JPH06327005A - Moving object detecting device - Google Patents

Abstract

PURPOSE:To detect a moving object without necessitating a background picture, and generating uncovered background. CONSTITUTION:Frame delay circuits 2 and 4 are serially continuing frames, and pixel data at the spatial same position are extracted. The data of the (n)th frame are subtracted from the data of the(n-1)th frame by a subtracting circuit 3, and a difference signal is supplied through an absolute value circuit 6 to a comparator circuit 7. The comparator circuit 7 generates an output '1' when the absolute value of the difference signal is more than a threshold value TH. The similar processing is operated to the (n-2)th frame and the (n-1)th frame, and the output is generated from a comparator circuit 9. A switch circuit S1 is controlled by a discriminating circuit 10, so that when the output signals of the comparator circuits 7 and 9 are both '1', the output signal of the frame delay circuit 2 can be selected, and when either of them is '0', the signal of a '0' level can be selected.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a moving object detecting device for detecting a moving object using an interframe difference signal of a digital image signal or an interframe vector signal.

[0002]

2. Description of the Related Art The most classical method for detecting a moving object from a digital image signal is to use a differential signal between two consecutive frames. this is,
The difference signal of the pixel data at the same position between the two frames is subjected to the threshold value processing, so that a pixel having a large difference is determined as a moving pixel and a pixel having a small difference is determined as a still pixel.

Further, an image called a background image which does not include any moving object is prepared by some method, and a moving object is detected by thresholding a difference signal between the background image and the current frame image. A method called, has also been proposed.

[0004]

However, in the method using the differential signal between two consecutive frames, the background appearing from the shadow of the moving object called the uncovered background is used as the moving pixel in addition to the true moving pixel. There is a problem that it is erroneously detected.

This will be described with reference to FIG. An image of a ball moving in front of a stationary background, Figure 2a
The image data in the nth frame, FIG.
1) Image data in a frame. According to a method of performing motion detection using a differential signal between two consecutive frames, that is, a differential signal obtained by subtracting data of the nth frame from data of the (n−1) th frame, the result is 2c, it can be seen that the uncovered background is erroneously detected as a moving object in addition to the true moving object.

On the other hand, the method of preparing the background image in the image for detecting the moving object and using the difference signal between this background image and the current frame image does not cause the problem of uncovered background. However, depending on the application, it may be difficult to create the background image in advance. Further, the background image needs to be sequentially updated in order to change the illumination, remove the influence of the shadow, and the like, but it is actually difficult to ideally perform this update. For example,
It takes some time for the background image to be updated to the complete state even after the image returns to the stationary state after the illumination change and the brightness change due to the shadow. Therefore, the determination accuracy deteriorates until the complete state is updated.
Further, the method using the difference signal between the background image and the current frame image has a problem that motion is not detected when the background image and the moving object have the same level.

SUMMARY OF THE INVENTION Therefore, an object of the present invention is to provide a moving object detecting apparatus which does not use the background image and does not cause the problem of uncovered background.

[0008]

According to a first aspect of the present invention, there is provided a moving object detecting apparatus for detecting a moving object by using an inter-frame difference signal of a digital image signal.
-2m) a circuit for detecting a first inter-frame difference signal between the frame image and the (n-m) th frame image, and a second inter-frame difference between the (n-m) th frame image and the n-th frame image A moving object detection device comprising a circuit for detecting a signal and a circuit for detecting a moving object in a (n−m) th frame using the first and second difference signals. is there.

According to a second aspect of the present invention, in a moving object detecting device for detecting a moving object using an inter-frame vector signal of a digital image signal, a (n-2m) th frame image and a (n-m) th frame image are provided. A circuit for detecting a first interframe vector signal of an image, a circuit for detecting a second interframe vector signal of an (n-m) th frame image and an nth frame image, And a circuit for detecting a moving object in the (n−m) th frame using the second vector signal.

[0010]

Function: Motion detection is performed by using image signals of three consecutive frames. M at each pixel of the digital image signal
= 1, the (n-2) th frame image and the (n-) th frame image
1) The inter-frame difference signal of the frame image and the (n-
1) Only when both the inter-frame difference signals of the frame image and the nth frame image are large, the pixel is determined to be a moving pixel. When only one of the difference signals is large, it means that an uncovered background appears, and therefore the pixel is determined as a still pixel. As described above, the problem of uncovered background can be avoided by the method of detecting motion using the image signals of three consecutive frames, that is, the inter-frame difference signal.

As another method of detecting motion using image signals of three consecutive frames, in each pixel of the data image signal, a small area centered on the pixel of interest is the (n-2) th frame. The motion vector signals of the image and the (n-1) th frame image and the motion vector signals of the (n-1) th frame image and the nth frame image are calculated, and both of the calculated motion vector signals are (0, 0 If it is not, the pixel is determined to be a moving pixel.

When both of the calculated motion vector signals are (0,0), or when one of the calculated motion vector signals is not (0,0), it means that an uncovered background appears, and The pixel is determined to be a stationary pixel. As described above, the problem of uncovered background can be avoided by the method of performing motion detection using image signals of three consecutive frames, that is, inter-frame vector signals. In addition, erroneous determinations due to illumination changes and shadows, which are likely to occur in the method using the interframe difference signal as the determination reference, can be greatly reduced by the method using the interframe vector signal as the determination reference.

[0013]

Embodiments (1) First Embodiment A first embodiment of the motion detecting device according to the present invention will be described below. The first embodiment shows the overall structure of a motion detection device using inter-frame difference signals. In FIG. 1, reference numeral 1 is an input terminal to which a digital image signal is supplied. The digital image signal is supplied to the frame delay circuit 2. The frame delay circuit 2 delays the digital image signal by one frame. The output signal of the frame delay circuit 2 and the digital image signal from the input terminal 1 are supplied to the subtraction circuit 3.

In the subtraction circuit 3, the digital image signal from the input terminal 1, that is, the image signal of n frames is subtracted from the output signal of the frame delay circuit 2, that is, the image signal of (n-1) frame. This subtraction process
This is done for the levels of two pixels at the same spatial position. The output signal of the subtraction circuit 3 is supplied to the absolute value conversion circuit 6.

The absolute value conversion circuit 6 converts the data into absolute values. The output signal of the absolute value conversion circuit 6 is supplied to the comparison circuit 7 and compared with a preset threshold value TH.
When the output signal of the absolute value conversion circuit 6 is larger than the threshold value TH, the comparison circuit 7 outputs `1`, otherwise,
The comparison circuit 7 outputs "0". The output signal of the comparison circuit 7 is supplied to the determination circuit 10.

The circuit operation described above will be described with reference to FIG. The image shown in FIG. 3a is an image signal of n frames,
That is, it is assumed that the digital image signal is supplied from the input terminal 1 to the subtraction circuit 3. On the other hand, it is assumed that the image shown in FIG. 3b is an (n-1) frame image signal, that is, a digital image signal supplied from the frame delay circuit 2 to the subtraction circuit 3. Subtraction is performed by the subtraction circuit 3, absolute value conversion is performed by the absolute value conversion circuit 6, and comparison processing is performed by the comparison circuit 7. As an output signal of the comparison circuit 7, the image signal shown in FIG. That is, when the absolute value of the difference signal is larger than the threshold value TH, an image signal which becomes "1" and which otherwise becomes "0" is obtained.

The output of the frame delay circuit 2 is also supplied to the frame delay circuit 4. The frame delay circuit 4 delays the digital image signal by one frame. The output signal of the frame delay circuit 4 and the output signal of the frame delay circuit 2 are supplied to the subtraction circuit 5.

The subtraction circuit 5 subtracts the output signal of the frame delay circuit 4, that is, the image signal of (n-2) frames from the output signal of the frame delay circuit 2, that is, the image signal of (n-1) frames. It The output signal of the subtraction circuit 5 is supplied to the absolute value conversion circuit 8.

The absolute value conversion circuit 8 converts the data into an absolute value. The output signal of the absolute value conversion circuit 8 is supplied to the comparison circuit 9 and compared with a preset threshold value TH. If the output signal of the absolute value conversion circuit 8 is larger than the threshold value TH, the comparison circuit 9 outputs `1`, and if not, the comparison circuit 9 outputs` 0`. The output of the comparison circuit 9 is supplied to the determination circuit 10.

The above circuit operation will be described with reference to FIG. It is assumed that the image shown in FIG. 4a is an (n-1) frame image signal, that is, a digital image signal supplied from the frame delay circuit 2 to the subtraction circuit 5. On the other hand, it is assumed that the image shown in FIG. 4b is an (n−2) th frame image signal, that is, a digital image signal supplied from the frame delay circuit 4 to the subtraction circuit 5. Subtraction circuit 5
The subtraction, the absolute value conversion circuit 8 performs absolute value conversion, and the comparison circuit 9 performs comparison processing. As an output signal of the comparison circuit 9, the image signal shown in FIG. That is, when the absolute value of the difference signal is larger than the threshold value TH,
An image signal which becomes "1" and which otherwise becomes "0" is obtained.

The determination circuit 10 outputs "1" when the output signals of the comparison circuit 7 and the comparison circuit 9 are both "1", and outputs "0" otherwise. That is, the determination circuit 10 is an AND gate.

The output signal of the frame delay circuit 2, that is, the image signal of the (n-1) th frame is also supplied to one input terminal of the switch circuit S1. A signal of level "0" from the ground 11 is supplied to the other input terminal of the switch circuit S1.

The switch circuit S1 is controlled by the output signal of the determination circuit 10, and the final output signal is taken out from this. In the switch circuit S1, the output signal of the determination circuit 10 is
When it is `1`, the output signal of the frame delay circuit 2, that is, the image signal of (n-1) frame is outputted as the final output signal, and when the output signal of the judgment circuit 10 is` 0`, the signal from the ground 11 (`0 'level signal) is output as the final output signal. The output signal of the switch circuit S1 is taken out to the output terminal 12.

The circuit operation described above will now be described with reference to FIG. As described above, the image shown in FIG. 5a is the output signal of the comparison circuit 7 (FIG. 3c), and the image shown in FIG. 5b is the output signal of the comparison circuit 9 (FIG. 4c). The determination circuit 10 performs an AND gate process on the output signal of the comparison circuit 7 and the output signal of the comparison circuit 9. The final output image shown in FIG. 5c is obtained by switching the (n-1) frame image and the `0 'level signal based on the result output from the determination circuit 10.

(2) Second Embodiment Next, a second embodiment of the present invention will be described. The second embodiment shows the overall structure of a motion detection device using interframe vectors. In FIG. 6, reference numeral 21 is an input terminal to which a digital image signal is supplied. The digital image signal is supplied to the frame delay circuit 22.
The frame delay circuit 22 delays the digital image signal by one frame. The digital image signal from the input terminal 21 is supplied to the area cutout circuit 23a, and the output signal of the frame delay circuit 22 is supplied to the area cutout circuit 23b.

The area cutout circuit 23a extracts an area centered on the pixel of interest indicated by ● as shown in FIG. 7 from the digital image signal supplied from the input terminal 21, that is, the image signal of n frames. Here, it is assumed that the pixels of the (5 × 5) block centering on the pixel of interest are extracted. Further, the area cutout circuit 23b, like the area cutout circuit 23a, extracts the pixels of the (5 × 5) block from the digital image signal supplied from the frame delay circuit 22, that is, the image signal of the (n−1) th frame. To do.
The output signal of the area cutting circuit 23a and the area cutting circuit 2
3b output signal, that is, an image signal of n frames and (n
-1) The image signal of the frame is the motion vector calculation circuit 2
6 is supplied.

The motion vector calculation circuit 26 is for calculating a so-called pixel-based motion vector in a block centered on the pixel of interest. The motion vector represented by (x, y) is composed of a horizontal motion vector x and a vertical motion vector y. In the motion vector calculation circuit 26, the block centering on the pixel of interest of the n frame extracted by the area cutout circuit 23a using the block matching method and the pixel of interest of the (n-1) frame extracted by the area cutout circuit 23b. An inter-frame vector is calculated with respect to the block centered at. In the motion vector calculation circuit 26, the vector detection range is, for example, ± 10 pixels in both the horizontal and vertical directions.

Although it is described here that the block matching method is used in the calculation of the inter-frame vector, in practice, the calculation method of the motion vector is not particularly limited, and the gradient method or the phase correlation method may be used. If the calculated motion vector is (0,0), the motion vector calculation circuit 26
If `0 'is output and a result different from (0,0) is calculated,` 1' is output. Motion vector calculation circuit 26
Is supplied to the decision circuit 28.

On the other hand, the output of the frame delay circuit 22 is also supplied to the frame delay circuit 24. The frame delay circuit 24 delays the digital image signal by one frame. The output signal of the frame delay circuit 22 is supplied to the area cutout circuit 25a, and the output signal of the frame delay circuit 24 is supplied to the area cutout circuit 25b.

The area cutout circuits 25a and 25b have the same functions as the area cutout circuits 23a and 23b, and focus on the pixel of interest from the output signal of the frame delay circuit 22, that is, the image signal of the (n-1) th frame. The extracted area is extracted. Here, similarly to the area cutout circuit 23, it is assumed that the pixels of the (5 × 5) block centering on the target pixel are extracted. The output signal of the area cutout circuit 25 a and the output signal of the area cutout circuit 25 b, that is, the image signal of the (n−1) frame and the image signal of the (n−2) frame are supplied to the motion vector calculation circuit 27.

The motion vector calculation circuit 27 has a function similar to that of the motion vector calculation circuit 26. The motion vector calculation circuit 27 has a block centered on the pixel of interest of the (n-1) frame extracted by the region extraction circuit 25a and a region. Cutting circuit 25
An inter-frame vector is calculated between the block centered on the pixel of interest of the (n−2) frame extracted by b. The motion vector calculation circuit 27 outputs `0 'when the calculated motion vector is (0, 0), and outputs (0,
If a result different from 0) is calculated, `1 'is output. The output of the motion vector calculation circuit 27 is supplied to the determination circuit 28 like the output of the motion vector calculation circuit 26.

Both the output signal of the motion vector calculation circuit 26 and the output signal of the motion vector calculation circuit 27
When 1 ′ is supplied to the judgment circuit 28, the judgment circuit 28
Outputs `1`, and if not, the decision circuit 2
8 outputs `0`. That is, the determination circuit 28
It is an AND gate.

The output signal of the frame delay circuit 22, that is, the image signal of the (n-1) th frame is the switch circuit S2.
Is also supplied to one input terminal of. A signal of level "0" is supplied from the ground 29 to the other input terminal of the switch circuit S2.

The switch circuit S2 is controlled by the output signal of the judging circuit 28, and the final output signal is taken out from this. The output signal of the determination circuit 28 is output to the switch circuit S2.
When it is `1`, the output signal of the frame delay circuit 22, that is, the image signal of (n-1) frame is output as the final output signal, and when the output signal of the judgment circuit 28 is` 0`, the signal from the ground 29. (`0 'level signal) is output as the final output signal. The output signal of the switch circuit S2 is taken out to the output terminal 30.

In the above-described second embodiment, the block centering on the pixel of interest is extracted by the area cutout circuit and the motion vector (1 pixel precision) is calculated for that block. The motion may be determined by dividing the screen into blocks like a blocking circuit and calculating motion vectors between the blocks. In this method, the accuracy of the motion determination is the size of the divided blocks, so the accuracy is somewhat degraded, but the amount of calculation can be significantly reduced.

As described above, in the second embodiment, the determination using the interframe vector is performed instead of the determination using the interframe difference described in the first embodiment.

Comparing the first embodiment with the second embodiment, since the brightness of the entire screen generally changes in the illumination determination, the motion using the inter-frame difference of the first embodiment is used. / In the stillness determination, even if no motion has occurred, it is likely to be determined to be motion due to a change in the brightness value. However, when the interframe vector of the second embodiment is used, the brightness value basically changes, although the brightness value changes.
Since the change uniformly removes the offset, when the motion vector is obtained, when there is no motion, it is calculated as a (0, 0) vector. Therefore, not only the interframe difference but also the motion / stillness determination by determining whether the interframe vector is (0, 0) can realize a moving object detection device that is resistant to illumination changes.

On the other hand, when the change of the brightness value due to the shadow occurs, unlike the case of the change of the illumination, the change in the brightness value of the entire screen is unlikely to be uniform. However, when observed in a local area, in the local area,
It can be regarded as a roughly uniform change in the luminance value, and if the block for calculating the interframe vector is made small, whether the interframe vector is (0, 0) or not as in the case of the above-mentioned illumination change. By performing the motion / stillness determination based on such determination, it is possible to realize a moving object detection device that is resistant to changes in the brightness value due to shadows.

However, when the judgment is made only by the normal motion vector between the two frames, the problem of uncovered background cannot be avoided as in the case where the judgment is made by the frame difference between the two frames. Absent. Therefore, in the method of the present invention, continuous / three frames are used to perform motion / still determination based on the inter-frame difference and inter-frame vector, and it is resistant to changes in brightness value due to changes in illumination, shadows, etc.
We have realized a moving object detection method that does not generate background.

[0040]

According to the invention of the first embodiment, since the background image is not required, the influence of the illumination change and the shadow is small. In particular, since the difference between the frames which is equal to or more than the frame interval is not used, the influence of the illumination change and the shadow does not have a time lag. Further, there is an advantage that an uncovered background such as a motion detection using a difference between two continuous frames does not occur.

In addition, according to the second embodiment of the invention, the change in illumination is used for determining the motion / stillness, not the conventional interframe difference but whether or not the interframe vector is (0, 0). , It is possible to greatly reduce the false judgment due to the shadow. Also,
Since the determination is performed using the data for three consecutive frames, there is an advantage that an uncovered background such as the motion detection using the data for two ordinary frames does not occur.

[Brief description of drawings]

FIG. 1 is an example of a block diagram according to the first embodiment of the invention.

FIG. 2 is a diagram for explaining an operation of a conventional motion detection method using a difference between two consecutive frames.

FIG. 3 is a diagram for explaining the operation of the motion detection method according to the present invention.

FIG. 4 is a diagram for explaining the operation of the motion detection method according to the present invention.

FIG. 5 is a diagram for explaining the operation of the motion detection method according to the present invention.

FIG. 6 is an example of a block diagram according to the invention of a second embodiment.

FIG. 7 is a schematic diagram for explaining a pixel of the motion detection method of the present invention.

[Explanation of symbols]

 2,4 Frame delay circuit 3,5 Subtraction circuit 7,9 Comparison circuit 10 Judgment circuit

Claims (7)

[Claims] 1. A moving object detecting apparatus for detecting a moving object using an inter-frame difference signal of a digital image signal, wherein a (n-2m) th frame image and a first (n-m) th frame image are inter-framed. Means for detecting a difference signal, means for detecting a second inter-frame difference signal between the (n−m) th frame image and the nth frame image, and the first and second difference signals described above. Using the (n-
m) A means for detecting a moving object in a frame, and a moving object detecting device. 2. A moving object detecting device for detecting a moving object using an inter-frame vector signal of a digital image signal, wherein a (n-2m) th frame image and a first frame between the (nm) th frame image are detected. Means for detecting a vector signal, means for detecting a second inter-frame vector signal of the (n−m) th frame image and the nth frame image, and the above first and second vector signals Using the (n
-M) A moving object detection device, characterized in that it comprises a means for detecting a moving object in a frame. 3. The moving object detecting device according to claim 1 or 2, wherein m = 1. 4. The moving object detecting device according to claim 1, wherein a pixel in which both the first and second difference signals are larger than a threshold value is detected as a motion. 5. The moving object detection device according to claim 1, wherein when both the first and second difference signals are larger than a threshold value, the input digital image signal is selected, and when not, a predetermined fixed value is set. A moving object detection device comprising means for selecting. 6. The moving object detecting device according to claim 2, wherein a pixel in which both the first and second vector signals are larger than zero is detected as a motion. 7. The moving object detection device according to claim 2, wherein when both the first and second vector signals are larger than zero, the input digital image signal is selected, and when not, a predetermined fixed value is selected. A moving object detection device comprising means.