JP6091319B2 - Image processing device for monitoring - Google Patents

Description

  The present invention relates to a monitoring image processing apparatus that monitors the appearance of a notification event such as an intruder.

  In a conventional image processing apparatus, for example, as disclosed in Patent Document 1, a change area is extracted using a “brightness background difference” method, and the change area is within a monitoring range and has a predetermined area. Or an intruder is determined based on conditions such as a predetermined speed. Also, in order to prevent erroneous intruder determination (false report), whether or not a disturbance has occurred due to the fact that the area of the change area is a specific area, the dimensions of the circumscribed rectangle of the change area, etc. If it is determined and it is determined that a disturbance has occurred, the reporting process is restricted.

Japanese Patent Laid-Open No. 9-214939

Since the conventional monitoring image processing apparatus is configured as described above, it is possible to efficiently determine moving objects and reduce unnecessary detection.
However, in the conventional monitoring image processing apparatus, particularly when performing monitoring work outdoors, there is a problem that false alarms are generated due to unnecessary detection due to the influence of a change area caused by the movement of a headlight of a car at night. there were.
Headlights illuminate a wide area with high brightness, creating a clear change area. In addition, the size is indefinite, and the shape may be misidentified as an intruder, and the conventional monitoring image processing device is not determined to be a disturbance and is subject to notification. However, in the human eye, the “intruder” and the “headlight” are clearly different, and due to this gap, the headlight alert is regarded as an unnecessary notification and a false alarm.

  The present invention has been made to solve the above-described problems, and it is determined that the change area caused by the headlight, which has been difficult to avoid false alarms, is effectively a disturbance, and unnecessary detection is corrected internally before the alarm is issued. It is an object of the present invention to obtain a monitoring image processing apparatus that can cope with the situation so as not to cause an alarm.

A monitoring image processing apparatus according to the present invention includes a video input unit that receives a current monitoring image of a monitoring target region as a video signal, and A that quantizes the video signal received by the video input unit to obtain current image data A A notification event appears from the / D converter, the current image data calculation storage unit that stores the current image data quantized by the A / D converter, and the current image data quantized by the A / D converter. A comparison image data calculation / storage unit for creating and storing comparison image data indicating the monitoring image that is not in a state, and comparing the current image data with the comparison image data, and a change area in the monitoring image A change region extraction unit that extracts the change region, a recognition processing unit that analyzes the change region to determine whether a notification event has occurred, and outputs a notification event occurrence notification, and the change A disturbance determination unit that determines whether a disturbance has occurred in the area and outputs a disturbance occurrence notification, the notification event occurrence notification output by the recognition processing unit, and the disturbance occurrence notification output by the disturbance determination unit And a reporting unit that performs a reporting process, wherein the disturbance determination unit acquires a circumscribed rectangle of the change area extracted by the change area extraction unit, and the ferret diameter check A block dividing unit that divides the circumscribed rectangle acquired by the unit into a plurality of blocks, a block gray measuring unit that calculates a gray value for each of the plurality of blocks divided by the block dividing unit, and a block gray measuring unit that calculates On the basis of the gray value for each block, a gray vertex block that is a block in which there is no block having a gray value larger than itself in an adjacent surrounding block is searched, and the gray vertex block is detected. A light source position calculation unit that determines a light source as a light source, a road area that is preliminarily stored as a road on the monitoring image, a road area verification unit that determines whether the light source belongs to the road area, and the road area For the light source determined by the matching unit to belong to the road area, the time during which the shade vertex block continues to exist in the road area, the moving direction range in which the shade vertex block has moved, the density value of the shade vertex block, And a headlight determination unit that determines whether the shade vertex block is due to a headlight based on a speed range in which the road block area moves and a generation time of the light source .

  According to the present invention, the change area caused by the headlight, which is difficult to avoid false alarms, is effectively determined as a disturbance, and unnecessary detection is performed at the same scale as the conventional apparatus, at a low cost and without greatly reducing the processing speed. It is possible to take measures so that it does not lead to a report by correcting it internally before the report.

1 is a configuration diagram of a monitoring image processing apparatus according to Embodiment 1 of the present invention. FIG. It is a block diagram of the disturbance determination part 7 which concerns on Embodiment 1 of this invention. It is a flowchart of the alerting | reporting process of the monitoring image processing apparatus which concerns on Embodiment 1 of this invention. It is a flowchart explaining the disturbance determination in the disturbance determination part which concerns on Embodiment 1 of this invention. It is a figure which shows the example of the present image accumulate | stored in the present image data calculation storage part, and the example of the comparison image (background image) accumulate | stored by the image data calculation storage part for a comparison. It is a figure which shows the example of the change area extracted by the change area extraction part. It is a figure explaining a ferret diameter. It is a figure which shows the example which arrange | positioned the small block of "8 pixels x 8 pixels" densely in the range where a ferret diameter is hidden so that a ferret diameter may be included. It is a figure which shows the result in which the block shade measuring part calculated the shade value of all the blocks. It is a figure which shows the example of a light / dark vertex block. It is a figure which shows the road area set beforehand. It is a block diagram of the image processing apparatus for monitoring which concerns on Embodiment 2 of this invention. It is a flowchart of the alerting | reporting process of the monitoring image processing apparatus which concerns on Embodiment 2 of this invention. It is an example of the monitoring image of the situation where a headlight and a pedestrian (intruder) were mixed. It is the figure which piled up the ferret diameter on each of the headlight shown in FIG. 14, and a pedestrian. It is the figure which marked the light-dark vertex of each ferret diameter shown in FIG.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
Embodiment 1 FIG.
FIG. 1 is a block diagram of a monitoring image processing apparatus according to Embodiment 1 of the present invention.
The monitoring image processing apparatus includes a video input unit 1, an A / D converter 2, a current image data calculation storage unit 3, a comparison image data calculation storage unit 4, a change area extraction unit 5, and a recognition processing unit 6. A disturbance determination unit 7 and a reporting unit 8.

The video input unit 1 receives a current monitoring image of a monitoring target area output from an imaging device (not shown) such as a TV camera installed in an electric station as a video signal.
The A / D converter 2 quantizes the video signal received from the video input unit 1. That is, the analog data received from the video input unit 1 is converted into digital data.
The current image data calculation accumulation unit 3 receives and accumulates current image data (referred to as current image data) quantized by the A / D converter 2. The current image data may be recorded as necessary, and the received current image data may be output to the change area extraction unit 5 as it is without accumulating the current image data.

  The comparison image data calculation accumulating unit 4 receives the output from the A / D converter 2 and performs a corresponding calculation, and shows a comparison image indicating a monitoring target area in a state where no alarm event (for example, an intruder) has appeared. Data (for example, image data acquired by the A / D converter 2 at a timing when a notification event does not appear) is accumulated. The comparison image data may be recorded as necessary. The comparison image data is not accumulated, the comparison image data is created from the current image data received from the A / D converter 2, and the change area is extracted as it is. The image data may be output to the unit 5, or the comparison image data created artificially may be acquired from the outside and output to the change region extraction unit 5 as it is.

The change area extraction unit 5 compares the current image data acquired from the current image data calculation storage unit 3 with the comparison image data acquired from the comparison image data calculation storage unit 4, and changes in the monitoring target region. To extract the change area. Generally, a binarization processing unit is built in here, and the output is converted into binary data.
The recognition processing unit 6 analyzes the change area detected by the change area extraction unit 105 and performs a process of determining whether or not a notification event has appeared. If it is determined that a reporting event has occurred, a notification event occurrence notification is output to the reporting unit 8. Note that the report target is, for example, a person, and is determined in advance by the operator.
The disturbance determination unit 7 receives the change region extracted by the change region extraction unit 5 and determines whether a disturbance has occurred. When it is determined that a disturbance has occurred, a disturbance occurrence notification is output to the reporting unit 8.

The reporting unit 8 receives instructions from the recognition processing unit 6 and the disturbance determination unit 7 and performs reporting processing.
When it is determined by the disturbance determination unit 7 that no disturbance has occurred, if it is determined by the recognition processing unit 6 that a notification event has appeared, a notification that a notification event has appeared is issued. Perform the process. In addition, when it is determined by the disturbance determination unit 7 that a disturbance has occurred, even if it is determined by the recognition processing unit 6 that a notification event has appeared, a notification that a notification event has appeared is issued. Regulate the processing to be reported.
In general, the recognition processing unit 6, the disturbance determination unit 7, and the reporting unit 8 are processed by S / W.

FIG. 2 is a configuration diagram of the disturbance determination unit 7 according to Embodiment 1 of the present invention.
The disturbance determination unit 7 includes a ferret diameter confirmation unit 20, a block division unit 21, a block density measurement unit 22, a light source position calculation unit 23, a road area verification unit 24, and a headlight determination unit 25.
The ferret diameter confirmation unit 20 receives the change region extracted by the change region extraction unit 5, and acquires the ferret diameter (the circumscribed rectangle of the change region).
The block division unit 21 divides the ferret diameter output from the ferret diameter confirmation unit 20 into a plurality of small blocks.

The block density measuring unit 22 calculates the average density value for each block divided by the block dividing unit 21.
The light source position calculation unit 23 identifies the block most likely to be a light source from the average of the gray value of each block calculated by the block gray level measurement unit 22.
The road area matching unit 24 determines whether or not the most likely light source block calculated by the light source position calculation unit 23 exists on the road.
The headlight determination unit 25 receives the determination from the road area collation unit 24, observes the time history of the most likely light source block on the road, and determines whether it is a headlight.

Next, the operation will be described.
FIG. 3 is a flowchart of the reporting process of the monitoring image processing apparatus according to the first embodiment of the present invention. First, the operation up to issue will be described with reference to FIG.
The video input unit 1 receives a video signal (video signal) output from an external imaging device such as a TV camera installed in an electric station or the like, and sends it to the A / D converter 2 (step ST1).
The A / D converter 2 converts the video signal received from the video input unit 1 into digital data, and outputs the image data to the current image data calculation storage unit 3 and the comparison image data calculation storage unit 4 (step ST2). . Specifically, conversion into multi-value data of 8 bits per pixel is common.

  The current image data calculation / storage unit 3 creates and stores "current image data" based on the image data input from the A / D converter 2 (step ST3). In many cases, the input data itself is “current image data”. However, this is only an example, and “current image data” may be created by other methods. The current image data may be recorded as necessary, and the received current image data may be output to the change area extraction unit 5 as it is without accumulating the current image data.

  The comparison image data calculation storage unit 4 creates and stores “comparison image data” based on the image data input from the A / D converter 2 (step ST4). The “comparison image data” indicates a monitoring image of a monitoring target area in a state in which no reporting event (for example, an intruder) has appeared. However, this is only an example, and “comparison image data” may be created by other methods. Further, the comparison image data may be recorded as necessary. The comparison image data is not accumulated, the comparison image data is created from the current image data received from the A / D converter 2 and the change area is extracted as it is. The image data may be output to the unit 5, or the comparison image data created artificially may be acquired from the outside and output to the change region extraction unit 5 as it is.

The change area extraction unit 5 acquires “current image data” and “comparison image data” from the current image data calculation storage unit 3 and the comparison image data calculation storage unit 4 and compares them (step ST5).
When there is a different part between “current image data” and “comparison image data” (in the case of “YES” in step ST5), the change area extraction unit 5 extracts the different part as a change area, and the recognition processing unit 6 And output to the disturbance determination unit 7. The change area is an image obtained by binarizing the difference between the current image and the background image in pixel units. The change area is generally a moving object in the video, but the brightness change of the headlight of the car is also extracted as the change area.
On the other hand, when there is no different portion between the “current image data” and the “comparison image data” (in the case of “NO” in step ST5), the processing is ended.

The recognition processing unit 6 receives the change area data from the change area extraction unit 5 and determines whether or not it is a report target, that is, whether or not a report event has occurred, and a report event occurs. If it is determined that the notification event has occurred, the notification unit 8 is notified of the occurrence of the notification event (step ST7). This determination is similarly made according to whether or not a conventional monitoring image processing apparatus satisfies a predetermined condition. If the predetermined condition is satisfied, it is determined that the image is to be issued.
Examples of the predetermined conditions include the following conditions.
(1-1) The change area is within the monitoring range (1-2) The change area is a predetermined area (1-3) The change area is a predetermined vertical and horizontal dimension (1-4) The change area is a predetermined area Note that this is only an example, and it may be determined whether or not the report is to be issued under other conditions.

The disturbance determination unit 7 receives the change area data from the change area extraction unit 5, and determines whether there is a disturbance in the received change area data (step ST8). Details of this determination will be described later.
The reporting unit 8 receives instructions from the recognition processing unit 6 and the disturbance determination unit 7 and determines whether or not it is a reporting target (step ST9). That is, in step ST7, it is determined whether or not the recognition processing unit 6 has determined that it is a notification target, specifically, whether or not a notification of the generation of a notification target has been received from the recognition processing unit 6.

If it is determined in step ST9 that it is not a notification target, that is, if it is determined that a notification of the generation of a notification target has not been received (in the case of “NO” in step ST9), the notification unit 8 performs the following processing. Is skipped and the process ends.
In step ST9, when it is determined that it is a notification target, that is, when it is determined that a notification of generation of a notification target has been received (in the case of “YES” in step ST9), the reporting unit 8 generates a disturbance. It is determined whether or not (step ST10). Specifically, in step ST8, it is determined whether or not a notification that a disturbance has occurred has been received from the disturbance determination unit 7.
Details of the disturbance determination by the disturbance determination unit 7 will be described later.

When it is determined in step ST10 that there is no disturbance, that is, when it is determined that a notification of the occurrence of disturbance has not been received (in the case of “NO” in step ST10), the reporting unit 8 needs to report. Judgment is performed, and a notification process is performed (step ST11).
Specifically, for example, a specific signal is sent to a host PC or the like, or a buzzer of the apparatus is sounded.
On the other hand, if it is determined in step ST10 that there is a disturbance, that is, if it is determined that a notification of the occurrence of a disturbance has been received (in the case of “YES” in step ST10), the reporting unit 8 does not require reporting. It is determined that there is, and the notification process is restricted (step ST12).

  That is, even if the recognition processing unit 6 determines that the notification target exists, the notification unit 8 determines that the target is an unnecessary detection target if the disturbance determination unit 7 determines that a disturbance has occurred. , Regulate the alert. Conversely, if the recognition processing unit 6 determines that the notification target exists and the disturbance determination unit 7 does not detect the occurrence of the disturbance, the notification unit 8 determines that the notification target determined by the recognition processing unit 6 is true. Judge as the subject of the report and issue a report. The specific contents of the regulation of notification will be described later.

FIG. 4 is a flowchart illustrating disturbance determination in disturbance determination unit 7 according to Embodiment 1 of the present invention.
When the change region extraction unit 5 extracts a change region in the camera monitoring area, the disturbance determination unit 7 determines whether the change region is due to the headlight of the car. Hereinafter, the operation of the disturbance determination unit 7 will be described with reference to FIG.
The ferret diameter confirmation unit 20 of the disturbance determination unit 7 acquires the ferret diameter from the change region extracted by the change region extraction unit 5 (step ST101).

FIG. 5 is a diagram showing an example of the current image stored in the current image data calculation storage unit 3 and a comparison image (background image) stored in the comparison image data calculation storage unit 4. FIG. 5A shows an example of the current image, and FIG. 5B shows an example of a comparison image (background image).
FIG. 5 illustrates an example of a vehicle that travels with a headlight turned on at night on a road. That is, the image showing the vehicle (FIG. 5A) is the current image, and the image not showing the vehicle (FIG. 5B) is the background image.
Note that the current image in FIG. 5A is a video temporally continuous with the background image.

FIG. 6 is a diagram illustrating an example of the change area extracted by the change area extraction unit 5.
In the background image shown in FIG. 5B, a normal road image can be seen, but in the current image (FIG. 5A) at the next moment, a car appears behind the image.
In this case, the change area viewed from the monitoring image processing apparatus is “headlight” which is a difference between the current image and the background image, as shown in FIG. That is, in FIG. 6, the dyed area is the change area.

Here, a circumscribed rectangle of the change region as shown in FIG. 6 is called a ferret diameter.
Specifically, the solid frame line written in FIG. 7 is a circumscribed rectangle (Ferre diameter) derived from the change region of the “headlight”. In the conventional image processing apparatus for monitoring, in this case, if this “headlight” is determined to be a notification target, the notification processing is performed in response to the determination.
In step ST101, the ferret diameter checking unit 20 acquires the ferret diameter shown in FIG.

The block division unit 21 of the disturbance determination unit 7 defines a plurality of small blocks so as to include the ferret diameter acquired by the ferret diameter confirmation unit 20 in step ST101 (step ST102). Here, as an example, small blocks of “8 pixels × 8 pixels” are densely arranged in a range where the ferret diameter is hidden.
FIG. 8 shows a state in which small blocks of “8 pixels × 8 pixels” are densely arranged in a range in which the ferret diameter is hidden.

The block density measurement unit 22 of the disturbance determination unit 7 calculates the density of the change region corresponding pixels for each small block divided by the block division unit 21 in step ST102, and calculates the luminance average value of the change region corresponding pixels in the small block. Obtained (step ST103). Here, the luminance average value calculated by the block gray level measuring unit 22 is referred to as a “light gray value”. In general, the magnitude of luminance is called shading, but it may be a color.
FIG. 9 is a diagram showing a result of the block density measuring unit 22 calculating the density values of all blocks.

The light source position calculation unit 23 of the disturbance determination unit 7 searches for the “dark vertex block” based on the gray value of each small block calculated by the block gray level measurement unit 22 in step ST103 (step ST104).
The light / dark vertex block is defined as “a block in which there is no block having a lighter / darker value than the adjacent surrounding blocks”, and this light / dark vertex block is assumed to be a position where the headlight light source exists. That is, the light / dark vertex block indicates the position most likely to be the light source of the headlight.
Actually, the road surface reflection may have a larger gray value, but as long as it is derived from a headlight as a light source, the determination may be made.
The light source position calculation unit 23 determines a light / dark vertex block and sets the light / dark vertex block as a “light source”.
FIG. 10 is a diagram illustrating an example of a light / dark vertex block. The blocks marked in FIG. 10 are shaded vertex blocks.
As shown in FIG. 10, the light / dark vertex block is not limited to one block, and increases or decreases depending on the presence of a plurality of defined headlight light sources and the degree of reflection.

FIG. 11 is a diagram illustrating a road area.
A “road area” as shown in FIG. 11 refers to a range regarded as a road on the monitoring image, is set in advance, and is stored in the road area matching unit 24 of the disturbance determination unit 7.
The road area matching unit 24 determines whether or not the shade vertex block defined by the light source position calculation unit 23 in step ST104 belongs to the “road area” stored in itself (step ST105). In addition, when there are a plurality of shade vertex blocks, it may be determined whether any one shade vertex block belongs to the “road area”. If it is determined whether or not it belongs, a result can be obtained with good accuracy.

In step ST105, when the road area matching unit 24 determines that the light / dark vertex block, that is, the “light source” is not in the “road area” (in the case of “NO” in step ST105), it is determined that the disturbance is not caused by the headlight. The subsequent processing is skipped and the disturbance determination is terminated.
In step ST105, when the road area matching unit 24 determines that the light / dark vertex block, that is, the “light source” is in the “road area” (in the case of “YES” in step ST105), in many cases, it is a vehicle headlight. In order to increase the accuracy, the headlight determination unit 25 of the disturbance determination unit 7 uses, for example, the following (2-1) for the light / dark vertex block that the road area matching unit 24 determines to belong to the road area. Based on the conditions of (2-5), a final determination is made as to whether the headlight is used (step ST106).

(2-1) The shade vertex block continued to exist in the road area for a predetermined duration (eg, 0.5 seconds).
(2-2) light and shade moving direction range apex block is predetermined: moved to (eg road uplink and downlink directions).
(2-3) The density vertex block is equal to or higher than a predetermined density value (eg, the density value is 200 or more).
(2-4) The shade vertex block moved in the road area within a predetermined speed range (eg, 10 pixels / second or more).
(2-5) The generation time is from sunset to sunrise.

The headlight determination unit 25 of the disturbance determination unit 7 determines whether or not a disturbance due to the headlight has occurred for the shade vertex block determined by the road area matching unit 24 to belong to the road area (step ST107). Specifically, when all of the conditions (2-1) to (2-5) are satisfied, the corresponding light / dark vertex block is a headlight light source, that is, “currently a disturbance due to the headlight occurs”. Judgment is made ("YES" in step ST107), and the occurrence of disturbance is notified to the reporting unit 8 (step ST108).
On the other hand, when all of the conditions (2-1) to (2-5) are not satisfied, it is considered that no disturbance has occurred ("NO" in step ST107), and nothing is notified to the reporting unit 8.
The conditions (2-1) to (2-5) are merely examples, and the conditions for determining the disturbance can be set as appropriate.

As a result of the disturbance determination by the disturbance determination unit 7 described above, when the notification of the occurrence of the disturbance is received, the reporting unit 8 determines that the disturbance has occurred and restricts the reporting process. Therefore, although the “headlight” has been determined to be reported retroactively, the reporting unit 8 regulates the reporting and suppresses the occurrence of unnecessary detection by the “headlight”.
On the other hand, when there is no disturbance occurrence, that is, when a notification of disturbance occurrence is not received from the disturbance determination unit 7, the reporting unit 8 performs normal reporting processing.

Here, the regulation of the reporting process may be, for example, that the reporting unit 8 may not perform reporting while a disturbance is occurring, or may omit part of the reporting while a disturbance is occurring. Further, the reporting unit 8 may instruct the recognition processing unit 6 to correct the condition for determining whether or not a reporting event has occurred during the occurrence of a disturbance, or there has been a disturbance. Such information may be notified to a host PC or the like. What is necessary is just to restrict the normal alarm processing so that it is not recognized as a false alarm.
In the first embodiment, the disturbance determining unit 7 notifies the reporting unit 8 of the occurrence of the disturbance and the reporting unit 8 regulates the reporting process. However, the disturbance determining unit 7 generates the disturbance. May be notified to the recognition processing unit 6, and the recognition processing unit 6 may be configured to perform a process that is not a notification target when a disturbance occurs.

  As described above, according to the first embodiment, when a light / dark vertex block having an area having a specific condition appears, it is determined that headlight disturbance has occurred and the notification is restricted. Unnecessary detection can be efficiently reduced in the process.

Embodiment 2. FIG.
FIG. 12 is a block diagram of a monitoring image processing apparatus according to Embodiment 2 of the present invention.
Here, a detailed description of the same configuration as that of the monitoring image processing apparatus according to the first embodiment shown in FIG. 1 is omitted, and only a portion different from FIG. 1 is described.
In FIG. 12, the monitoring image processing apparatus further includes a mixed state determination unit 9.
When two or more ferret diameters exist in the monitoring image, the mixed situation determination unit 9 determines whether to give priority to the occurrence of disturbance or to give priority to pedestrian (intruder) reporting.

FIG. 13 is a flowchart of the reporting process of the monitoring image processing apparatus according to the second embodiment of the present invention. Hereinafter, a description will be given with reference to FIG.
In FIG. 13, the processing of steps ST1 to 12 is the same as that of steps ST1 to 12 of FIG.
In step ST8, when the disturbance determination unit 7 determines whether or not there is a disturbance in the change area data received from the change area extraction unit 5, the mixture determination unit 9 determines that the disturbance determination unit 7 For the change area determined to be present, a mixed state determination is made as to whether the headlight and the intruder are not mixed in the monitoring image (step ST13).

FIG. 14 is an example of a monitoring image in a situation where headlights and pedestrians (intruders) are mixed. In FIG. 14, a vehicle with a headlight is approaching from the back of the screen, and a pedestrian is crossing the road in front of the screen.
FIG. 15 is a diagram in which the ferret diameter is superimposed on each of the headlight and the pedestrian illustrated in FIG. 14. As shown in FIG. 15, the ferret diameter overlaps the headlight at the back and the pedestrian in front.
For example, when the monitoring image is as shown in FIG. 14, in the monitoring image processing apparatus described in the first embodiment, the disturbance determination unit 7 determines that there is a disturbance due to the headlight (step ST8 in FIG. 3). The reporting unit 8 regulates normal reporting (steps ST9, 10, 12 in FIG. 3).
As a result, it cannot be reported by a pedestrian in front, and will be lost.

Therefore, in the monitoring image processing apparatus according to the second embodiment, when two or more ferret diameters appear in the monitoring image, the mixed status determination unit 9 performs the mixed status determination.
In the case of a monitoring image as shown in FIG. 14, the mixed situation determination unit 9 first checks the “dark vertex” of each ferret diameter in FIG. 15. As shown in FIG. 16, the marked block is a light and dark vertex of each ferret diameter.

Next, for example, the mixed status determination unit performs the determination under the following conditions (3-1) and (3-2).
(3-1) The shade vertex block is equal to or higher than a predetermined density value (eg, shade value 200 or more). (3-2) The shade vertex block continues to exist in the predetermined road area (3- When all of 1) and (3-2) are satisfied, the mixed situation determination unit 9 determines that the shade vertex block is a headlight. When the conditions (3-1) and (3-2) are not satisfied, the mixed situation determination unit 9 determines that the shaded vertex block is a pedestrian (intruder).
The conditions (3-1) and (3-2) are merely examples, and it may be determined whether the headlight or the pedestrian (intruder) is in other conditions.

When the determination of the shade vertex block of each ferret diameter is completed, the mixed situation determination unit 9 determines whether there is a ferret diameter determined as “pedestrian (intruder)”.
Here, if there is a ferret diameter determined to be “pedestrian (intruder)”, the mixed situation determination unit 9 may determine that there is a ferret diameter determined to be “headlight” in the monitoring image. It is not a disturbance. That is, it corrects so that the notification part 8 may not be notified to the reporting unit 8. Specifically, the notification notifying the occurrence of the disturbance output by the disturbance determination unit 7 is canceled.
As a result, the reporting unit 8 performs the reporting process (step ST11), assuming that there is no disturbance (step ST10).

  As is apparent from the above, according to the second embodiment, even in a situation where a headlight disturbance and a pedestrian (intruder) coexist, necessary notification can be performed without degrading accuracy.

  In the present invention, within the scope of the invention, any combination of the embodiments, or any modification of any component in each embodiment, or omission of any component in each embodiment is possible. .

  DESCRIPTION OF SYMBOLS 1 Video input part, 2 A / D converter, 3 Current image data calculation accumulation | storage part, 4 Comparison image data calculation accumulation | storage part, 5 Change area extraction part, 6 Recognition processing part, 7 Disturbance determination part, 8 Notification part, 9 Mixed state determination unit, 20 ferret diameter confirmation unit, 21 block division unit, 22 block density measurement unit, 23 light source position calculation unit, 24 road area verification unit, 25 headlight determination unit.

Claims (3)

A video input unit for receiving a current monitoring image of the monitoring target area as a video signal;
An A / D converter that quantizes the video signal received by the video input unit into current image data;
A change area extraction unit that compares the current image data with comparison image data indicating the monitoring image in a state in which a notification event does not appear in the current image data, and extracts a change area in the monitoring image; ,
Recognizing the occurrence of a notification event by analyzing the change region, a recognition processing unit that outputs a notification event occurrence notification,
It is determined whether a disturbance has occurred in the change region, and includes a disturbance determination unit that outputs a disturbance occurrence notification,
The disturbance determination unit includes a ferret diameter confirmation unit that acquires a circumscribed rectangle of the change region extracted by the change region extraction unit;
A block dividing unit that divides the circumscribed rectangle obtained by the ferret diameter checking unit into a plurality of blocks;
A block density measuring unit that calculates a density value for each of the plurality of blocks divided by the block dividing unit;
Based on the shade value for each block calculated by the block shade measurement unit, the adjacent vertex block is searched for a shade vertex block that does not have a block with a shade value greater than itself, and the shade vertex block is used as a light source. A light source position calculation unit to determine;
A road area that is preliminarily stored as a road on the monitoring image and that determines whether the light source belongs to the road area;
For the light source determined by the road area matching unit to belong to the road area, the time during which the shade vertex block continues to exist in the road area, the moving direction range in which the shade vertex block has moved, the shade vertex block And a headlight determination unit for determining whether the light source is based on a density value, a speed range in which the light / dark vertex block has moved in the road area, and a generation time of the light source. An image processing apparatus for monitoring. Based on the notification event occurrence notification output by the recognition processing unit and the disturbance occurrence notification output by the disturbance determination unit, further includes a notification unit that performs a notification process,
The reporting unit is:
When receiving the notification event occurrence notification from the recognition processing unit, perform the notification processing,
Even when the notification event occurrence notification is received from the recognition processing unit, the headlight determination unit receives the disturbance occurrence notification output by determining that the light source is the headlight The monitoring image processing apparatus according to claim 1, wherein the notification processing is restricted. The disturbance determination unit performs a mixed state determination as to whether a headlight and an intruder are not mixed in the monitoring image determined to have a disturbance,
The monitoring image processing apparatus according to claim 1, further comprising: a mixed state determination unit that cancels the disturbance occurrence notification when the headlight and the intruder are mixed.

Images