JPH10136341A - Method for processing intruder detection image - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide the method for intruder detection processing image used for a mounting CCTV equipment in which mis-recognition is avoided in the case of occurrence of vibration or the like in an area sensor camera and mis- recognition due to a contrast difference is reduced. SOLUTION: The monitoring CCTV equipment has an are sensor camera 2 and an image processor 3 provided with a master image memory 6, a detection image memory 7, and a CPU 8. Then in the intruder detection and image processing method in which a master image and a detected image are compared and calculated, an intruder or an intruded object is detected to discriminate the presence of intrusion. The image processor 3 has a data memory section 9 that stores required position coordinate data dividing a storage are of each image memory into a required number in advance, and the storage/read of each image memory are conducted from each division area of a required number based on position coordinate data stored in the data memory section 9.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an improved intruder detection image processing method for detecting an intruder or an intruder in a monitoring area in a monitoring CCTV apparatus for monitoring an intruder or an intruder. is there.

[0002]

2. Description of the Related Art When an intruder or an intruder is monitored by using a monitoring CCTV device, a monitoring area may be, for example,
In the case where people or cars are constantly entering and leaving the entrance of the facility, people must constantly monitor the video monitor that displays the image captured by the TV camera, but for example, people and cars almost disappear. At the entrances and exits of the facilities at nighttime, the interior of the warehouse, etc., is not suitable for human monitoring because there is no change even if the images captured by the television camera are constantly monitored by a video monitor. Therefore, when there is a change in the image of the monitoring area captured by the television camera, the image processing device may detect the change in the image and generate an alarm, and many means have conventionally been developed and implemented. I have.

Conventionally, a monitoring CCT for monitoring a required area
The V device stores an imaging lens through which incident light from a required monitoring area passes, an area sensor camera that photoelectrically converts the incident light from the imaging lens to output a video signal, and a video signal input from the area sensor camera. An image processing device for comparing a video signal stored in the master image memory and the detected image memory to detect an intruder or an intruder and determine intrusion; And a video monitor for monitoring the image of the area. In a conventional intruder detection image processing method using the monitoring CCTV device, a reference image captured by an area sensor camera and having no intruder or intruder in a monitoring area is stored in a master image memory. Thereafter, an image for monitoring the intrusion of an intruder or an intruder in the same monitoring area is stored in the detected image memory, and the master image stored in the master image memory and the detected image stored in the detected image memory are read out. The contrast between the image and the image is used to determine the difference in the contrast of the image (the difference in the level of the video signal), and the presence or absence of an intruder or an intruder is determined based on the sum of the differences in the contrast.

[0004]

A conventional intruder detection image processing method according to the prior art is based on a detection image output from an area sensor camera in a monitoring CCTV apparatus when an area sensor camera for capturing an image of a monitoring area is shaken by vibration or the like. A shift occurs in the image between the detected image stored in the memory and the master image stored in the master image memory. When the difference is obtained, the difference amount increases. Therefore, even when there is no intruder or intruder, the sum of the shade differences becomes large, and there is a problem that it is determined that there is an intrusion, that is, erroneous recognition is performed. In addition, in the intruder detection image processing method according to the related art, the intrusion of an intruder or an intruder is determined based on the sum of the gray level differences between the master image and the detected image. Even if the overall image is brighter or darker than the image, there is a difference in gray level.Therefore, the memory of the master image memory is updated frequently, but erroneous recognition due to the sudden change in brightness despite the update There was a problem that occurs. The present invention solves the above problem, eliminates erroneous recognition when there is a vibration or the like in an area sensor camera, and can reduce erroneous recognition due to a difference in brightness between a master image and a detected image. It is an object of the present invention to provide an intruder detection image processing method used for a computer.

[0005]

In order to achieve the above object, an intruder detection image processing method according to the present invention comprises an area sensor camera for imaging a required monitoring area, and a reference image from the area sensor camera. A master image memory that stores in advance as a master image, a detection image memory that stores an image monitored from the area sensor camera as a detection image, and a central processing unit that controls storage, reading, and the like of each image memory. A monitoring CCTV device having an image processing device, wherein a comparison is made between a master image read from the master image memory and the detected image memory and a detected image, an operation is performed, an intruder or an intruder is detected, and an intruder is detected. In the intruder detection image processing method for determining presence / absence, the image processing apparatus may reduce a storage area of each image memory to a required number. A data memory unit that stores in advance the required position coordinate data to be divided, and stores and reads out each of the image memories under the control of the central processing unit, based on the position coordinate data stored in the data memory unit. This is performed by dividing into divided areas.

More specifically, in the intruder detection image processing method according to the present invention, the data memory section divides a storage area of the image memory into a required number to form a required divided area, and Position coordinate data having a required correction area and a required correction width are stored in advance. Further, the intruder detection image processing method according to the present invention may be arranged such that one divided area of the master image read out from the master image memory with required position coordinate data as a required divided area, The calculation is performed for all the divided regions by comparing the detected image read out with the position coordinate data having the additional correction width with one divided region and shifting one pixel at a time.

[0007]

DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of an intruding object detection image processing method in a monitoring CCTV apparatus according to the present invention is shown in FIGS.
This will be described with reference to FIG. FIG. 1 is a block diagram showing a monitoring CCTV device using the intruder detection image processing method of the present invention. In FIG. 1, reference numeral 1 denotes an imaging lens for imaging a required monitoring area for monitoring an intruder or an intruder;
Reference numeral 2 denotes an area sensor camera such as a television camera which photoelectrically converts incident light from the imaging lens 1 and outputs a video signal;
Denotes an image processing device for processing a video signal input from the area sensor camera 2 to detect an intruder or an intruder;
Denotes a video monitor for monitoring an image picked up by the area sensor camera 2. In the image processing device 3, reference numeral 4 denotes an A / D converter for converting a video signal input from the area sensor camera 2 into a digital video signal. Vessel 5
A memory input switching unit that switches a digital video signal input from the A / D converter 4 by a control signal from a central processing unit (hereinafter, referred to as a CPU) and outputs the digital video signal to a required memory. 5 is a master image memory for storing video signals to be switched and input by 5; 7 is a detected image memory for storing video signals to be switched and input by the memory input switching unit 5; 8 is a CPU for controlling the image processing device 3; A data memory unit for storing position coordinate data of the memory division,
Reference numeral 10 denotes a D / A converter that converts a digital video signal input from the A / D converter 4 into an analog video signal and outputs the analog video signal to a video monitor 11, and 12 denotes an external alarm output terminal.

In a monitoring CCTV apparatus using the intruder detection image processing method according to the present invention, incident light from a required monitoring area that has passed through the imaging lens 1 enters the area sensor camera 2. The area sensor camera 2 photoelectrically converts the incident light, performs necessary signal processing and level adjustment on the video signal output by the photoelectric conversion, and outputs a video signal of a predetermined level to the A / D conversion unit 4 of the image processing device 3. Output. A / D
The conversion unit 4 converts a video signal input from the area sensor camera 2 into a digital video signal, and outputs the digital video signal to the D / A converter 10 and the memory input switching unit 5. One D / A converter 10 converts the digital video signal input from the A / D converter 4 into an original analog video signal, and outputs it to the video monitor 11. The video monitor 11 displays an image of a required monitoring area captured by the area sensor camera 2.

The other memory input switching section 5 has a separate CP
It is controlled by a control signal input from U8, and outputs a digital video signal input from the A / D converter 4 to the master image memory 6 or the detected image memory 7. Note that C
In the data memory unit 9 connected to the PU 8, position coordinate data (XB _1s , YB _1s ), (XB _1e , YB _1e )... Shown in FIG.
(XB _ns , YB _ns ), (XB _ne , YB _ne ) and FIG.
Position coordinate data (XC _1s , YC _1s ) including a correction width D for correcting a displacement due to vibration shown in FIG.
(C _1e , YC _1e ) (XC _ns , YC _ns ), (XC _ne ,
YC _ne ) is stored, and when a master image is created in the master image memory 6, data of a read processing area is created under the control of the CPU 8. FIG. 2 shows the divided area B ₁
.About.B _n and shows the relationship of the detection area C ₁ -C _n, FIG. 2 (a) those entire split state, FIG. 2 (b) showing an enlarged portion shown in FIG. 2 (a) It is.

The master image stored in the master image memory 6 is
The creation is performed by setting the memory input switching unit 5 to a control signal from the CPU 8.
To input a signal to the master image memory 6, and
Split area B₁ From B_n Data memo with position coordinate data up to
Read from the file section 9 and divided area B₁ From B_n In the position
Reads image data stored in the master image memory unit 6
3A, first, as shown in FIG.₁ In
Average image density B_1mIs calculated, and shown in FIG.
As shown in FIG.₁ Level difference within
ΔB_1m1 ~ ΔB_1mn Is calculated, and the master image memory unit 6
The image data is rewritten and stored. This divided area B₁ ~
B_n Calculate the average density of the image in
Divided area B₁ ~ B_n Calculate the level difference between
Rewrite and store the image data in the image memory unit 6 by ΔB
_1m1 ~ ΔB_1mn From ΔB_nm1 ~ ΔB_nmn Create up to.
As a result, the average level fluctuation of the image due to light and dark can be calculated.
It is possible to extract the change in shading that has been removed. Sa
Further, the CPU 8 executes the division stored in the data memory unit 9.
Area B₁ From B_n Up to the position coordinate data and the correction width D
Data is read and the detection area C₁ To C_n Position coordinates
Data (XC_1s, YC_1s), (XC_1e, YC_1e) ・ ・ ・
(XC_ns, YC _ns), (XC_ne, YC_ne) And calculate
Data area 9 as detection area data.

Next, detection of an intruder or an intruder will be described. The reference video signal from the area sensor camera 2 is stored as a master image in the master image memory 6 as described above. A video signal to be monitored from the area sensor camera 2 is converted into a digital video signal by the A / D converter 4, passes through the memory input switching unit 5 switched to the detection image memory 7 side under the control of the CPU 8, and passes through the detection image memory 7. Is stored. Next, within the range of the position coordinate data (XC _1s , YC _1s ) and (XC _1e , YC _1e ) of the detection area C ₁ stored in the data memory unit 9,
Sectional areas B ₁ is shifted by one pixel to compare. This comparison, sectional areas B ₁ and the detection area C _1-1, as shown in FIG. 4, C _1-2 performed by the ··· C _1-n, the average concentration C _1-1 detection area C _1-1 i, and further calculate the average density C
_1-1 calculates a difference ΔC _1-1 i _s ~ΔC _1-1 i _e of each pixel from i, B ₁ part data ΔB of the master image memory unit 6 _1m1
The ~ΔB _1mn, calculates the similarity r ₁ by the following equation (1).

(Equation 1)

This calculation is performed based on the detection areas C _1-1 , C _1-2,.
··· Repeat until C _1-n to find a position with high similarity. As a result of obtaining the similarity, if the similarity with the data in the master image memory unit 6 is low and the matching position cannot be determined, it can be determined that there is an intruder or an intruder.
This similarity determination is performed for C _{1 to} C _n , and an alarm such as an intrusion to the outside is output via the alarm output terminal 12 according to a predetermined number of detection areas where an intruder or an intruder exists. Further, as a result of obtaining the similarity, if the similarity with the data in the master image memory unit 6 is high and the matching position can be determined, it can be determined that there is no intruder or intruder. The data is transferred to the master image memory unit 6, and the master image memory unit 6 is updated. In the monitoring area, for example, if a moving object such as a tree branch in a certain area in the monitoring area is known in advance, the detection of the corresponding area is stopped and the detection area is limited. In addition, it is possible to reduce erroneous detection by limiting the processing part.

Further, the processing of the above-described embodiment will be described with reference to the flowchart of FIG. This FIG.
Until the divided regions B ₁ .about.B _n is a flowchart of after registering the master image. Detection of intruders or intruders is
In “image input”, a video signal from the area sensor camera 2 is input to the detected image memory 7. In the divided area “calculation of average density of B _n ”, the sum of the levels of all pixels in the image portion of B _n is calculated, and the result is divided by the total number of pixels to obtain an average density. Further, in the “calculation of the density change of B _n ”, the density change is calculated by subtracting the average density for all the pixels of B _n . Detection Area "average density calculation of C _{n n"} is
C _n average concentration of _n is calculated, and then calculates the density variation of the "C _{n n} shade variation calculation" in C _{n n.} More thereto by shading variation of obtained B _n and C _{n n} gray variation in the "B
wherein in _n and C _{n n} similarity calculation (r _n) "(1) performing a similarity calculation of expression. This obtains the similarity to all the C _{n 1} to C _{n n,} obtains the highest r _n MAX of similarity. The presence or absence of an intruder or an intruder is determined based on whether or not the value of the similarity r _n MAX obtained here is greater than or equal to a specified value. At this time, if there is no intruder or intruder,
In "Moving Images of C to B", it was high the highest similarity C _n
the _n image transferred to B _n as the master image, and updates the register. If it is determined that there is an intruder or an intruder, an alarm is output to the outside by, for example, a contact signal or the like in “alarm output”. These processes are performed by using master images B _{1 to} B ₁
Perform up to _n .

[0014]

According to the present invention, there is provided a monitoring CCTV apparatus capable of eliminating erroneous recognition due to vibration or the like in an area sensor camera and reducing erroneous recognition due to a difference in brightness between a master image and a detected image. An intruder detection image processing method to be used can be provided.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an embodiment of a monitoring CCTV device using an intruding object detection image processing method of the present invention.

FIG. 2 is a conceptual diagram of image processing of an embodiment of an intruder detection image processing method according to the present invention.

FIG. 3 is a conceptual diagram of image processing of an embodiment of an intruder detection image processing method according to the present invention.

FIG. 4 is a diagram illustrating a comparison between a master image and a detected image according to an embodiment of the intruder detection image processing method of the present invention.

FIG. 5 is a flowchart of an embodiment of an intruder detection by the intruder detection image processing method of the present invention.

[Description of Signs] 1 imaging lens, 2 area sensor camera, 3 image processing device, 4 A / D conversion unit, 5 memory input switching unit, 6 master image memory, 7 detection image memory, 8 CPU, 9 data memory unit, 10 D / A converter, 11 Video monitor, 12 Alarm output terminal.

Claims (3)

[Claims] 1. An area sensor camera for capturing an image of a required monitoring area, a master image memory for previously storing a reference image from the area sensor camera as a master image, and detecting an image to be monitored from the area sensor camera A monitoring CCTV device comprising: a detection image memory for storing as an image; and an image processing device including a central processing unit for controlling storage, reading, and the like of each of the image memories, wherein the master image memory and the detection image are stored. The intruder detection image processing method of comparing a master image read from a memory with a detected image and performing an operation to detect an intruder or an intruder and determine the presence or absence of an intruder, wherein the image processing device comprises: A data memory unit for previously storing required position coordinate data for dividing the storage area of the memory into a required number; An intruding object detection image processing method, wherein, under the control of the central processing unit, storage and reading of each of the image memories are performed by dividing into a required number of divided areas by position coordinate data stored in the data memory unit. . 2. The data memory unit according to claim 1, wherein the data memory unit divides a storage area of the image memory into a required number and divides the storage area into a required division area; An intruder detection image processing method characterized by storing in advance position coordinate data having a correction width. 3. The image processing apparatus according to claim 1, wherein one divided area of the master image read out from the master image memory with required position coordinate data to be a required divided area; An intruding object characterized by performing a comparison with one divided region of a detected image read out with position coordinate data having a required correction width added to the divided region and shifting by one pixel for all divided regions. Detection image processing method.