JP2007133470A - Monitoring terminal device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a monitoring terminal device that prevents undetection and excessive detection by detecting detection targets correctly even when the conditions of image data from a monitoring camera change from time to time with changing natural phenomena such as sunshine and shade. <P>SOLUTION: When single fields of image signals imaging a monitored area at predetermined time intervals are compared, the image signals are divided into a plurality of blocks and luminance is smoothed correspondingly to provide contrasts between the blocks. A threshold used in the image comparison can take a plurality of values independent of changing sunshine conditions. Such a value is set in each block to ensure discrimination between an intruder image and a background image. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a monitoring terminal device that determines the entry of an intruder or the like into a monitoring area by analyzing the movement of image data sent from, for example, a monitoring camera.

  A conventional monitoring terminal device stores input image data sent from a monitoring camera or the like at an arbitrary timing for confirmation of an intruder or the like as a non-delayed signal, and further, one field or one frame. Store the delayed signal delayed for a period, obtain a difference value between the delayed signal and the non-delayed signal, and if the difference value is equal to or greater than a preset threshold value, determine that an intruder has entered, A start signal is output so that the video deck device can be recorded for a certain period of time or an alarm can be output (for example, see Patent Document 1).

JP-A-2-2486

  The image data sent from a surveillance camera installed at an arbitrary location may change the state of the sun and shade due to the movement of the sun and clouds, etc., or the luminance state may change due to blinking of night illumination. The state of the image (contrast, etc.) is greatly changed. Therefore, in the conventional monitoring terminal device, the difference value between the non-delayed signal and the delayed signal obtained from the image data may fluctuate more than a preset threshold value (luminance difference threshold value), which is a detection target. There were problems such as oversight of intruders and excessive detection.

  The present invention solves the above-mentioned problems, for example, correctly detects the detection target even in a state where the state of the image data from the surveillance camera sequentially changes with changes in natural phenomena such as the sun and shade, An object of the present invention is to provide a monitoring terminal device that does not cause oversight or overdetection.

  The monitoring terminal device of the present invention is a device that determines whether or not a new object has entered the monitoring area because the image signal obtained by imaging the monitoring area is compared at each predetermined first time interval. Means for dividing an image signal for one field or one frame into block signals having a plurality of ranges, means for smoothing the image signal in each block signal to obtain each block signal value, Means for integrating the block signal values at a second time interval to obtain a block signal integral value; means for obtaining an average value (total integral average value) of the divided block integral values; Means for obtaining a difference (signal deviation) between the average value and each block integration value, and whether or not a new object has entered the monitoring area using one or both of the integration average value and the signal deviation. Each threshold is judged And wherein points having a means for obtaining for each range.

  According to the monitoring terminal device of the present invention configured as described above, since the integration process is performed for each divided range using each block signal value and image signal value, instantaneous luminance by an intruder or the like The block luminance signal integrated value that responds sensitively to relatively slow luminance signal levels such as the sun and shade is obtained with almost no response to changes, resulting in intrusion without being affected by natural sunlight fluctuations. The effect that the detection of the person or the like can be accurately performed is obtained.

Embodiment 1 FIG.
Hereinafter, Embodiment 1 of the present invention will be described in detail with reference to the drawings. FIG. 1 is a schematic diagram of a camera monitoring system using the monitoring terminal device of the present invention, and FIG. 2 is a conventional camera monitoring system in which the sunshine and shade conditions at the time of shooting are the luminance difference between the background of the shooting screen and the detection object ( FIG. 3 is a block diagram showing a schematic configuration of the monitoring terminal device according to the present invention for determining the entry of the detection object, and FIG. 4 is performed in the monitoring terminal device according to the present invention. It is a schematic diagram which shows an image processing state. 1 to 4 are assigned the same reference numerals.

  As shown in FIG. 1, the camera monitoring system 1 includes a monitoring camera 2 and a monitoring terminal device 3 of the present invention. The monitoring camera 2 captures an arbitrary monitoring area as a target, and provides the captured image data to the monitoring terminal device 3 by wire or wirelessly. Based on the image data from the monitoring camera 2, the monitoring terminal device 3 determines that a suspicious person or the like has entered the monitoring area (the presence of a new object). The monitoring terminal device 3 also includes a communication line 31 such as a LAN for externally reporting that a suspicious person or the like has entered the monitoring area as necessary, and image data relating to the recorded suspicious person or the like. Alternatively, an alarm output terminal 32 is provided for informing a suspicious person.

  By the way, the existing surveillance camera used for a general camera surveillance system has the following characteristics. Here, the influence of the sunshine and shade conditions at the time of shooting on the luminance difference (contrast) between the background screen of the screen shot by the surveillance camera and the screen of the object to be detected such as a suspicious person will be organized. It is necessary for a normal surveillance camera to cover a wide range of illuminances ranging from about 1 to 2 looks of candle light to the seaside and so on that reach 1000 looks or more, and obtain a clear photographed image. For this reason, in order to optimize the dynamic range of an imaging element (for example, a CCD array element) that determines the imaging state of the imaging area of the surveillance camera, an AGC (Auto Gain Control) for the auto iris and the imaging element is mounted. Designed to automatically adjust to the correct brightness. The procedure for setting the photographing conditions can be started by adjusting the light amount with the auto iris of the surveillance camera. In addition, if the light amount is adjusted only with auto iris, the screen may be too dark, or on the contrary, overexposure may occur. Is adjusted to the brightness. However, AGC that is generally widely used is designed so that, for example, the average value of the luminance signal of the element group of the array element, which is a photographing element, is constant, so that the entire screen is dark, bright, or shaded. When there is a large contrast in the sun, the contrast between the background screen and the detection object screen is greatly different. This will be further described with reference to the conceptual diagram of FIG.

  FIG. 2 shows a state in which a black box 41 and a white box 42, which are detection objects, are arranged and photographed on a uniform background screen (with no irregularities and a monochrome background other than white and black) using a surveillance camera. First, a thought experiment is performed in the case where a uniform strong light or a weak light is irradiated to the photographing space with a uniform illuminance. The AGC adjusts the gain to a constant brightness so that the luminance signal value becomes an average value for almost the entire field. For this reason, as shown in FIG. 2A, the photographing screen displays a uniform background screen having substantially the same luminance signal level in both cases of strong light irradiation and weak light irradiation. Therefore, the luminance difference between the black box and the background screen is a substantially constant value regardless of the portion of the screen. Similarly, the luminance difference between the white box and the background screen is a substantially constant value regardless of the portion of the screen.

  Next, a thought experiment is performed in the case where strong light is irradiated to the half of the shooting space and weak light is irradiated to the other half. The AGC adjusts the gain so that the average value of the luminance signal values becomes a constant brightness for almost the entire field. For this reason, as shown in FIG. 2B, the photographing screen becomes overexposed without being suppressed in the luminance signal level in the portion irradiated with strong light, and conversely the luminance signal level is sufficient in the portion irradiated with weak light. Since it cannot be lifted, a dark background screen is copied. As a result, the background screen has two environments, a portion irradiated with strong light and a portion irradiated with weak light. Therefore, the black box in the portion irradiated with the weak light has a feature that the luminance difference from the background screen is reduced, and the white box in the portion irradiated with the strong light has a feature that the luminance difference from the background screen is reduced. Since this point will be described more specifically, problems of the conventional camera monitoring system will be described.

  Under the condition that there are black box and white box in the presence of two environments of Hinata 43 and Shade 44, the brightness difference between the shaded background screen and the black box becomes small as shown in Fig. 2-2. And the brightness difference becomes larger. Also, the luminance difference between the background screen in the sun and the white box is small, and conversely, the luminance difference from the black box is large. The change of the sun position that moves with time is accompanied by a large change in the position of the sun 43 and the shade 44 and the luminance state, and greatly changes the luminance difference between the black box and the white box with respect to the background screen. Therefore, in the “method for obtaining the luminance difference between the detection object and the background screen and comparing the luminance difference with a preset luminance difference threshold value (fixed value)” employed in the conventional camera monitoring system, the black It can be seen that boxes and white boxes cannot always be confirmed (identified). Therefore, it is considered that the difference in luminance signal level generated between the sunflower 43 and the shade 44 needs to be appropriately set / updated according to the background where the luminance changes as in the present invention to be clarified. .

  As shown in FIG. 3, the monitoring terminal device 3 according to the first embodiment of the present invention includes an AD converter 301, a buffer memory 302, a block data smoothing unit 303, a block data integration unit 304, and a luminance difference threshold setting unit. 305, a state change detection unit 306, a motion detection control unit 307, an image data recording processing unit 308, a notification processing unit 309, and an information source encoding unit 310. The operation of each part will be described below.

  The A / D converter 301 receives the image data d taken by the surveillance camera 2, and luminance data (Y) and color difference data (Cb, Cr) included in the received image data. For example, a component signal such as 13.5 MHz or 6.75 MHz is sampled, so that it is converted into 8-bit digital luminance data dy and digital color difference data dcb, dcr.

  The buffer memory 302 converts the digital luminance data dy converted by the A / D converter 301 into 1 image signal of 240 lines (480 lines: when interlace scanning is not performed) having 720 dots per line, for example. This is assigned to the luminance signal sy which is a form, and is temporarily stored as one field (one frame). Also, the digital color difference data dcb and dcr are assigned to 240 lines of color difference signals scb and scr having pixels of 360 dots per line, for example, and temporarily assigned as one field (one frame). store. In the following description, the image processing in the field using the luminance signal sy will be described without particular notice. However, the same handling is performed in the case of using the color difference signals scb and scr and the image processing in the frame. Is possible.

  The block data smoothing unit 303 divides the luminance signal sy for one field stored in the buffer memory 302 into a plurality of ranges with 16 dots × 16 lines (256 pixel range) as one block unit. Thus, the block signals in the respective ranges can be reproduced. That is, the luminance signal sy for one field is divided into 45 × 15 (total 675) blocks. Further, the block data smoothing unit 303 obtains, for example, a block luminance signal value nb (a block signal value obtained from the luminance) corresponding to the average value of the luminance of 256 pixel groups possessed by each block from the luminance signal sy. Obtained for each block and temporarily stored as a list Nb of block luminance signal values. Note that although the luminance signal for one field is divided in units of one block here, it is not always necessary to perform it in units of one block, and it is only necessary to be able to grasp the range after being divided appropriately.

  The block data integration unit 304 is a functional unit that integrates the block luminance signal value list Nb obtained by the block data smoothing unit 303 in the time axis direction to extract a background screen. Perform integration processing.

  The block data integration unit 304 includes, for example, a 16-bit register, and each block luminance signal value nb obtained at an appropriate sampling interval w2 corresponding to the second time interval and the 16-bit register value nbs. The difference value ns is obtained, multiplied by a preset coefficient k, and then returned to the 16-bit register. That is, the block data integration unit 304 calculates a value obtained by multiplying each difference value ns corresponding to the change value (nb−nbs) of the luminance signal level obtained at an appropriate sampling interval w2 by a preset coefficient k. Then, this is added to the block luminance signal integrated value nbs of the register and integrated to obtain the block luminance signal integrated value nbs having a relatively slow time response. The block luminance signal integration value nbs is calculated in each of a total of 675 blocks forming one field, and the list is corrected at an appropriate sampling interval w2 and stored as a block luminance signal integration value. As the preset coefficient k, a coefficient value that does not overflow the digit of the 16-bit register used in the integrator may be selected. For example, an appropriate sampling interval w2 corresponding to the second time interval is set to 0. In the case of 5 seconds, it may be set by appropriately selecting from the range of about 0.05 to 0.25. As a result, the block luminance signal integrated value nbs which responds to relatively slow luminance signal level changes such as the sun and the shade without almost responding to the instantaneous luminance change caused by an intruder or the like is obtained, resulting in 675. From the list Nbs of each block luminance signal integrated value composed of blocks, it is possible to generate a background image map in which the luminance difference between the sun and the shade is accurately extracted without responding to the instantaneous luminance change by the intruder. There is.

  Note that the appropriate sampling interval w2 corresponding to the predetermined second time interval necessary for the integration process may be a time interval considering the speed of movement of the intruder or the like, and more preferably the background state. It is sufficient that the time interval is as long as possible within a range in which the change can be followed. For example, the sampling interval w2 when the width of the monitoring area of the monitoring camera 2 is about 10 m and the intruder is a person will be estimated. Since the speed at which a person can act is 10 m / sec or less per second, the time interval required for detecting the intruder state change can be estimated to be several hundred msec or more. Therefore, for example, by measuring the level difference of the luminance signal between an image obtained by delaying one field by a sampling interval w1 (several hundreds msec time interval) and a non-delayed image, an intruder that is an object to be detected in the photographing screen. It can be seen that the presence (entrance) or the like can be detected (sensing). Therefore, the sampling interval w2 may be about 0.1 seconds, or may be several seconds longer than that.

  The luminance difference threshold value setting unit 305 refers to the list Ns of block luminance signal integrated values stored in the block data integrating unit 304, and performs smoothing, for example, an average value of block luminance signal integrated values nbs in all blocks (for example, The total integrated average luminance signal value (nbsav) is obtained and temporarily stored. Further, a signal deviation nac which is an absolute value of a difference value between the total integral average luminance signal value nbsav and each block luminance signal integrated value nbs is obtained, and a list Nac of the signal deviations is temporarily stored. A register corresponding to all 675 (45 × 15) blocks for setting a luminance difference threshold value nh weighted based on a predetermined condition corresponding to the total integrated average luminance signal value and the list Nac of the signal deviations. Is provided. For example, at a certain arbitrary time, a difference value (signal deviation) between each block luminance signal integrated value ns and the total integrated average luminance signal value nbsav stored in the block data integration unit 304 for comparison or each block luminance. The ratio of the total integrated average luminance signal value nbsav to the signal integrated value nbs is obtained (the second comparison value is obtained), and the difference value or ratio (corresponding to the second comparison value) as the comparison result is determined in advance. If the value is smaller than the predetermined value (separation determination value), a value obtained by correcting the first luminance difference threshold value nh1 set in advance by the coefficient a determined by the total integral average luminance signal value nbsav is selected and set in the register of the luminance difference threshold value setting unit 305. When the difference value or the ratio is greater than a predetermined value, the total integrated average luminance signal value nbsav is set to the second luminance difference threshold value nh2 that is smaller than the first luminance difference threshold value nh1. The value corrected by the determined coefficient a select set in the register. Thereby, a plurality of preset brightness difference threshold values can be associated with each divided range, and a list Nh of brightness difference threshold values as a result is temporarily stored. In addition, in response to the case where the difference value or ratio becomes a large value stepwise without providing the predetermined value, a plurality of luminance difference threshold values nh are set in the register so as to become small values stepwise. Alternatively, the following processing may be performed. Since the difference value or ratio (second comparison value) obtained for each block is referred to, the luminance difference threshold value nh corresponding to each block may be selected and assigned. In this case, a special effect that a more detailed luminance difference threshold nh can be appropriately selected and set for each block is obtained. The coefficient a determined by the total integral average luminance signal value nbsav corrects the detection value sensitivity for darkness in which AGC is saturated or white stripe image in which AGC is not effective at all, thereby improving detection accuracy. It may be provided according to the necessity from the monitoring camera installation conditions.

  The state change detection unit 306 includes a list Nbb of non-delayed block luminance signal values obtained by the block data smoothing unit 303, and an appropriate sampling interval w1 corresponding to the first time interval to be compared (for example, an intruder or the like). Therefore, there is provided a means for temporarily storing a list Nba of delayed block luminance signal values obtained by the block data smoothing unit 303 after a predetermined time interval). The state change detection unit 306 also compares the non-delayed block luminance signal value list Nbb constituting the non-delayed block luminance signal value list Nbb and the block luminance constituting the delayed block luminance signal value list Nba for comparison. Each difference value or each ratio is obtained for each block from the signal value nba (a third comparison value is obtained). Further, the difference value or the ratio is compared with the luminance difference threshold value nh (stored in the luminance difference threshold value list Nh) of each block set in the register of the luminance difference threshold value setting unit 305. It is determined that there has been a change in state, and *, which is a detected symbol 14, is set. In other words, a status field image made up of 1 field and 675 blocks as shown in FIG. 4 (3) in which the symbol 14 is set to the change block having the status change can be output. This makes it possible to determine the presence or absence of a state change caused by an intruder or the like with respect to the monitoring target area displayed by the surveillance camera 2 and to express the result by the presence or absence of a symbol * as necessary. Therefore, it can be understood that the approximate size and shape of the intruder 13 can be estimated, and the entry determination can be made easier. In the following, the state of the entry determination of the intruder 13 will be further described with reference to the schematic diagram 4.

  FIG. 4 shows a 1-field shooting screen composed of 675 (45 × 15) blocks with 16 dots × 16 lines (256 pixels) reproduced based on the image data sent from the surveillance camera 2 as one block unit. This is shown schematically. In FIG. 4A, the tree 11 and the house 12 are shown on the photographing screen of the image data A sent from the surveillance camera 2 at a certain arbitrary time. FIG. 4 (2) shows the image data B sent from the monitoring camera 2 after elapse of, for example, a sampling interval w1 (predetermined first time interval) from an arbitrary time in FIG. 4 (1). In addition to the tree 11 and the house 12, an intruder 13 is shown on the shooting screen. FIG. 4 (3) shows a difference value for each block from a list Nbb of non-delayed block luminance signal values obtained from the image data A and a list Nba of delayed block luminance signal values obtained from the image data B. Since the ratio is obtained and this is compared with the luminance difference threshold value nh set in the register of the luminance difference threshold value setting unit 305, the symbol * is set for the change block in which the state has changed, and this is displayed on the screen. It is the displayed state field picture. From the arrangement of the symbols * caused by the intruder 13, it can be seen that the approximate size and approximate shape of the intruder 13 can be estimated.

  The motion detection control unit 307 compares the number of detection determination thresholds nx set in advance with the number of changed blocks (number of symbols *), and when the number of changed blocks is larger than a preset value, the motion detection control unit 307 Determine that there is an intrusion. In addition, the motion detection control unit 307 immediately notifies the image data recording processing unit 308 and the notification processing unit 309 that an intruder or the like has entered using a detected signal. The detection determination threshold number nx is a parameter determined by the angle of view of the monitoring camera 2 and the distance at which the detection target is captured. The number of blocks detected by capturing an image of an intruder in advance. You can check the settings and use them as a guide.

  The image data recording processing unit 308 converts, for example, compressed image data encoded by the information source encoding unit 310 from the image data temporarily stored in the buffer memory 302 into the motion detection control unit 307. In response to the notification of the detection presence signal issued from, recording can be started on a storage medium such as an HDD or a semiconductor memory.

  The notification processing unit 309 receives the notification of the detection presence signal from the motion detection control unit 307, and a speaker or the like (not shown) connected to the output terminal 32 with an alarm that warns an intruder from the monitoring terminal device 3. ), Or a mobile phone (not shown) or the like via a communication line such as the LAN 31.

  Since the processing as described above is performed, the block data smoothing unit 303 has a function of canceling a luminance change caused by scattering that closes within a range, such as a seaside wave or a tree leaf shake. effective. Further, by performing the process of the luminance difference threshold setting unit 305 with respect to the integration process over time in the block data integration unit 304, chattering that occurs due to repetitive luminance changes such as seaside waves and tree leaf shakes. There is an effect to avoid. Furthermore, the state change detection unit 306 has an effect that it is possible to estimate the approximate size and shape of the intruder while improving the excessive motion detection other than the detection target, and to improve the oversight of the intruder.

  In the processing of the block data integration unit 304 and the luminance difference threshold value setting unit 305, when a state change is detected for each block segmented by the block data smoothing unit 303, it corresponds to the luminance change of the background screen. Thus, since the brightness difference threshold value nh can be set, it is possible to accurately detect an intruder and reduce detection misses.

  In the first embodiment of the present invention, the input image from the monitoring camera 2 is used as a component signal. However, if a composite signal such as NTSC or PAL is input to the monitoring terminal device 3 and decoded, the same form is obtained. It can be input to the AD converter 301. Further, the AD converter 301 converts the luminance signal sy into 13.5 MHz and the color difference signal sc into 8-bit digital data sampled at 6.75 MHz. These are also converted into the performance of the monitoring camera 2 and the monitoring system. Different parameters may be set depending on the intended use. Further, the image data received by the A / D converter 301 does not necessarily have to be received directly from the specific monitoring camera 2, and is an image obtained by photographing another arbitrary monitoring area. It goes without saying that the same processing can be performed even with data.

  Further, the block data smoothing unit 303 according to the first embodiment divides the luminance data dy for one field of the buffer memory 302 into a range of 45 × 15 with 16 dots × 16 lines as one block. However, an appropriate value may be set as appropriate according to the monitoring area and the speed and size of the detection target. Further, the block data smoothing unit 303 and the block data integration unit 304 process only the luminance signal sy as image data, and the state change detection unit 306 has a signal difference between the delayed and non-delayed luminance signals sy. Although the presence / absence of the state change is detected from the above, the same effect can be obtained even if the determination is made with the comprehensive difference value including the color difference (Cb, Cr) signal.

  Further, as a means for processing each part of the present invention, a block data smoothing part that requires high-speed signal processing is processed by hardware such as PLD (Program Logic Device), and necessary operations for various judgments thereafter. An efficient monitoring terminal device can be obtained by processing up to the detection control unit with firmware of a CPU (Central Processing Unit).

It is the schematic of the camera monitoring system using the monitoring terminal device of this invention. It is a schematic diagram which shows the contrast of the imaging | photography screen of the conventional camera monitoring system. It is a block diagram which shows schematic structure of the monitoring terminal device of this invention. It is a schematic diagram which shows the image processing state of the monitoring terminal device of this invention.

Explanation of symbols

1 camera monitoring system, 2 monitoring camera, 3 monitoring terminal device, 11 tree, 12 house, 13 intruder, 14 symbol *, 31 output terminal, 32 LAN,
41 Black box, 42 White box, 43 Hinata, 44 Shade 301 AD converter, 302 Buffer memory,
303 block data smoothing unit, 304 block data integration unit,
305 brightness difference threshold setting unit, 306 state change detection unit, 307 motion detection control unit,
308 Image data recording processing unit, 309 Notification processing unit, 310 Information source encoding unit

Claims (7)

  In the monitoring terminal device for determining whether or not a new object has entered the monitoring area by comparing the image signal obtained by imaging the monitoring area at each predetermined first time interval, one field or one Means for dividing an image signal for a frame into block signals having a plurality of ranges; means for smoothing the image signal in each block signal to obtain each block signal value; and Means for obtaining a block signal integrated value by integrating at time intervals, means for obtaining an average value of each of the divided block integrated values, and obtaining a signal deviation which is a difference between the average value and each of the block integrated values Means and means for obtaining, for each divided range, a threshold value for determining whether or not a new object has entered the monitoring area using one or both of the average value and the signal deviation. Features a point That monitoring terminal equipment.   The second time interval for integrating each block signal value of the monitoring terminal device according to claim 1 is processed at the same or the same as or more than the first time interval for comparing the image signal at predetermined time intervals. The monitoring terminal device according to claim 1, further comprising:   Each block signal value delayed by the first time interval and each non-delayed block signal value are compared for each divided range to obtain each third comparison value. 2. The monitoring according to claim 1, further comprising means for determining whether or not there is a state change of each delayed block signal value by comparing with a threshold value selected and determined for each of the divided ranges. Terminal device.   It is determined whether or not a new object has entered by comparing the means for obtaining the number of blocks of the block signal value having a state change with the number of blocks and a preset detection determination threshold number. The monitoring terminal device according to claim 3, further comprising: means.   5. The monitoring terminal device according to claim 1, wherein the image signal is a luminance signal or a color difference signal.   The magnitudes of the plurality of threshold values set in advance are smaller as the absolute value difference between each block signal value and the corrected total smoothed signal value or the ratio of each block signal value divided by the corrected total smoothed signal value increases. The monitoring terminal device according to claim 1, wherein the monitoring terminal device is set as follows. Means that digital image data is encoded as information source to form compressed image data, image data recording means for controlling the compressed image data to be recorded on a recording medium, and input image has moved. 7. The monitoring terminal device according to claim 1, further comprising at least one of reporting means for outputting the information to the outside.

Images