JPH06243259A - Person recognizing method - Google Patents

Abstract

PURPOSE:To reduce erroneous report in an intruder monitor system by utilizing the figure peculiar to a person, which is generated at the time of walking, to accurately recognize the person. CONSTITUTION:The difference picture between an input picture for confirmation of the presence or the absence of an intruder and a background picture is obtained by a difference processing 3 and is by binary processing 4, and the presence or the absence of the intruder is discriminated from the change area of this binary difference picture. In this method, a circumscribed rectangle of the change area is obtained by a change area segmenting processing 5. The circumscribed rectangle is longitudinally equally divided into three by a change area dividing processing 6, and the projection histogram obtained by projecting the change area in this divided circumscribed rectangle onto the axis of abscissas is generated by a projection histogram generating processing 7. A feature quantity from the projection histogram is extracted by a feature quantity extracting processing 8, and the person is recognized by a intruder discriminating processing 9 in accordance with relations between the extracted feature quantity and the aspect ratio in the circumscribed rectangle of the change area.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a person recognition method for recognizing a person in an intruder monitoring system for detecting an intruder from an image picked up by a television camera (hereinafter referred to as a TV camera) or the like.

[0002]

2. Description of the Related Art Conventionally, as a technique in such a field,
For example, some documents were described in the following documents. References: 1990 IEICE Spring National Convention Proceedings 6D-436, Yoshiro Sakuma, Kiyoshi Ito, Isao Masuda “Detecting Intruding Objects by Difference Between Frames of Differential Image” 6-438 Conventionally, as a person recognition method in an intruder monitoring system, as described in the above document, an input image for confirming the presence or absence of an intruder and a background image are input by a TV camera, and the image is input. The change area is obtained from the difference binary image between the background image and the background image, the presence or absence of an intruder is determined from the feature amount of the change area, and if there is an intruder, an alarm is issued. Here, as a method of determining the presence or absence of an intruder from the characteristic amount of the changed region, the presence or absence of an intruder is determined using the area of the changed region and the circumscribed rectangle as parameters. In other words, if the area of the change area is equal to or larger than a certain value and the aspect ratio of the circumscribing rectangle (vertical length / horizontal length of the circumscribing rectangle) is within a range, the intruder is determined. It was

[0003]

However, the person recognition method in the conventional intruder monitoring system has the following problems. Image the intruder with a TV camera,
When the presence or absence of an intruder is detected based on the image, an intruder far from the installation location of the TV camera has a small area of the change region on the image. On the other hand, in the vicinity of the installation location of the TV camera, the area of the change region becomes large even if a small object moves. Therefore, if the threshold value is set low in the area of the change area in order to recognize an intruder far from the installation location of the TV camera, the change area caused by the movement of a small object other than a person near the TV camera is recognized. It may not be possible to make a distinction and may give false alarms. Also, the aspect ratio of the circumscribed rectangle, which is another parameter, is about 2-3 times larger when the person's normal walking motion is captured from the side when the foot and arm are closed than when they are opened. Become. Therefore, even when the intruder is determined only based on the aspect ratio, there are many false recognitions. As described above, if the presence or absence of an intruder is determined based only on the area of the change region and the aspect ratio of the circumscribing rectangle, many false alarms may occur. The present invention solves the problem of erroneous recognition caused by the method of determining the presence or absence of an intruder from the area of the change area and the aspect ratio of the circumscribing rectangle, as a problem that the above-mentioned conventional art has, It is intended to provide a person recognition method in an intruder monitoring system for accurately recognizing a person by using a shape and reducing false alarms.

[0004]

In order to solve the above problems, the present invention binarizes a difference image between an input image for confirming the presence or absence of an intruder and a background image, and the difference binarized image. In the method of determining the presence or absence of an intruder from the change area, the change area in each of the rectangles A1, A2, and A3 obtained by vertically dividing the circumscribed rectangle of the change area is obtained by projecting on the horizontal axis. The person is recognized based on the relationship between the following features (1) to (6) which are the feature amounts of the projection histogram and the aspect ratio of the circumscribed rectangle of the change region. Feature (1); Number of histogram blocks in each circumscribed rectangle A1, A2, A3 Feature (2); Average value F of each block Feature (3); Difference between circumscribed rectangles of average value F of each block Feature (4); Maximum value number in each block Feature (5); Area ratio R in each block R Feature (6); Position M of maximum value in divided circumscribed rectangle A3

[0005]

According to the present invention, since the person recognition method is configured as described above, the features (1) to (6) represent the features of the shape peculiar to the person that occur during walking. Accurate recognition of a person can be performed by using the shape of. Therefore, the above problem can be solved.

[0006]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 shows an embodiment of the present invention and is a diagram showing a general processing procedure of a person recognition method in an intruder monitoring system. In this person recognition method, a background image and an input image are input by an image input device such as a TV camera,
Image input processing for storing them in the background image memory 1 and the input image memory 2 is performed, and difference processing 3, binarization processing 4, change area cutout processing 5, change area division are performed on the input background image and input image. Process 6, projection histogram creation process 7, feature amount extraction process 8, and intruding object determination process 9
And outputs the recognition result for the intruder. Hereinafter, each of these processes will be described with reference to FIGS.

Image Input Processing to Memories 1 and 2 FIG. 2 is a diagram showing an example of a background image, and FIG. 3 is a diagram showing an example of an input image. First, using an image input device such as a TV camera, as shown in FIG. 2, only the background in which no intruding object exists is photographed,
The background image is stored in the background image memory 1 in advance. Then, the background image is photographed at the same position as that photographed, and the photographed input image as shown in FIG. 3 is stored in the input image memory 2.

After the difference processing 3 image input processing, in the difference processing 3, in order to detect a change from the background image in the input image, the background image and the input image input from the background image memory 1 and the input image memory 2 respectively. Then, the absolute value of the difference is calculated, the difference image is obtained, and then the process proceeds to the binarization processing 4.

Binarization Process 4 FIG. 4 is a diagram showing an example of a differential binarized image. In the binarization process 4, the difference image obtained in the difference process 3 is binarized by using an appropriate threshold value, and a difference binarized image as shown in FIG. 4 is obtained. In this difference binarized image, the pixel value of the area that has changed from the background image in the input image is set to 1, and the pixel value of the area that has not changed is set to 0, and the area change is extracted. However, some intruding objects may locally have the same density as the background image. In the binary difference image in that case, as shown in FIG. 4, only that part is not recognized as an intruding object, and the change region is cut off. Therefore, each change area is recognized separately, and is not recognized as one original intruding object. Therefore, the following change area cutout processing 5 is performed.

Change Area Cutting Process 5 FIGS. 5A to 5F are diagrams showing examples of patterns when a person walks. In the change area cutout process 5, the difference binarized image is expanded several times and then shrunk the same number of times in order to combine the change areas representing one intruding object into one. The expansion is to convert a pixel having a value of 0 into a pixel having a value of 1 when there is a pixel having a value of 1 around the pixel having a value of 0. In other words, it expands around the change area by one pixel. Contraction is the opposite, and reduces the area around the change area by one pixel. By this work, the change area that seems to represent one intruding object,
It is extracted as one area. Then, as shown in FIGS. 5A to 5F, the change region thus obtained is
Enclose with a circumscribed rectangle, and the area surrounded by this circumscribed rectangle is the difference 2
Cut out from the binarized image. When determining whether or not the intruding object is a person from this clipped area, if it is attempted to make a determination based only on the aspect ratio of the circumscribing rectangle, the aspect ratio of the circumscribing rectangle changes greatly, as shown in FIG. Can't recognize people. However, the aspect ratio of the circumscribed rectangle has some restrictions. That is, the aspect ratio of the circumscribed rectangle becomes smaller (that is, the circumscribed rectangle spreads horizontally) when the person opens his / her arms and legs (FIG. 5 (d)), and the aspect ratio becomes larger. This is the case when the arms and legs are closed (FIG. 5 (b)). However, since there are individual differences in how the arms are opened during walking, it cannot be said that the degree of opening of the arms and the change in the aspect ratio are always related. On the other hand, the opening / closing of the foot and the change of the aspect ratio always have the relationship as described above. Therefore, if such regularity of the aspect ratio of the circumscribed rectangle and the shape of the portion corresponding to the foot in the change region can be captured, it can be accurately recognized whether or not the intruding object is a person. Therefore, the following change area division processing 6 is performed.

Change Region Division Processing 6 FIG. 6 is a diagram showing an example of a change region divided into three equal parts. In the change area division processing 6, as shown in FIG. 6, the circumscribed rectangle of the change area is vertically divided into three equal parts. This is to make it easier to capture the shape of the change area by dividing the human arm and foot parts that cause the aspect ratio change into different areas in the change area representing the person. The rectangles formed by this trisection are referred to as A1, A2, and A3 from the top. Then, the process proceeds to the next projection histogram creating process 7.

Projection Histogram Creation Process 7 FIGS. 7A to 7F are diagrams showing examples of projection histograms of divided circumscribed rectangles corresponding to FIGS. 5A to 5F, respectively. The horizontal axis of this projection histogram is the circumscribed rectangle A
1, A2, A3 corresponding to the horizontal axis, the vertical axis is A1, A2, A
3 corresponds to the frequency HIST of each histogram. 8 and 9 are diagrams showing the block BK in FIG. 7.
FIGS. 10A to 10F are diagrams showing average values in the projection histograms of FIGS. 7A to 7F. Also, FIG.
1 to 14 are diagrams showing the block BK in FIG. 7.

In the projection histogram creation process 7, the pixels of the value 1 representing the change region are vertically added to each of the divided circumscribed rectangles, and as shown in FIGS. The projection histogram of the change area on the horizontal axis is obtained. Then, this histogram is smoothed so that the features can be easily captured. Then, FIG. 7 (a)
As is clear from (f), the divided circumscribing rectangles A1, A2,
The shape feature of the change region in A3 appears as the shape feature of the projection histogram. If there is no change in the shape of the change area with respect to the change in the aspect ratio as in the rectangle A1,
The shape of the projection histogram also hardly changes. Also,
When the shape of the upper part from the chest of such a person is projected, the shape of the projection histogram is a block B close to a triangular shape.
It becomes K. However, as shown in FIG. 7 (f), the block BK is a block in which the frequency HIST is 0, which is divided at a frequency HIST of 0 in the projection histogram. As shown in FIG. The range of the horizontal axis in BK is the base L, and the maximum value is the height H. In the rectangle A1, this block BK does not extremely exist at the end of the horizontal axis.

In the rectangle A2, the aspect ratio of the circumscribing rectangle is large, and when the human arm is closed, the shape of the projection histogram becomes a trapezoidal shape as shown in FIG. As the human arm portion in the change region expands as the aspect ratio of the circumscribing rectangle decreases, both ends of the base L of the trapezoid of the block BK, which has a shape close to a trapezoid, extend as the shape of the projection histogram increases. , The shape changes like a top hat (Fig. 7 (a) → Fig. 7 (d)). However, depending on the person, there are cases in which the arm is hardly opened during walking, and the shape of the histogram at that time hardly changes regardless of the change of the aspect ratio like the rectangle A1 of the block BK having a trapezoidal shape. It does not exist at the extreme end of the horizontal axis.

In the rectangle A3, when the aspect ratio of the circumscribing rectangle is large, the foot portion of the change region is closed and becomes a lump, so that the shape of the projection histogram becomes a trapezoidal shape of the rectangle A2. The aspect ratio of the circumscribed rectangle is reduced, and
As shown in FIG. 7D, when the foot portion of the change area is opened and the foot shape is an inverted V shape, the shape of the projection histogram is as shown in FIG.
As shown in (d), it has a shape like a mountain having two maximum values. Further, when the foot portion has a “he-shaped” shape as shown in FIG. 5 (e), the histogram has one mountain at the end of the block as shown in FIG. 7 (e), and is otherwise flat. It becomes a block BK that looks like a shape. In this way, by paying attention to the change in the shape of the projection histogram of the change region in the circumscribing rectangles A1, A2, and A3 divided into three equal parts, the walking state of the person can be expressed, and the person can be accurately recognized. Therefore, the following feature amount extraction processing 8 is performed.

Feature amount extraction process 8 In the feature amount extraction process 8, the following feature amounts are represented by rectangles A1 to A3.
From the projection histogram of. Feature (1): Number of blocks of histogram in each circumscribed rectangle A1, A2, A3 Feature (2): Average value F of each block BK Feature (3): Between circumscribed rectangles of average value F of each block BK Difference Feature (4); Maximum value number in each block BK Feature (5); Ratio of area S in each block BK R Feature (6); Position M of the maximum value in the rectangle A3 However, the average of each block BK here The value F is shown in FIG.
It is the x-coordinate of the horizontal axis of the average value F shown in, and is calculated by the following equation (1). The vertical axis of FIG. 9 indicates the frequency HIST of the histogram.
Is.

[0017]

[Equation 1] As shown in FIG. 9, the ratio R of the area S of each block BK means that the area of a quadrangle ABCD formed by multiplying the base L of the block BK by the height H is 100%.
It is the ratio of the area S of K to the square ABCD. This is calculated by the following equation (2).

[0018]

[Equation 2] Further, the position M of the maximum value is as shown in FIG. The reason for extracting such a feature amount is as follows.
As can be seen from FIG. 7, the number of blocks of the histogram in each of the divided circumscribed rectangles A1, A2, and A3 of the feature (1) is 1 in every histogram. The average value F of the feature (2) is not extremely close to the end positions in the histograms of the rectangles A1 to A3 as shown in FIG. The difference is small between the rectangles A1 to A3 and is almost a straight line. That is, the features (1) to (3) represent the overall features of the change region, and the change regions that are completely different from the human shape can be distinguished from each other based on these feature amounts.

On the other hand, the features (4) to (6) are for each block B.
The characteristic of the shape of K is shown. The maximum value number in each block BK of the feature (4) is 1 in the rectangles A1 and A2 regardless of the change in the aspect ratio of the circumscribing rectangle. In the rectangle A3, it becomes 2 or 1 according to the change of the aspect ratio. The relationship will be described later. Area S of feature (5)
As for the ratio R of the rectangle A1, the shape is almost as shown in FIG. 11 regardless of the change of the aspect ratio of the circumscribing rectangle. Therefore, when the area of the quadrangle abcd is 100%, the area of the triangle edc is 50%. %, The area ratio R is 50% or more. In the rectangle A2, when the aspect ratio of the circumscribing rectangle is large, as shown in FIG.
Since the shape of K is close to a trapezoid, the area ratio R is 5
Greatly exceeds 0%. In addition, the aspect ratio becomes smaller, and
Even if the block BK has a top hat-like shape as in 3, the area ratio is 50% or more. Rectangle A3
In the above, since there is also a relation with the aspect ratio of the circumscribed rectangle, it will be described in detail below. Features of rectangle A3 (4)
The relationship between (6) and the aspect ratio of the circumscribed rectangle is as follows.

The maximum value number in the block BK of the feature (4) in the rectangle A3 is 2 when the aspect ratio is very small as shown in FIG. 7 (d), and as shown in FIGS. 7 (a) and 7 (d). When the aspect ratio is very large, it becomes 1. However, FIG.
Within a certain range of aspect ratios such as (c), (e), and (f), the maximum number is 2 (in the case of FIG. 7C), or
1 (in the case of FIGS. 7E and 7F). In the case where the aspect ratio has a maximum value of 1 within a certain range, FIGS.
The block BK has a shape in which one peak exists at the end of the block BK as typified by FIG. This is different from the shape of the block BK as shown in FIG. Therefore, area S of feature (5)
In the shape as shown in FIG. 7 (a), the ratio becomes 50% or more, but in the shape as shown in FIG. 7 (e),
As shown in FIG. 14, it becomes 50% or less, and the absolute value of the difference between the position M of the maximum value and the center BO of the block BK becomes large. That is, the features (4) to (6) in the rectangle A3
If the aspect ratio of the circumscribed rectangle is summarized, the maximum value is 2 when the aspect ratio is small. Within a certain range of the aspect ratio, the number of maximum values is 2 or 1. In this case, if the number of maximum values is 1, the area ratio is 50% or less, and the position M of the maximum value exists at the end of the block BK. . When the aspect ratio is large, the maximum number is 1, and the area ratio is 50 at this time.
That is, the position M having the maximum value or more and the maximum value is near the center BO of the block BK.

Intruding object determination processing 9 Finally, in the intruding object determination processing 9, the change area is determined from the relationship between the six feature amounts obtained in the feature amount extraction processing 8 and the aspect ratio of the circumscribed rectangle of the change area. Then, it is determined whether the intruding object is a person. Flow charts of the person determination procedures 1 and 2 are shown in FIGS. 15 and 16. First, FIG.
In step S1 of 1., it is checked whether the aspect ratio of the circumscribed rectangle is within a certain range. Next, in step S2, it is determined whether the number of blocks in the three projection histograms is all 1, and in steps S3 and S4,
The average values F of the blocks of the three projection histograms are all within a certain range, and between the divided circumscribed rectangles (A1 and A2, A
It is checked that the difference between the average values F of 2 and A3, A1 and A3) is within a certain range. Then, in step S5
In S6, it is checked that the maximum value number in each block BK of the projection histograms of the rectangles A1 and A2 is 1 and that the ratio R of the area S is within a certain range.
Next, the boundary value of the aspect ratio is set as a1 <a2 <a3 <a4, and the processing in the rectangle A3 depending on the aspect ratio is distributed in steps S7 to S9 in FIG. In the section where the aspect ratio is from the boundary value a1 to the boundary value a2, it is checked in step S10 whether there are two maximum values. In the section where the aspect ratio is the boundary value a2 to a3, there are two maximum values in step S11, or if the maximum value is one, the position M of the maximum value and its block BK are determined in steps S13 and S14. The ratio R of the area S is 50% within a certain range (e1 or more and e2 or less) of the difference from the center
It is checked whether each is within the following range (f1 or more and f2 or less). If the aspect ratio is within the range of the boundary values a3 to a4, it is checked in step S15 and S16 whether the number of maximum values is one and the area ratio is within a certain range of 50% or more (d1 or more and d2 or less). .

Only when all of these conditions are satisfied, the intruding object is recognized as a person, the recognition result is output, and the person recognition processing is ended. As described above, in this embodiment, the aspect ratio of the circumscribing rectangle of the change area and each change area in each of the rectangles A1, A2, and A3 obtained by vertically dividing the circumscribing rectangle into three equal parts are projected onto the horizontal axis. Since the person is recognized based on the features (1) to (6) which are the feature amounts of the projection histogram obtained in this way, the person can be accurately recognized, and false alarms can be reduced in the intruder monitoring system. The present invention is not limited to the above embodiment, and various modifications are possible, for example, the intruding object determination process 9 of FIG. 1 is processed by a procedure other than the person determination procedures 1 and 2 of FIGS. 15 and 16. Is.

[0023]

As described above in detail, according to the present invention, the aspect ratio of the circumscribing rectangle of the change area and the respective rectangles A1, A2, A3 obtained by vertically dividing the circumscribing rectangle into three equal parts. Since the person is recognized based on the features (1) to (6), which are the feature amounts of the projection histogram obtained by projecting each change region on the horizontal axis, the person can be accurately recognized, and the intruder can be monitored. It is possible to reduce false alarms in the system.

[Brief description of drawings]

FIG. 1 is a diagram showing a schematic processing procedure of a person recognition method according to an embodiment of the present invention.

FIG. 2 is a diagram showing an example of a background image.

FIG. 3 is a diagram showing an example of an input image.

FIG. 4 is a diagram showing an example of a difference binarized image.

FIG. 5 is a diagram showing an example of a pattern when a person walks.

FIG. 6 is a diagram showing an example of three equally divided regions.

FIG. 7 is a diagram showing an example of a projection histogram of a divided circumscribed rectangle.

8 is a diagram showing a block BK in FIG. 7. FIG.

9 is a diagram showing a block BK in FIG. 7. FIG.

10 is a diagram showing average values in the projection histogram of FIG.

11 is a diagram showing a block BK in FIG. 7. FIG.

FIG. 12 is a diagram showing a block BK in FIG. 7.

FIG. 13 is a diagram showing a block BK in FIG. 7.

FIG. 14 is a diagram showing a block BK in FIG. 7.

15 is a flowchart showing a person determination procedure 1 in the intruding object determination processing 9 of FIG.

16 is a flowchart showing a person determination procedure 2 in the intruding object determination processing 9 of FIG.

[Description of Codes] 1 background image memory 2 input image memory 3 difference processing 4 binarization processing 5 changing area cutout processing 6 changing area dividing processing 7 projection histogram creation processing 8 feature amount extraction processing 9 intruding object determination processing

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁵ Identification code Internal reference number FI Technical display location G08B 23/00 N 9377-5G

Claims (1)

[Claims] 1. A method of binarizing a difference image between an input image and a background image for confirming the presence or absence of an intruder, and determining the presence or absence of an intruder from a change area of the difference binarized image, wherein the change The following features (1) -features, which are the feature amounts of the projection histogram obtained by projecting each change region in each rectangle A1, A2, A3 obtained by vertically dividing the circumscribed rectangle of the region into three equal parts. A person recognition method characterized in that a person is recognized based on a relationship between (6) and an aspect ratio of a circumscribed rectangle of the change area. Feature (1); Number of histogram blocks in each circumscribed rectangle A1, A2, A3 Feature (2); Average value F of each block Feature (3); Difference between circumscribed rectangles of average value F of each block Feature (4); Maximum value number in each block Feature (5); Area ratio R in each block R Feature (6); Position M of maximum value in divided circumscribed rectangle A3