transmission means) 26 and a CPU (Central Processing Unit) 27 that controls each functional unit are provided.

  The monitoring terminal device 3 includes a LAN connection unit 31 that transmits and receives image data and additional information of the monitoring camera 2 via a transmission network such as a LAN, a control unit 32 that includes a CPU and a memory, and acquired compressed image data. And a storage device 33 for recording. Further, a compressed data decompression unit 34 for decoding the compressed image data and a display driver 35 for displaying an image on the display device 4 are provided.

  Next, the detailed configuration and function of each unit will be described. The lens unit 21 of the surveillance camera 2 used in the surveillance camera system 1 varies depending on the installation environment, such as a fixed focus with a focal length of 3 to 9 mm, a varifocal, or a telephoto lens with several tens of millimeters. The surveillance camera 2 is effectively installed by using a wide-angle lens having a horizontal field of view of about 90 degrees and a vertical field of view of about 70 degrees so that the surveillance area can be photographed widely like the photographed image shown in FIG. Is.

  The image sensor 22 is an element that converts the visible light collected by the lens unit 21 into image data for each pixel by a sensor array such as a CCD or CMOS manufactured by a semiconductor process. The digital signal is converted by the instrument and output. The conversion pixel size is generally compatible with VGA of 640 pixels in the horizontal direction and 480 lines in the vertical direction. However, in recent years, the evolution of the semiconductor process and the demand for high-definition images are combined, and the horizontal: 1280 pixels and vertical: In addition to 1024 line SXVGA, an image sensor with horizontal: 2048 pixels and vertical: 1536 lines is also used as a 3 mega pickle sensor. In the present embodiment, the description will be made with an image using horizontal: 1280 pixels, vertical: 960 lines, that is, a VGA quadruple density image sensor.

  The image data processing unit 23 temporarily stores at least one frame of image data of horizontal: 1280 pixels and vertical: 960 lines output from the image sensor 22 in an internal buffer memory, and outputs full screen data. It has a function that can. When the moving object tracking processing unit 24 detects a moving object, the template block data is extracted based on the template block information from the moving object tracking processing unit 24, and the image size and resolution of each template block are extracted. Set up. The template block will be described later.

(Operation)
Next, the operation of the monitoring camera system 1 will be described. First, the necessity of the resolution setting process performed in this embodiment will be briefly described. As shown in FIG. 1, the surveillance camera system 1 generally connects a plurality of surveillance cameras 2 to a surveillance terminal device 3 via a LAN. As a result, under the condition that the effective transmission capacity of the LAN is about 50 Mbps, a megapixel image compressed to about 25 Mbps can be acquired at a high rate if one monitoring camera 2 and the monitoring terminal device 3 are directly connected. When the number of cameras increases, it is clear that the amount of information must be reduced in order for the monitoring terminal device 3 to accurately acquire the images of each monitoring camera 2, and it is necessary to reduce the resolution and / or the number of frames. In surveillance camera systems, an extremely small frame rate is inappropriate to reliably acquire the moment of occurrence of an abnormality, and a system that is more suitable for the purpose when the resolution is reduced to VGA or QVGA (320 × 240 pixels) in normal times. It can be said.

  Therefore, in the present embodiment, SXVGA full-screen video data (first image) having a number of imaging pixels of 1280 × 960 pixels (SXVGA resolution) as shown in FIG. Are stored in the buffer memory of the image data processing unit 23. Further, the image data processing unit 23 samples the SXVGA full-screen video data only in odd-numbered pixels in both the line direction and the vertical direction, and has a pixel size of 640 × 480 pixels (VGA resolution) as shown in FIG. The image data is converted into VGA screen video data (second full screen video data) and output to the image data compression unit 25 as full screen video data. In FIG. 4, ● represents an effective pixel, and x represents a rejected pixel. In the above description, an odd value pixel is used, but an even value pixel may be used. Further, if 640 × 480 pixels are generated with an average value obtained from the four pixels on the left, right, top, and bottom, a higher quality image is obtained.

  Next, processing for detecting a moving body (suspicious person) in the monitoring area will be described. In the present invention, the high-definition image data for one frame temporarily stored in the buffer memory of the image data processing unit 23, that is, the SXVGA full-screen video data is utilized, and the conventional wide angle and pan / tilt are performed with one camera. A monitoring image system capable of supporting the functions realized by two compatible telephoto cameras is realized by the following means.

  First, processing for detecting a moving object (suspicious person) from a captured image of a wide monitoring area will be described. The current SXVGA full screen video data temporarily stored in the buffer memory of the image data processing unit 23 is input to the moving body tracking processing unit 24. The moving body tracking processing unit 24 compares the current SXVGA full-screen video data (for example, FIG. 5) with the previous SXVGA full-screen video data (for example, FIG. 2) stored in the memory and detects the difference between them. Perform the difference. Alternatively, the current SXVGA full-screen video data (for example, FIG. 5) and the SXVGA full-screen video data with no moving object, or the background image generated by integrating the SXVGA full-screen video data up to the previous time in the time axis direction. A state change is detected by a background difference that compares (for example, FIG. 2) and detects the difference. Then, the state change groups are separated by a labeling (changed dot connection detection) method, and an image in which a state change group having a predetermined size or more is detected is set as a state change presence image, and “1” is set in the state change detection flag.

  In the moving body tracking processing unit 24, if “1” continues for a predetermined number of times or more in the state change detection flag detected at predetermined time intervals, the state change patterns are compared on a plurality of screens, and the movement (position change) exceeds a predetermined value. If so, “1” is set in the moving object detection flag of the current image data. Further, as shown in FIG. 5, based on the image data position and the gravity center information of the state change group detected as the moving body 50, the range including the state change image with respect to the current image data, that is, the detected movement A range including the body is set as a template block 80. The moving body tracking processing unit 24 includes template block setting means.

  The predetermined size for determining an image in which a state change group is detected by the labeling method as a state change image is a size that can eliminate a dog, a cat, a small bird, etc. when a subject to be tracked as a moving object is a human. Setting it can reduce over-detection. In addition, it is not always necessary to perform processing at a full rate of 30 frames / second for the predetermined time interval of state change or moving object detection, taking into account the time when the moving object to be captured is normally shot and the processing load of the CPU. What is necessary is just to set to an appropriate value within the range to about several hundred mSec period.

  Whether the detected moving object is a new moving object or a tracking moving object is determined by pattern matching of a template block described later, and if it is detected as a new moving object, the center of gravity or image of the set template block is detected. Information such as data position information and brightness values in the template block is stored with a label (symbol) attached to a buffer memory space provided in the moving object tracking processing unit 24. Here, the information in the buffer memory is used as identification information for independently tracking the mobile object for each template block, and a memory amount is secured for the mobile object extraction screen to be processed at one time.

  Next, means for tracking the detected mobile object will be described. The moving body tracking processing unit 24 processes the first full-screen video data and detects the current moving body within a predetermined area close to the center of gravity of each template block and the position of the image data. A search is made as to whether there are several types of template blocks which are assigned labels and stored in the buffer memory.

  If there is no existing template block in the predetermined area, a label is assigned as described above as a new moving object, and the template block related information is stored in the buffer memory. On the other hand, if there is an existing template block stored in the buffer memory in the predetermined area, the template block detected by processing the brightness value of the stored template block moving object and the current image data is processed. The luminance values of the body are sequentially compared from the existing template block having a close position by pattern matching including the image position and size. If there is similarity within a predetermined range as a result of the comparison, the template block information and the center of gravity of the template block detected by processing the current image data as the tracking block of the template block with the highest similarity among them and the image The data position information and the brightness value in the template block are updated.

  As described above, by repeating the process of the moving body tracking processing unit 24 at a predetermined time interval of state change detection, for example, the moving body 50 at the center of the upper end of the SXVGA full-screen video data as shown in FIG. In addition, a label “A” is given as the template block 80, and tracking can be performed while adjusting the block size on the locus of the template block. Further, the moving body 51 appearing at the right end of the middle stage of the screen can be tracked by giving a label “B” as the template block 81 of the second moving body. Note that the template block once detected as a moving object and given a label is subjected to pattern matching extraction of the presence or absence of a similar moving object at every predetermined time interval of state change detection. In addition, a similar pattern may not appear with a probabilistic period. If the template block is not updated without detecting the similar pattern continuously for a plurality of periods, all the template block information of the label is cleared from the buffer memory.

  Next, means for tracking and displaying a moving object according to the template block set by the moving object tracking processing unit 24 will be described. When the moving object is detected by the moving object tracking processing unit 24, the image data position and size from which the moving object is extracted, that is, the coordinate information of the template block, is set based on the center of gravity of the moving object and the image data position information. (23t). For example, when the moving object A shown in FIG. 4 is detected from high-definition SXVGA full-screen video data of horizontal: 1280 pixels and vertical: 960 lines used in this embodiment, the pixel map shown in FIG. Correspondingly, as the image data for tracking the moving body, the center is horizontal: 400 pixels, vertical: 90th line, the upper left diagonal start point is 240 pixels, vertical: 10th line, lower right diagonal end point An area having horizontal: 560 pixels and vertical: 170th line is set.

  It is desirable that the image data of the moving body tracking area, that is, the template block has a higher resolution, and the image data processing unit 23 extracts all the captured pixels in the area from the SXVGA full-screen video data without thinning out the image data compression unit. 25. The image data processing unit 23 includes template block extracting means. As a result, in the present embodiment, the template block is not thinned out for VGA full-screen video data that has been thinned from 1280 pixels × 960 lines to 1/2 in the vertical and horizontal directions. An image with a resolution having the above is sent from the surveillance camera 2.

  The image data compression unit 25 compresses the VGA full-screen video data and template block image data input from the image data processing unit 23 to about 1/20 by time-division video compression or still image compression such as motion JPEG. Encode. The LAN connection unit 26 time-division-multiplexes or selectively switches the compression-coded VGA full-screen video data and template block image data, and sends them to the monitoring terminal device 3 located outside. Note that the compression encoding of the VGA full-screen video data and the image data of the template block is time-division processed for each screen, and the compression rate is also set separately.

  The size of the template block set by the moving body tracking processing unit 24 is set to an integral multiple of the encoding unit of the image compression method to be used (for example, DCT block size in the case of the JPEG compression method), and is included in the action range of the suspicious person. If limited, efficient image compression can be realized, and the amount of information transmitted to the LAN can be further reduced. Further, when the moving body approaches the monitoring camera 2, an image is taken as a large subject, so that even if the pixels are thinned out, an image sufficient for specifying the moving body can be obtained. Therefore, when the template block exceeds a predetermined size, the amount of information transmitted to the LAN can be appropriately controlled by performing pixel thinning by the image data processing unit 23 and compression encoding by the image data compression unit 25. That is, the image data processing unit 23 may include template block thinning means for thinning template block data according to the data size of the template block, and the thinning may be increased as the data size of the template block increases.

  The monitoring terminal device 3 controls the compressed image data and the additional information from the monitoring camera 2 received from the LAN connection unit 31 via a transmission network such as a LAN by using a CPU (Central Processing Unit), a memory, and the like. The compressed image data of the template block is recorded in the built-in storage device 33 in association with event information such as a moving body detection number together with the received VGA full-screen video data. Further, the compressed data decompression unit 34 decodes the template block data compression-encoded by the surveillance camera 2 and the VGA full-screen video data, and sends them to the external display device 4 via the display driver 35. The display device 4 receives and displays the data processed by the monitoring terminal device 3. When the image data of the template block is smaller than a predetermined size for far-distance shooting, the image data of the template block is doubled by vertical / horizontal interpolation processing and enlarged and displayed. As a result, the tracking image of the moving object is zoomed 16 times by the amount without pixel decimation and interpolation enlargement.

(effect)
The surveillance camera system according to the present embodiment employs a high-definition camera even when monitoring a wide range with a single camera, thinning out full-screen video data, detecting a moving object, tracking a moving object, and By performing appropriate pixel decimation according to the shooting size of the moving object, the object is detected when the moving object is detected while transmitting the minimum necessary image data from the camera to the LAN network during normal shooting of only the background image. Only high-definition images can be captured and recorded and displayed. Therefore, a large-scale and expensive surveillance camera system using a plurality of cameras can be realized in a small size and at a low cost, and a high-accuracy image can be obtained for a plurality of moving objects, and the network network can be transmitted. There is an effect to reduce the load.

<Embodiment 2>
(Constitution)
The configuration of the surveillance camera system according to the present embodiment will be described with reference to FIG. The moving body tracking processing unit 24 in the monitoring camera 2 sets the tracking setting area in an arbitrary specific area of the SXVGA full-screen video data, so that the moving body passes through the tracking setting area or passes in a certain direction. It has a function of detecting the moving object only at the time. The monitoring terminal device 3 is connected to an input device (not shown) such as a mouse for setting a tracking setting area, and the set tracking setting area is notified to the monitoring camera 2 via the LAN. It has a possible configuration. Since other configurations are the same as those of the first embodiment, detailed description thereof is omitted here.

(Operation)
Next, the operation of the surveillance camera system according to the present embodiment will be described with reference to FIG. 1 showing the configuration of the surveillance camera system and FIG. 7 showing the images taken by the surveillance camera 2 attached in the bank. To do.

  First, using the input device connected to the monitoring terminal device 3 while viewing the entire screen on the display device 4, the tracking setting area is set with coordinates on the image data. For example, it is assumed that the area near the front entrance where only visitors enter as shown by the dot pattern in FIG. 7 is set as the tracking setting area 40. The coordinate information of the set tracking setting area 40 is notified to the monitoring camera 2 via the LAN, and as a template block provision condition of the moving object tracking processing unit 24, it is added that the moving object passes through this area from the entrance. The

  By adding the above conditions, the moving body tracking processing unit 24 only applies templates to moving bodies that pass through the tracking setting area 40 from the entrance, that is, moving bodies 60 to 63 that indicate visitors who enter and move from the general customer entrance. Blocks 90 to 93 are assigned to be tracked. Accordingly, the actions of the moving bodies 70 to 72 indicating the employees who have entered from the employee entrance 41 are not tracked because they do not pass the tracking setting area 40 even if they move within the shooting screen.

  Since other operations are the same as those in the first embodiment, a detailed description thereof is omitted here.

(effect)
The surveillance camera system 1 according to the present embodiment can obtain high-precision images for a plurality of moving objects while realizing the small size and low cost of the surveillance camera system as in the first embodiment. The transmission load of the network can be reduced. Furthermore, by providing the tracking setting area in a specific area in the shooting screen, it is possible to narrow down the tracking target objects and to improve the monitoring efficiency. That is, it is possible to reduce the processing load of the mobile body tracking processing unit 24 and the transmission load of the network.

1 is a configuration diagram of a surveillance camera system according to Embodiment 1. FIG. FIG. 3 is a diagram illustrating a captured image according to the first embodiment. 3 is a diagram illustrating a pixel arrangement of image data according to Embodiment 1. FIG. 3 is a diagram illustrating a pixel arrangement of image data according to Embodiment 1. FIG. 6 is a diagram for explaining an operation of a moving body tracking processing unit according to Embodiment 1. FIG. 6 is a diagram for explaining an operation of a moving body tracking processing unit according to Embodiment 1. FIG. FIG. 6 is a diagram illustrating a captured image according to the second embodiment.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 Surveillance camera system, 2 Surveillance camera, 3 Surveillance terminal device, 4 Display apparatus, 21 Lens unit, 22 Image sensor, 23 Image data processing part, 24 Moving body tracking process part, 25 Image data compression part, 26 LAN connection part, 27 CPU, 31 LAN connection unit, 32 control unit, 33 storage device, 34 compressed data decompression unit, 35 display driver, 40 tracking setting area, 41 employee entrance, 50, 51, 60 to 63, 70 to 72 mobile unit, 80, 81, 90-93 Template block.

Claims (7)

Image data generation means for imaging an imaging target and generating first full-screen video data;
Image data conversion means for converting the first full-screen video data into second full-screen video data having a predetermined pixel size;
Moving body detecting means for detecting a moving body from the first full-screen video data;
Template block setting means for setting a template block on the mobile body;
Template block extracting means for extracting the template block as image data of a predetermined size from the first full-screen video data;
Image data sending means for sending image data of the template block and the second full-screen video data to the outside by time-division multiplexing or selectively switching;
Means for tracking the moving body;
A template block-related information including a center of gravity of the template block and a position of image data, and a brightness value in the template block, and a memory in which a label is attached and stored ,
The means for tracking the moving body is:
When a moving body is detected by the moving body detection means, a search is made as to whether or not a template block stored in the memory exists within a predetermined area,
When there is no template block stored in the memory in the predetermined area, a label is attached to the template block related information of the template block set in the moving body detected by the moving body detecting means. Save it in the memory,
When there is a template block stored in the memory in the predetermined area, the luminance value of the moving body of the template block stored in the memory and the luminance of the moving body detected by the moving body detecting means If there is similarity within a predetermined range, the template block related information of the template block having the highest similarity is set in the moving body detected by the moving body detecting means. you of updates to the template block-related information,
Surveillance camera. Template block thinning means for thinning out the data of the template block according to the data size of the template block;
The surveillance camera according to claim 1. The template block decimation means increases decimation as the data size of the template block increases.
The surveillance camera according to claim 2. The moving body detecting means sets a tracking setting area in an arbitrary specific area of the first full-screen video data so that the moving body passes through the tracking setting area or only in a certain direction. Detecting the moving object;
The surveillance camera according to any one of claims 1 to 3. The number of imaging pixels of the first full-screen video data is 1 M pixels or more, and the number of imaging pixels of the second full-screen video data is VGA (630 × 480 pixels) or less.
The surveillance camera according to any one of claims 1 to 4. Image data compression means for compressing and encoding the template block data and the second full-screen video data;
The image data sending means sends the compression-coded template block data and the second full-screen video data to the outside.
The surveillance camera according to any one of claims 1 to 5. A surveillance camera according to claim 6;
A monitoring terminal device that receives and decodes the compression-encoded data of the template block and the second full-screen video data, and sends them to the outside;
A display device that receives and displays data processed by the monitoring terminal device,
Surveillance camera system.

Images