JPH06231252A - Method for tracking moving object on monitoring picture - Google Patents

Abstract

PURPOSE:To provide a monitored picture moving object tracking method capable of accurately tracking a moving object in a monitoring picture and improving performance for judging the existence of a moving object. CONSTITUTION:A picture monitoring device for monitoring a moving object by picking up the images of a monitoring area by an image pickup means, continuously entering and digitizing the picked up images at a prescribed time interval and computing the plural digitized images, a difference image of the digitized images is calculated, the calculated difference image is binarized to generate (S2) a binarized difference image, a change area is detected from the generated binarized difference image and registered in a change area gradation pattern dictionary as a moving object. After the lapse of a prescribed time, the contents of the gradation pattern dictionary are compared with the gradation pattern of an original image at the periphery of a position where the moving object on the original image inputted from the image pickup means exists, an area having high similirity is extracted and judged as a moving object to track the moving object.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a monitoring object moving object tracking method for tracking a moving object in a monitoring image in an image monitoring apparatus for monitoring a moving object in a monitoring area.

[0002]

2. Description of the Related Art An example of this type of image monitoring apparatus is shown in FIGS. 9 and 10, for example. In the figure, an ITV camera 1 as an image pickup means picks up an image in a monitoring area 2 and converts it into an electric signal. The image captured by the ITV camera 1 is processed by the transmission path 3 into the processing device 4,
Then, it is sent to a video tape recorder (hereinafter simply referred to as VTR) 5 as an image recording means. VTR
Reference numeral 5 indicates continuous monitoring images captured by the ITV camera 1,
Or moving object (hereinafter sometimes referred to as intruding object)
Record when detected.

On the other hand, the image signal sent to the processing device 4 is
The sampling pulse having a predetermined cycle output from the sampling pulse generation circuit 10 is converted into a digital signal by the A / D converter 11, and is stored in the image memory 12 as image data at time t = ti, for example.

Then, in the differential binarization circuit 13, the image data at the time t = ti is extracted before the image memory 12 in order to extract a change area, as will be described later.
The difference image is obtained by performing the inter-pixel difference calculation with the image data at time t = ti-1 that was captured in the difference binary image in which the changed pixel is represented by "1". And is stored in the differential binarized image memory 14.

A CPU (Central Processing Unit) 15 has a difference 2 in a difference binarized image memory 14.
When the changed area is analyzed using the binarized image and it is determined that the changed area is a moving object (intruding object), the alarm device 6 is sounded, or the supervisor moves the moving object on the screen of the display device 7. Is checked, and if the VTR 5 is not operating, it is operated to record a monitoring image. If it is determined that the object is not a moving object, the image is continuously captured and the above processing is repeated.

By the way, as an inter-pixel calculation for extracting a change area, in general, a difference from a background image when a moving object does not exist, or a time t before a predetermined time Δt.
Either one of the time-series differences from the image of = ti-1 is used. With the difference from the background image, it can be detected even when the moving object is stopped, but when the environment such as the brightness is changing, such as outdoors, the moving object exists unless the background image is updated. There is a feature that a change area is generated even if it is not used.

In the case of the time series difference, since the difference is generally made at a relatively short time interval, the followability to the environmental change is excellent, but it cannot be detected when the moving object stops, and further, Even if the moving object moves, the difference between the two screens is characterized in that both the disappearing portion and the occurring portion of the moving object are detected.

In order to eliminate the latter drawback of the time-series difference, in addition to the two screens of t = ti and t = ti-1,
= Ti + 1 consecutive three screens are used to perform the difference calculation on the front two screens and the rear two screens respectively, and by the logical product of the two difference images obtained, the moving object in the image t = ti In some cases, the method of extracting is taken.

In addition, when the CPU 15 analyzes a change area from a binary difference image, the value is "as a basic method of image processing in order to obtain a predetermined grouped area.
A serial number is given (labeling) to the closed area that is 1 ", and the area and distance of each partial area are calculated to extract a grouped area.
In many cases, the area such as 1 "is calculated and analyzed.
Performing this operation over the entire image area is extremely time consuming.

Therefore, for example, as disclosed in Japanese Patent Laid-Open No. 62-147891, a method has been proposed in which a changed area is roughly divided from a difference binary image. This is as shown in FIG. 11 in the difference binarized image.
The projection histogram is obtained in the X and Y axis directions, and in the obtained histogram, a segmented area whose height of the predetermined value Xth or Yth or more is equal to or more than the predetermined width Xw or Yw is obtained, and the inside is analyzed. Is a technique.

That is, the CPU 15 examines only the divided areas A11 to A22 shown in FIG. 11 and analyzes whether or not the changed area is a moving object. As described above, when the change area is analyzed, the labeling analysis can be performed, and as in the case of the divided areas A11 and A22 in FIG. 11, the divided area is often a circumscribed rectangle of the moving object. When the divided area has a predetermined size, it may be analyzed by a method such as directly inspecting its shape or obtaining the area of a portion whose value is "1" without performing processing such as labeling. .

The method of image difference and the method of analyzing a moving object have been described above. However, since the captured image contains a noise component, it is common to perform spatial filtering or logical filtering processing.

[0013]

In the conventional image monitoring apparatus, after the change area is determined to be a moving object, an alarm is sounded, or a video recording is performed on the VTR, and the like. It is determined whether or not the moving object moves from where and how it is a suspicious person. Therefore, compared with the initial image monitoring device in which the monitor always looks at the screen of the display device to find the moving object, the work is reduced, but it takes a considerable time for the monitor to make a judgment in the post-processing. Spend labor.

Therefore, not only the detection of the moving object but also the improvement of tracking the moving object and sounding the alarm only to a truly suspicious person can be expected to further reduce the work.

Therefore, according to the present invention, it is possible to accurately track a moving object in a monitoring image, improve the performance of determining whether or not the object is a moving object, and perform efficient image monitoring. The purpose is to provide a tracking method.

[0016]

According to the method of tracking a moving object in a surveillance image of the present invention, an image in the surveillance area is picked up by an image pickup means and inputted, and the inputted image in the surveillance area is set at predetermined time intervals. In the image monitoring apparatus for monitoring a moving object in the monitoring area, a differential image of the digitized image is obtained by continuously capturing and digitizing the digitized image and calculating a plurality of the digitized images. Then, the calculated difference image is binarized to generate a difference binarized image,
A change area is detected from the generated difference binarized image, the detected change area is used as a moving object, a grayscale pattern of the change area is stored in the storage means, and an original image input from the image pickup means after a predetermined time elapses. In the vicinity of the position where the moving object existed, by comparing the grayscale pattern in the storage means with the grayscale pattern of the original image, a region with high similarity is extracted to determine the moving object. It is characterized in that the moving object is tracked.

[0017]

According to the present invention, by tracking a moving object between time-series frames, the appearance location and destination of the moving object can be known, and whether the moving object appears and in which direction the true moving object It is possible to accurately determine whether or not the object is a moving object, and the performance of determining whether or not the object is a moving object can be improved. In addition, when tracking, it is calculated from the difference image so that it can be dealt with even when a plurality of moving objects move in an intersecting manner or when a large number of moving objects move by separating or merging. The grayscale pattern of the original image of the moving object is registered as dictionary information, and the original image and the dictionary information of the grayscale pattern are subjected to matching processing within the range of the preset prediction window, and a region having a high degree of similarity is set as a candidate. By using the moving area obtained from the difference image for selecting the candidate, more accurate tracking can be realized.

[0018]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. First, an outline of the present invention will be described before giving a specific description.

In order to analyze the movement of the moving object, it is considered that the moving object is extracted from the image at a certain time t = ti, and the corresponding area of the moving object is extracted from the images at the subsequent times. . That is, "tracking" means extraction of moving objects and extraction of corresponding parts (hereinafter referred to as tracking processing).
Can be divided into two steps. Where the first
The step of extracting a moving object can be realized by the method described in the conventional example. By adding the tracking process of the second step described in the present invention to the first step, a more effective image monitoring device can be realized as compared with the conventional example.

As described above, in order to track a moving object, it is necessary to correspond the moving object with images captured at different times. In order to take this correspondence, time t = t
An area (hereinafter referred to as a prediction window) in the image that is considered to have moved by the moving object at i at time t = ti + 1 is set, and corresponds to the moving object at time t = ti in the prediction window. The area may be extracted. That is, the method of setting the prediction window (the first point of the present invention) and the method of taking correspondence with the moving object at time t = ti within the prediction window at time t = ti + 1 (method of extracting the moving object) are important. Becomes

First, a method of setting the prediction window will be described. For example, as shown in FIG. 3A, time t = ti
It is assumed that the moving object is extracted by. Here, from the position of the moving object at time t = ti−1 and the position at time t = ti, the direction vector in which the moving object moves can be obtained. This direction vector is obtained, for example, by connecting the centers of gravity of the both. A movement prediction direction vector (arrow) 52 (x1> 0, y1) for predicting the next position from this direction vector
> 0).

In this case, since the moving object 51 has moved in the upper right direction from the previous frame (time t = ti−1), it should also move in the upper right direction in the next frame (time t = ti + 1). . Even when the direction is changed, the direction should gradually change if the image capturing interval is short, and the moving speed does not change so suddenly. Therefore, the prediction window may be set large in the upper right direction.

In practice, the current frame (time t = t
The circumscribed quadrangle 53 including the moving object 51 of i) is enlarged by setting W1 and W2 in the vertical direction with reference to the size of y1 and set W3 and W4 in the horizontal direction with reference to the size of x1. Then, the prediction window 54 is generated by enlarging.

Similarly, as shown in FIG. 3B, a moving object 55 is extracted, and a predicted movement direction vector (arrow) 56 (x2 <x2 <from the previous frame position and the current frame position is extracted.
0, y2> 0) is set, the circumscribing quadrangle 57 including the moving object 55 is enlarged in the vertical direction with reference to the size of y2, and the horizontal direction is the size of x2 (absolute value). The prediction window 58 is generated by enlarging the prediction window 58.

Next, a method of extracting a moving object within the prediction window will be described. The size, shape, and density of moving objects change as they move, but the moving time (image capture interval)
If is short, the change should be small.

Therefore, in the present invention, the grayscale pattern in the original image of the moving object at time t = ti is used as a reference pattern (hereinafter, referred to as a reference pattern).
(Referred to as a dictionary pattern), the dictionary pattern is moved in the prediction window to find the best matching grayscale pattern, and a region considered to be a moving object is extracted (second point of the present invention). Is this extraction good? A method of confirming whether or not there is a third point (the third point of the present invention) and a method of updating a dictionary pattern to cope with changes in size, shape, and density due to movement (the fourth point of the present invention) have been considered. .

First, a method of setting a dictionary pattern will be described. To obtain a grayscale pattern of a moving object means to perform a moving object extraction process and obtain a difference 2 of the extracted regions.
The binarized image is applied to the original image, only the data at the same position as the extraction region is left, and the others are set to "0", which is used as the dictionary pattern.

For example, FIG. 4 shows an example of a grayscale pattern of a moving object in the moving object extraction processing of the first step.
Here, the grayscale pattern of the moving area 32 in the input frame 31 is registered as a dictionary pattern. The dictionary area may be a dictionary pattern of a short shape including the background of a moving object obtained from the difference binarized image or a dictionary pattern of a person type excluding the background. It is more effective to create a person-type dictionary pattern excluding.

Next, a method of extracting a shade pattern in the prediction window will be described. For example, the image of FIG. 4 (time t = t
In the next frame of i), the image of FIG. 5 (time t = ti + 1)
Is taken in (the moving object 32 is moving to the position of 41). Assuming that the dictionary pattern 32 and the prediction window 33 are set at time t = ti as described above, the shade pattern and the dictionary pattern in the prediction window 42 (same as 33 of the previous frame) at time t = ti + 1. Compare with 32. Various comparison methods are conceivable, for example, a template matching method in which dictionary patterns are sequentially shifted within a prediction window and compared in a brute force manner.

By the above comparison (matching), the prediction window 4
When areas having the highest degree of similarity with the dictionary pattern 32 in 2 are obtained, these areas are considered to be moving objects.
It should be noted that it is possible to extract a plurality of candidates instead of one area having a high degree of similarity.

Further, in order to confirm whether or not the region considered to be a moving object really corresponds to the moving object, the above-mentioned moving object extraction method is combined. That is, time t = t
At i + 1, a difference binarized image (moving region is "1" at a predetermined threshold value and "0" other than that) is obtained by the same method as the conventional method, and is extracted from the difference binarized image. The validity of the extracted area is confirmed by comparing the area with the position of the area extracted using the original image in the prediction window to determine whether or not the object is a moving object.

Conventionally, the moving object is detected only by the difference binary image, but as described above, the matching process is performed on the original image, and the difference binary image is used to judge whether or not it is correct. By using it, it is very effective for tracking multiple people.
As a specific determination method, a ratio of a moving area included in an area having a high degree of similarity, a line omission in the X direction and the Y direction, and the like can be considered.

Next, a method of updating the dictionary pattern will be described. The update of the dictionary pattern is performed every frame because it corresponds to the change due to the movement of the moving object and the change in the size due to the perspective of the camera 1. That is, as shown in FIG. 6, the moving object in the current frame (time t = ti) has a large size or a small size in the next frame (time t = ti + 1), and thus has a high degree of similarity. It is necessary to enlarge or reduce the size of the dictionary pattern based on the area extracted as to update the dictionary pattern. The change in size, shape, and density should be smaller as the image capture interval is shorter, but the size changes considerably when the user steps on his / her foot or shakes his / her hand.

Therefore, when the dictionary pattern is updated, the differential binarized image is referred to, and if a moving area exists beyond the matching position as in the pattern 61 in FIG. 6, the size 62 of D1 to D3 is used. Expand the size of the dictionary pattern to the size of. In addition, pattern 6 in FIG.
When the moving area is smaller than the matching position as in 3, the size of the dictionary pattern is reduced to the size of 64 of D4 to D6.

At this time, the person dictionary described above may be rectangular information including a background or a person type excluding the background. As described above, by updating the dictionary pattern for each frame and using the dictionary pattern and the size of the previous frame in matching, the matching performance in the next frame can be improved. In addition to enlarging or reducing the dictionary pattern using the difference image, it is also possible to change the size of the dictionary pattern using affine transformation.
Next, one embodiment of the present invention will be specifically described.

FIG. 1 shows the structure of an image monitoring apparatus for carrying out the present invention. The difference between the method of the present invention and the conventional method is that t = ti stored in the image memory 12.
Difference image is obtained by performing the difference calculation between the pixels of the image of No. 1 and the image captured before that, and the difference 2 in which the changed pixel in the difference binarization circuit 13 is represented by "1". After conversion into a binarized image, the CPU 15 obtains a change area and determines that the object is a moving object, the moving object is registered (stored) in a person dictionary (storage means) 16 in a light and shade pattern, and the next frame and thereafter. Then, the prediction window is set in the moving direction of the moving object, the dictionary pattern in the entry dictionary 16 is matched with the original image in the prediction window, and the moving object is moved to the region having the highest degree of similarity. That is, the moving object is tracked.

The position of the moving object and the size of the prediction window in each frame are stored in the tracking table 17. Not only the detection of a moving object, but the history of the moving object is obtained so that the appearance location and the destination of the moving object can be known, and even when two or more people cross each other, they can be tracked. It will be possible.

Next, the procedure for tracking a moving object will be described with reference to the flow chart shown in FIG. The processing is
It is roughly divided into new object extraction processing and tracking processing. First, the new object extraction process will be described. Enter the image (S
1) The difference binarization circuit 13 generates a difference binarized image using the time series difference or the background difference method (S2). Next, if the flag (new object extraction or tracking success) stored in the tracking table 17 is checked and there is no tracking object (S3), the generated differential binary image is used to determine the X and Y directions. A moving object is extracted by projecting onto the object or extracting a closed region (S4).

If there is a moving object (S5), as described above, a person dictionary pattern of the shade pattern of the extracted moving area is generated and registered in the person dictionary 16 (S6). The extracted moving object appears from the edge of the frame or the shadow of the object, but since it is possible to predict the direction of the next movement depending on the appearance site, the prediction window as described above is set in the predicted movement direction. The flag is set and stored in the tracking table 17 together with the new object extraction flag (FLG) and position information for tracking in the next frame and later (S7).

As for the movement prediction direction, considering that there is no shadow of a building or the like in the frame, the appearance of a moving object is predicted from four frame edges (top and bottom, left and right (or two edges such as upper right)). , (0, -1), (0, 1), (1, 0), (-1,
0) and ((X, Y): X direction, Y direction) movement prediction directions are set ((−1, −1) in the case of upper right).

In FIG. 4, since the moving object 32 has emerged from the lower part of the frame 31, the predicted movement direction (arrow) 34
(0, 1) is given and the prediction window 33 is set. The size of the prediction window 33 must be taken into consideration by the image capturing interval and the moving speed of the moving object, but since the moving speed of the newly appearing moving object is unknown, a larger prediction window is set.

Further, when the moving object is tracked in the next frame, the direction vector from the previous frame position to the next frame position is used as the predicted motion direction. For example, in FIG.
In the next frame of the image of FIG.
If it is moved to the position of, the movement prediction direction is an arrow 43.

Next, the tracking process will be described. In the flowchart of FIG. 2, after inputting an image (S1), a differential binarized image is generated (S2), and a flag (new object extraction or tracking success) stored in the tracking table 17 is checked to find a tracking object. If there is (S3), tracking processing is performed by the following method.

First, the prediction window 42 set in the previous frame
The light and shade pattern inside is compared with the light and shade pattern registered in the person dictionary 16 (portion 32 in FIG. 4) (S8,
First matching process). Here, for example, the template matching method is used to extract the region in the prediction window 42 that has the highest degree of similarity with the grayscale pattern registered in the person dictionary 16.

However, since it is possible that the shape and density of the moving object may change as the moving object moves, or the size may change slightly due to the difference in distance from the camera 1, a region having a high degree of similarity does not necessarily become a moving object. . Therefore, as a method of determining whether the extracted region is a moving object,
As described above, the determination is performed by comparing the position of the extracted region in the difference binarized image with the position of the region thought to be a person extracted using the original image in the prediction window.

In addition to the determination using the above-described differential binary image, the moving direction of the moving object up to that time (3 in FIG. 4).
4) and the direction vector (43 in FIG. 5) from the previous frame position to the presently extracted position may be compared for reference, but the moving object does not necessarily move in the same direction at the same speed. It is not limited to moving, so if there are other candidates, it is better to use them in the same direction. But,
The determination using this direction is effective when a similar moving object moves from the opposite direction and crosses and passes.

In the first matching process (S8), the comparison result is determined based on the difference between the line omission in the difference area, the ratio of the moving area, and the direction vector. If it is determined that the matching has failed, it is necessary to track the moving object by another method. Here, when the first matching process fails, the logical sum of the difference binarized image and the original image is obtained, and the matching process is performed again on the original image in the difference area (S9, second matching process). Also in this case, the template matching method is used similarly to the first matching processing, and the evaluation method also uses the line omission in the difference area, the ratio of the moving area, and the difference in the direction vector.

Further, if the matching also fails in the second matching process, the size of the prediction window is changed (S10). In particular, with emphasis on the moving direction, areas other than the moving direction are reduced, the prediction window is set large in the moving direction, and the matching process is performed again on the original image of the difference area (S).
11, third matching processing). For example, when a moving object is present in the previous frame at the position 41 in FIG. 5 and the prediction window 45 is set and the matching fails, a larger prediction window 44 is set and the matching process is performed.

The evaluation method of the third matching processing is also the first.
Similar to the matching process of No. 3, and if the third matching process also fails, it is possible to search in another range, but here, the tracking is terminated.

A moving object that has failed tracking should be extracted as a new moving object in a new object extracting process (S4 and later) to be performed later, and a post-processing (S12) associates the tracking-failed object with the new object. Keep track of things.

If the matching with the dictionary pattern is successful in the above-described first to third matching processes and the moving object can be tracked, the dictionary pattern is updated and registered in the person dictionary 16. (S14), the prediction window and the tracking success flag used in the next frame are set (S1).
5), save in the tracking table 17. On the other hand, the first to
When the matching with the dictionary pattern fails in the third matching process, the tracking failure flag is stored in the tracking table 17 (S16).

After the tracking process is completed, a new moving object is further extracted from the image obtained by deleting the tracking object extracted by the tracking process from the difference binary image (S4). Figure 7
FIG. 7A is a differential binarized image of a ti frame, FIG. 7B is a differential binarized image of a (ti + 1) frame, and FIG. 7C is “binary differential binarization of a (ti + 1) frame. An example of "image-tracking object" is shown. The range of the tracking object to be deleted may be a person type excluding the background or a rectangle including a person. Here, a rectangle including a person is used. The subsequent "new object extraction process" performed on the image of "(ti + 1) frame difference binary image-tracking object" is as described above.

After the tracking process and the new object extraction process are completed, post-processing (S12) is performed. In the post-processing, when a moving object that has failed in tracking is extracted as a new moving object, tracking is successful, but when it is determined as another moving object, it is checked using the information stored in the tracking table 17. And correct the result. Then, after the post-processing, the processing ends (S13). An example of the tracking table 17 is shown in FIG.

Although the tracking process has been described above, not only a moving object is present, but by tracking the moving object by the above method, an intruding object (intruder) heading for a dangerous area or an unauthorized area, or It is possible to detect an intruding object that is moving suspiciously, and when an suspicious intruding object is detected, sound an alarm to enable effective monitoring. Further, even when a plurality of intruders cross each other in the frame, they can be tracked because the matching process is performed using the original image.

As described above, according to the above embodiment, when the moving object is tracked by the image monitoring apparatus, the region is extracted from the difference image, and the dictionary information of the grayscale pattern is used to perform the matching process with the original image in the prediction window. By doing, it is possible to know where the moving object has moved, and to track the moving object with high accuracy.

Further, by having the dictionary information of the original picture, even when there are a plurality of moving objects, it is possible to correlate them.
Even if the object is moved in a different direction, the same object can be determined by the matching process with the original image. In the matching process with the original image, a search window can be shortened by providing a prediction window according to the moving speed and direction of the moving object and limiting the range to perform the matching process.

Furthermore, the person dictionary can be improved in matching accuracy in each frame by expanding or reducing the size using the difference image and updating it for each frame.

From the above, it is possible to realize an efficient image monitoring apparatus by tracking a moving object as in the present invention as compared with a conventional apparatus for detecting a moving object.

[0059]

As described in detail above, according to the present invention, it is possible to accurately track a moving object in a monitoring image, improve the performance of determining whether or not the object is a moving object, and perform efficient image monitoring. It is possible to provide a possible moving object tracking method of a monitoring image.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of an image monitoring apparatus to which an embodiment of the present invention is applied.

FIG. 2 is a flowchart for explaining tracking processing.

FIG. 3 is a diagram illustrating a method of setting a prediction window in tracking processing.

FIG. 4 is a diagram showing an example of an image of a ti frame in a tracking process.

FIG. 5 is a diagram showing an example of an image of a (ti + 1) frame in a tracking process.

FIG. 6 is a diagram illustrating a method of updating a person dictionary in tracking processing.

FIG. 7 is a diagram for explaining a new object extraction process in the tracking process.

FIG. 8 is a diagram showing an example of a tracking table in tracking processing.

FIG. 9 is a configuration diagram showing an example of an image monitoring device.

FIG. 10 is a block diagram showing a configuration of a conventional image monitoring device.

FIG. 11 is a diagram illustrating a region extraction method using projection.

[Explanation of symbols]

1 ... ITV camera, 2 ... monitoring area, 3 ... transmission path,
4 ... Processing device, 5 ... VTR, 6 ... Alarm device, 7 ...
Display device, 10 sampling pulse generation circuit, 1
1 ... A / D converter, 12 ... Image memory, 13 ... Difference binarization circuit, 14 ... Difference binarized image memory, 15 ...
... CPU, 16 ... person dictionary, 17 ... tracking table.

Claims (4)

[Claims] 1. An image in a surveillance area is picked up by an image pickup means and inputted, and the inputted image in the surveillance area is continuously taken in at a predetermined time interval and digitized, and the digitized plurality of images are obtained. In the image monitoring apparatus that monitors the moving object in the monitoring area by calculating the image of the difference image, the difference image of the digitized image is calculated, and the calculated difference image is binarized to obtain the difference binary value. A changed image is generated, a change region is detected from the generated difference binarized image, the detected change region is used as a moving object, and a grayscale pattern of the change region is stored in a storage unit, and after a predetermined time has elapsed, the image pickup unit. In the vicinity of the position where the moving object exists in the original image input from, the grayscale pattern in the storage means and the grayscale pattern of the original image are compared to obtain a high degree of similarity. Moving object tracking method of monitoring an image, characterized in that tracking the moving object by determining said moving object by extracting frequency. 2. A tracking window centering on the current position is set according to the speed obtained from the frame position immediately before the moving object and the current position after tracking the moving object, and the prediction window is set in the next frame. 2. The method according to claim 1, wherein the grayscale pattern in the storage means is compared with the grayscale pattern of the original image in the prediction window. 3. The gray level pattern in the storage means is compared with the gray level pattern of the original image to extract an area having a high degree of similarity, and then the moving area obtained from the difference image is used to obtain the high degree of similarity. The moving object of the monitoring image according to claim 1, wherein it is determined whether or not the extracted area is a valid tracking object based on a ratio of the moving area included in the area or a shape of the moving area. Tracking method. 4. A grayscale pattern in the storage means and a grayscale pattern of the original image are compared to extract a region having a high degree of similarity, and the extracted region is enlarged or reduced using the difference image, 2. The moving object tracking method for monitoring images according to claim 1, wherein this is updated and stored for each frame in the storage means as a grayscale pattern for tracking the moving object in the next frame.